diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 27da2a18f1aeab619fdaa58f43db094d27d32768..8a1df5406fa12c14a650e02dab37e062795bef92 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": "20260604T120021Z", - "generated_at": "2026-06-04T12:10:12Z", + "snapshot_id": "20260604T180024Z", + "generated_at": "2026-06-04T18:11:31Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1148,364 +1148,6 @@ } ] }, - { - "cluster_id": "cluster-45363-3", - "summary": "Cluster of 3 related pull requests linked by soft_similarity.", - "status": "open", - "confidence": 0.5, - "canonical_issue_number": null, - "canonical_pr_number": 46339, - "issue_numbers": [], - "pr_numbers": [ - 45363, - 45666, - 46339 - ], - "evidence_types": [ - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 45363, - "right_pr_number": 45666, - "code_similarity": 0.392, - "size_similarity": 0.298, - "file_overlap": 0.571, - "area_overlap": 0.329, - "patch_similarity": 0.306, - "shared_filenames": [ - "fused_qwen_example.py", - "src/transformers/integrations/__init__.py", - "src/transformers/integrations/hub_kernels.py", - "src/transformers/modeling_utils.py" - ], - "shared_file_areas": [ - { - "filename": "fused_qwen_example.py", - "left_ranges": [ - [ - 1, - 30 - ] - ], - "right_ranges": [ - [ - 1, - 49 - ] - ] - }, - { - "filename": "src/transformers/integrations/__init__.py", - "left_ranges": [ - [ - 67, - 81 - ], - [ - 226, - 240 - ] - ], - "right_ranges": [ - [ - 67, - 73 - ], - [ - 222, - 228 - ] - ] - }, - { - "filename": "src/transformers/integrations/hub_kernels.py", - "left_ranges": [ - [ - 11, - 17 - ], - [ - 22, - 35 - ], - [ - 505, - 716 - ] - ], - "right_ranges": [ - [ - 11, - 53 - ], - [ - 286, - 295 - ], - [ - 521, - 1077 - ] - ] - }, - { - "filename": "src/transformers/modeling_utils.py", - "left_ranges": [ - [ - 3738, - 3747 - ] - ], - "right_ranges": [ - [ - 1330, - 1338 - ], - [ - 3763, - 3789 - ], - [ - 4223, - 4234 - ], - [ - 4628, - 4641 - ] - ] - } - ] - }, - { - "left_pr_number": 45363, - "right_pr_number": 46339, - "code_similarity": 0.504, - "size_similarity": 0.842, - "file_overlap": 0.286, - "area_overlap": 0.591, - "patch_similarity": 0.291, - "shared_filenames": [ - "src/transformers/integrations/hub_kernels.py", - "src/transformers/modeling_utils.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/integrations/hub_kernels.py", - "left_ranges": [ - [ - 11, - 17 - ], - [ - 22, - 35 - ], - [ - 505, - 716 - ] - ], - "right_ranges": [ - [ - 14, - 51 - ], - [ - 284, - 293 - ], - [ - 520, - 744 - ] - ] - }, - { - "filename": "src/transformers/modeling_utils.py", - "left_ranges": [ - [ - 3738, - 3747 - ] - ], - "right_ranges": [ - [ - 1346, - 1354 - ], - [ - 3785, - 3811 - ], - [ - 4245, - 4257 - ], - [ - 4651, - 4664 - ] - ] - } - ] - }, - { - "left_pr_number": 45666, - "right_pr_number": 46339, - "code_similarity": 0.423, - "size_similarity": 0.354, - "file_overlap": 0.667, - "area_overlap": 0.283, - "patch_similarity": 0.356, - "shared_filenames": [ - "src/transformers/core_model_loading.py", - "src/transformers/integrations/hub_kernels.py", - "src/transformers/modeling_utils.py", - "tests/kernels/test_kernels.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/core_model_loading.py", - "left_ranges": [ - [ - 19, - 25 - ], - [ - 80, - 86 - ] - ], - "right_ranges": [ - [ - 19, - 25 - ], - [ - 80, - 86 - ] - ] - }, - { - "filename": "src/transformers/integrations/hub_kernels.py", - "left_ranges": [ - [ - 11, - 53 - ], - [ - 286, - 295 - ], - [ - 521, - 1077 - ] - ], - "right_ranges": [ - [ - 14, - 51 - ], - [ - 284, - 293 - ], - [ - 520, - 744 - ] - ] - }, - { - "filename": "src/transformers/modeling_utils.py", - "left_ranges": [ - [ - 1330, - 1338 - ], - [ - 3763, - 3789 - ], - [ - 4223, - 4234 - ], - [ - 4628, - 4641 - ] - ], - "right_ranges": [ - [ - 1346, - 1354 - ], - [ - 3785, - 3811 - ], - [ - 4245, - 4257 - ], - [ - 4651, - 4664 - ] - ] - }, - { - "filename": "tests/kernels/test_kernels.py", - "left_ranges": [ - [ - 19, - 54 - ], - [ - 228, - 281 - ], - [ - 294, - 317 - ], - [ - 605, - 995 - ] - ], - "right_ranges": [ - [ - 21, - 28 - ], - [ - 33, - 39 - ], - [ - 90, - 103 - ], - [ - 217, - 301 - ], - [ - 314, - 337 - ] - ] - } - ] - } - ] - }, { "cluster_id": "cluster-43240-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", @@ -1972,6 +1614,81 @@ } ] }, + { + "cluster_id": "cluster-43656-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 43656, + "issue_numbers": [], + "pr_numbers": [ + 43656, + 43842 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43656, + "right_pr_number": 43842, + "code_similarity": 0.405, + "size_similarity": 0.5, + "file_overlap": 1.0, + "area_overlap": 0.0, + "patch_similarity": 0.036, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 30, + 36 + ], + [ + 315, + 323 + ], + [ + 665, + 671 + ], + [ + 931, + 937 + ], + [ + 1843, + 1849 + ], + [ + 1868, + 1874 + ] + ], + "right_ranges": [ + [ + 54, + 61 + ], + [ + 587, + 593 + ] + ] + } + ] + } + ] + }, { "cluster_id": "cluster-44018-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", @@ -2277,6 +1994,15 @@ "target_issue_number": 43240, "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." }, + { + "cluster_id": "cluster-43656-2", + "canonical_pr_number": 43656, + "duplicate_pr_numbers": [ + 43842 + ], + "target_issue_number": null, + "reason": "PRs in cluster-43656-2 are treated as duplicates because they share soft_similarity evidence." + }, { "cluster_id": "cluster-43698-3", "canonical_pr_number": 43779, @@ -2321,16 +2047,6 @@ "target_issue_number": 45081, "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." }, - { - "cluster_id": "cluster-45363-3", - "canonical_pr_number": 46339, - "duplicate_pr_numbers": [ - 45363, - 45666 - ], - "target_issue_number": null, - "reason": "PRs in cluster-45363-3 are treated as duplicates because they share soft_similarity evidence." - }, { "cluster_id": "cluster-45561-3", "canonical_pr_number": 45645, diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 5f44c7ae5349120930395102a0e76cbf562830d6..c5efecb55bc693660dd99c23b4354d6e94001670 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": "20260604T120021Z", - "generated_at": "2026-06-04T12:10:12Z", + "snapshot_id": "20260604T180024Z", + "generated_at": "2026-06-04T18:11:31Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -16,7 +16,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:24643", "issue:30277", @@ -106,15 +106,15 @@ "issue:45874|issue:46294", "issue:39290|issue:45081", "issue:45823|issue:45874", - "issue:34928|issue:41720", "issue:43425|issue:44292", + "issue:34928|issue:41720", "issue:43828|issue:45161", "issue:43541|issue:43716", "issue:43828|issue:44292", "issue:43716|issue:45685", "issue:44918|issue:45161", - "issue:45507|issue:45685", "issue:44292|issue:44918", + "issue:45507|issue:45685", "issue:43716|issue:44292", "issue:43039|issue:44918", "issue:41762|issue:43638", @@ -133,383 +133,422 @@ "issue:41762|issue:43596", "issue:39290|issue:41720" ], - "prepared_review_unit_hash": "ea5dff9f23e7d88ab3e55c4a8d243ee135d8389ac6c6e6d996397b10cfa2ac83", + "prepared_review_unit_hash": "53bd3fd6d0990d3a29343d0e99ae884846eb64591f640e9f1aa3fbb4f15259dc", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 64, - "serialized_chars": 46881, - "estimated_input_tokens": 11721, - "estimated_eval_tokens": 23698 + "serialized_chars": 46877, + "estimated_input_tokens": 11720, + "estimated_eval_tokens": 23696 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are largely unrelated bug reports and docs problems sharing only broad keywords like DeepSpeed, dtype, or model family names. I found no strong duplicate groups; the soft pairs should mostly be rejected.", - "confidence": 0.95, - "canonical_issue_reason": "Issue 34928 is the most substantial concrete bug report here, with the most discussion and a clear reproducible FSDP/activation-checkpointing failure, so it is the best representative issue even though the cluster is not truly duplicate-heavy.", + "summary": "Most pairs are unrelated. The only clear duplicate cluster is the DeepSpeed ZeRO-3 zero-sized tensor IndexError around Bert/Gemma model loading; the rest are distinct model-, backend-, or docs-specific issues.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43638 is the clearest representative of the ZeRO-3 zero-sized tensor IndexError cluster: it states the exact exception, the DeepSpeed ZeRO-3 context, and a concrete trigger (non-pretrained Bert).", "canonical_pr_reason": null, - "best_issue_reason": "Issue 34928 is the strongest standalone issue in the set because it is well-scoped, reproducible, and heavily discussed.", + "best_issue_reason": "issue:43638 is the most general and explicit reproduction of the shared ZeRO-3 loading bug, so it is the best canonical issue for that cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different failures: missing config field vs streaming chat/completions error." + "reason": "Different Qwen3.5 bugs: missing config attribute vs streaming chat/completions failure." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different hardware/code paths: AMD GPU failure vs MPS histogram op issue." + "reason": "Different hardware/backend failures: AMD GPU issue vs MPS histogram op limitation." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different models and symptoms; image dtype mismatch is unrelated to AMD GPU execution." + "reason": "Different models and failure modes; not the same bug." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both mention runtime errors, but the MoE autocast dtype issue is not the AMD GPU failure." + "reason": "Different symptoms: autocast dtype mismatch vs AMD GPU failure." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different underlying problem: AMD GPU support vs invalid method call in GraniteMoEHybrid." + "reason": "Different problems: AMD GPU runtime issue vs invalid method call in GraniteMoEHybrid." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different model/runtime contexts; NVFP4 run failure is not the AMD GPU bug." + "reason": "Different models and error paths; not a duplicate." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch and ONNX export are unrelated code paths." + "reason": "Distributed collective mismatch in training is unrelated to ONNX export failure." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "ModernBERT ONNX export and Qwen3Moe FP8 failure are distinct model/export issues." + "reason": "Export-to-ONNX bug and FineGrainedFP8 config failure are different code paths." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ONNX export error is unrelated to Qwen3 auto device-map cuda assertion." + "reason": "ONNX export issue vs auto device-mapping cuda assert; unrelated." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Table QA crash and outdated RAG examples are different documentation/runtime issues." + "reason": "Table QA crash and outdated RAG examples are unrelated." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs cleanups for v5, but they refer to different removed docs content." + "reason": "Both are v5 docs cleanups, but they target different removed docs content." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export failure does not match the GenerationMixin import regression." + "reason": "ONNX export bug and GenerationMixin import error are separate issues." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "TRL input-embedding unpacking and resuming local-trust-remote-code training are different problems." + "reason": "TRL SFT embedding unpacking and training resume with remote code are different failures." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Mask-shape IndexError and Gemma4 device mismatch are different failures." + "reason": "Generic mask-shape mismatch vs Gemma4 device-mapping mismatch; different bugs." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both are IndexErrors, but one is ZeRO-3/BertModel init and the other is a mask shape mismatch." + "reason": "Different exception types and different model/runtime contexts." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different zero3/BertModel init bug vs tensor-mask indexing bug." + "reason": "Zero-3 init IndexError is unrelated to mask/tensor shape mismatch." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask shape mismatch is unrelated to Qwen3.5 video token/feature mismatch." + "reason": "Mask-shape mismatch and Qwen3.5 video token mismatch are different problems." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Generic IndexError is not the same as GPT-OSS TP-only failure." + "reason": "Different failures: mask/indexing bug vs GPT-OSS TP issue." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video StopIteration failure and resume-training issue are unrelated." + "reason": "Video apply_chat_template failure is unrelated to resume/trust_remote_code training issue." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "ZeRO-3 BertModel init bug is unrelated to resuming from local model with remote code." + "reason": "DeepSpeed ZeRO-3 Bert loading bug is unrelated to training resume with remote code." }, { "left": "issue:43572", "right": "issue:46165", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is unrelated to training resume logic." + "reason": "StableLmConfig pad_token_idx regression is unrelated." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice_preset and ModernBERT ONNX export are unrelated." + "reason": "BarkProcessor voice_preset bug is unrelated to ModernBERT ONNX export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both involve distributed training, but the concrete errors and code paths differ." + "reason": "Both involve training/runtime issues, but the errors and code paths are different." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 index-out-of-bounds loading bug on Bert models; only the model wording differs." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross_entropy call and GPT-OSS tensor-parallel failure are unrelated." + "reason": "Liger Kernel cross_entropy call and GPT-OSS TP failure are different code paths." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Tokenizer detokenization bug and LoRA save/reload inconsistency are different issues." + "reason": "Incorrect detokenization and LoRA merge/reload mismatch are different bugs." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "NVFP4 runtime failure is unrelated to LoRA merge/save behavior." + "reason": "NVFP4 runtime failure is unrelated to LoRA save/reload inconsistency." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Different failures: recomputed tensor size mismatch vs mask indexing shape error." + "reason": "FSDP checkpointing size mismatch is unrelated to the mask indexing error." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Different DeepSpeed errors and different model setups." + "reason": "DeepSpeed weight-dimension error and collective mismatch are different failure modes." }, { "left": "issue:45859", "right": "issue:45902", "accept": false, - "reason": "Both touch the same class, but one is CPU offload compatibility and the other is uninitialized buffer materialization." + "reason": "Same model family, but CPU offload incompatibility and meta-to-empty uninitialized memory are different concrete bugs." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "Zero-3 Bert loading bug is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "Checkpointing/FSDP tensor size mismatch is unrelated to CustomTrainer compute_loss signature change." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "DeepSpeed ZeRO-3 Bert loading bug is unrelated to GPT-OSS TP failure." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Dtype mismatch in Mistral-3 preprocessing is unrelated to GraniteMoEHybrid invalid method calls." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Different Gemma4 input-embedding issue vs Qwen3.5 video chat-template mismatch." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3 config attr regression and Mistral tokenizer regex patch crash are unrelated." + "reason": "Gemma3/vLLM attribute error and Mistral regex patch crash are unrelated." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Gemma4 device-map mismatch and reverse embedding lookup are different code paths." + "reason": "Different Gemma4 bugs: device mismatch in forward vs expensive reverse embedding lookup." }, { - "left": "issue:34928", - "right": "issue:41720", + "left": "issue:43425", + "right": "issue:44292", "accept": false, - "reason": "FSDP activation checkpointing mismatch is unrelated to Qwen3 auto device-map assert." + "reason": "Torch/version incompatibility is unrelated to Qwen-3 NVFP4 runtime failure." }, { - "left": "issue:43425", - "right": "issue:44292", + "left": "issue:34928", + "right": "issue:41720", "accept": false, - "reason": "Torch/transformers compatibility and Qwen NVFP4 runtime failure are different problems." + "reason": "FSDP activation checkpointing mismatch and Qwen3 auto device-map assert are unrelated." }, { "left": "issue:43828", "right": "issue:45161", "accept": false, - "reason": "Phi MoE autocast dtype mismatch is unrelated to GPT-OSS tensor-parallel failure." + "reason": "Phi-tiny-MoE autocast dtype mismatch is unrelated to GPT-OSS TP failure." }, { "left": "issue:43541", "right": "issue:43716", "accept": false, - "reason": "Grouped_mm tracing error in Mixtral is unrelated to Mistral-3 image dtype mismatch." + "reason": "Mixtral grouped_mm tracing error and Mistral-3 preprocessor dtype mismatch are different bugs." }, { "left": "issue:43828", "right": "issue:44292", "accept": false, - "reason": "Different model families and different failure modes." + "reason": "Different models and different dtype/runtime failures." }, { "left": "issue:43716", "right": "issue:45685", "accept": false, - "reason": "Image preprocessor dtype mismatch is unrelated to the MPS histogram Int error." + "reason": "Image-preprocessor dtype mismatch is unrelated to MPS histogram op failure." }, { "left": "issue:44918", "right": "issue:45161", "accept": false, - "reason": "TRL SFT embedding unpacking and GPT-OSS TP failure are not the same bug." + "reason": "TRL embedding unpacking issue is unrelated to GPT-OSS tensor-parallel failure." }, { - "left": "issue:45507", - "right": "issue:45685", + "left": "issue:44292", + "right": "issue:44918", "accept": false, - "reason": "Invalid method call in GraniteMoEHybrid is unrelated to the MPS histogram op issue." + "reason": "Different models and code paths; not the same bug." }, { - "left": "issue:44292", - "right": "issue:44918", + "left": "issue:45507", + "right": "issue:45685", "accept": false, - "reason": "NVFP4 runtime failure and TRL SFT input-embedding issue are different." + "reason": "GraniteMoEHybrid invalid method call and MPS histogram op error are unrelated." }, { "left": "issue:43716", "right": "issue:44292", "accept": false, - "reason": "Different model-specific runtime errors with no shared code path." + "reason": "Mistral-3 dtype mismatch and Qwen-3 NVFP4 failure are different issues." }, { "left": "issue:43039", "right": "issue:44918", "accept": false, - "reason": "Liger Kernel cross_entropy issue is unrelated to Qwen3.5 embedding unpacking." + "reason": "Liger Kernel cross_entropy call is unrelated to TRL SFT embedding unpacking." }, { "left": "issue:41762", "right": "issue:43638", - "accept": false, - "reason": "Both use DeepSpeed ZeRO-3, but the Gemma3 load error and Bert init error are distinct." + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 index-out-of-bounds loading error; only the affected model differs." }, { "left": "issue:43827", "right": "issue:44512", "accept": false, - "reason": "Different v5 docs removals: pipeline tasks vs run command." + "reason": "Both are docs cleanup, but they refer to different removed features and should stay separate." }, { "left": "issue:43716", "right": "issue:44560", "accept": false, - "reason": "Mistral-3 image dtype mismatch is unrelated to video StopIteration." + "reason": "Mistral-3 dtype mismatch is unrelated to Qwen3.5 video template token mismatch." }, { "left": "issue:43828", "right": "issue:45685", "accept": false, - "reason": "Different model/device problems; autocast dtype mismatch is not the MPS histogram bug." + "reason": "Autocast dtype mismatch is unrelated to MPS integer histogram support." }, { "left": "issue:44918", "right": "issue:45685", "accept": false, - "reason": "TRL embedding unpacking and MPS histogram Int error are unrelated." + "reason": "TRL trainer issue is unrelated to MPS backend limitation." }, { "left": "issue:43039", "right": "issue:43828", "accept": false, - "reason": "Different models and code paths; both mention runtime math but not the same bug." + "reason": "Liger Kernel cross_entropy usage and Phi-tiny-MoE autocast mismatch are different bugs." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch and custom trainer signature mismatch are unrelated." + "reason": "DeepSpeed collective mismatch and compute_loss signature change are unrelated." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "FSDP/activation checkpointing recompute bug is unrelated to GenerationMixin import regression." + "reason": "FSDP checkpointing mismatch is unrelated to GenerationMixin import failure." }, { "left": "issue:45237", "right": "issue:46015", "accept": false, - "reason": "AMD GPU runtime failure and outdated RAG docs are unrelated." + "reason": "AMD GPU runtime issue is unrelated to outdated RAG docs." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger Kernel cross-entropy call and mask shape IndexError are different issues." + "reason": "Cross-entropy kernel usage and mask-shape mismatch are unrelated." }, { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Corrupted Ovis2 checkpoint and unloaded Sam3Model text encoder weights are different." + "reason": "Corrupted checkpoint and Sam3Model weight-loading regression are different issues." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "Generic torch/transformers compatibility and USE_DISTRIBUTED=0 breakage are not the same bug." + "reason": "Torch/transformers compatibility complaint and USE_DISTRIBUTED=0 regression are unrelated." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "PaddleOCR-VL/vLLM load failure is unrelated to distributed-build support regression." + "reason": "vLLM/PaddleOCR-VL load failure and distributed build regression are different bugs." }, { "left": "issue:41762", "right": "issue:43596", "accept": false, - "reason": "Both are ZeRO-3-related, but the Gemma3 load issue and Bert zero3 init issue are distinct." + "reason": "Closely related DeepSpeed ZeRO-3 IndexError reports, but the model/context differs enough to keep them separate without more evidence." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3 config attribute error is unrelated to Qwen3 auto device mapping assertion." + "reason": "Gemma3/vLLM attribute issue is unrelated to Qwen3 auto device-map cuda assert." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the soft-edge calls are generally conservative. The one clear duplicate cluster is the DeepSpeed ZeRO-3 zero-sized tensor IndexError around Bert/Gemma loading, while the rest are separated by clearly different model, backend, or docs contexts." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -524,7 +563,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:36010", "issue:39290", @@ -577,8 +616,8 @@ ], "soft_pairs": [ "issue:44589|issue:46304", - "issue:43901|issue:44512", "issue:45823|issue:46294", + "issue:43901|issue:44512", "issue:45237|issue:45901", "issue:44841|issue:45823", "issue:43039|issue:44560", @@ -610,28 +649,28 @@ "issue:41093|issue:41762", "issue:43296|issue:43366", "issue:43572|issue:45070", - "issue:44387|issue:45005", "issue:44805|issue:44918", + "issue:44387|issue:45005", "issue:43541|issue:45161", "issue:43828|issue:45464", "issue:44291|issue:45070", "issue:43366|issue:45464", "issue:43531|issue:45070" ], - "prepared_review_unit_hash": "95eb0a4e4965120aa65ed7c9f5887b9fe0b48d24e0d9be5b0877bf1651e22717", + "prepared_review_unit_hash": "0d10c028f016d239a66ca3499e4dfe0212ddf46e73cfa1385e13e6eb88278dbc", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39794, + "serialized_chars": 39796, "estimated_input_tokens": 9949, "estimated_eval_tokens": 20154 }, @@ -640,267 +679,237 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated Transformers issues across tokenization, model-loading, backend compatibility, and docs. A few pairs look like true duplicates with the same concrete runtime/error signature, but most are only superficially similar.", - "confidence": 0.78, - "canonical_issue_reason": "issue:45920 is the broadest and most general report of the v5 AutoTokenizer wrong-token-ID regression, covering multiple model families rather than a single narrow case.", + "summary": "Most pairs are clearly unrelated, with the only strong duplicate signal in the visible v5 AutoTokenizer wrong-token-ID regression cluster. The rest are mostly superficially similar error strings or broad subsystem overlaps without enough evidence of the same underlying bug.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45920 is the best canonical issue because it is the broadest visible report of the v5 AutoTokenizer wrong-token-ID regression and spans multiple model families; issue:45812 is the narrower Granite-specific instance of the same symptom.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45920 is the strongest representative issue overall: it is open, clearly scoped to a concrete regression, and its title captures the widest affected surface among the likely-duplicate tokenizer reports.", + "best_issue_reason": "issue:45920 is the best global issue anchor in this packet because it describes the broadest visible tokenizer regression and names several affected model families, making it the most representative report here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Different failure classes: missing Float8 storage vs USE_DISTRIBUTED=0 support regression." + "reason": "Different incompatibility classes: missing Float8 storage support versus a USE_DISTRIBUTED=0 regression." }, { - "left": "issue:43901", - "right": "issue:44512", + "left": "issue:45823", + "right": "issue:46294", "accept": false, - "reason": "Both are docs issues, but they target different outdated doc content and are not the same bug." + "reason": "Different models and failure surfaces: Gemma4 device-map mismatch versus Qwen3.5 video/chat-template token-feature mismatch." }, { - "left": "issue:45823", - "right": "issue:46294", + "left": "issue:43901", + "right": "issue:44512", "accept": false, - "reason": "Different models and different failure modes: device mismatch vs chat-template video token mismatch." + "reason": "Both are docs-related, but they concern unrelated documentation problems." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "Unrelated subsystems: AMD GPU support vs table-question-answering crash." + "reason": "Unrelated failures in different areas: GPT-OSS AMD GPU support versus table-question-answering crashes." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Different model-specific loading/processor issues; no clear shared code-path bug." + "reason": "Different models and bugs: Voxtral processor failure versus Gemma4 device mismatch." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross-entropy call and Qwen3-VL StopIteration are unrelated." + "reason": "Different code paths and symptoms: Liger cross_entropy dispatch versus Qwen3-vl video StopIteration." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Detokenization bug vs outdated RAG examples; different problem types." + "reason": "Tokenizer detokenization bug versus outdated RAG examples; not the same issue." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Same concrete regression class: AutoTokenizer returns wrong token IDs in v5 across affected model families." + "reason": "Same underlying v5 AutoTokenizer regression producing wrong token IDs across multiple model families." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint-loading bug is unrelated to documentation updates." + "reason": "Model-loading/architecture support issue versus outdated examples; unrelated root causes." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Both mention recent-version breakage, but they affect different paths and are not the same bug." + "reason": "Broad remote_code breakage is not the same as the distributed-build regression." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Corrupted checkpoint vs NemotronH loader failure are different issues." + "reason": "Checkpoint corruption versus implementation loading support; different causes." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes init argument incompatibility is unrelated to USE_DISTRIBUTED=0 support." + "reason": "bitsandbytes constructor mismatch is unrelated to the USE_DISTRIBUTED=0 regression." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Model loading/runtime error vs outdated docs; no shared underlying defect." + "reason": "Model-loading type error versus a documentation issue; no shared bug." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both involve sliding-window terminology, but the model/config contexts differ and the concrete bugs are not the same." + "reason": "Both mention sliding_window, but they affect different models and likely different code paths." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Similar IndexError wording, but one is a mask-shape mismatch and the other is a zero-sized tensor under ZeRO-3; not enough to merge." + "reason": "Similar IndexError wording, but the shapes and contexts differ enough to treat as separate bugs." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Different model-loading failures for different checkpoints." + "reason": "Different load failures: unit-test model loading problem versus checkpoint corruption." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper load failure and table-question-answering crash are unrelated." - }, - { - "left": "issue:43425", - "right": "issue:45070", - "accept": false, - "reason": "General torch incompatibility and pydantic PretrainedConfig regression are different breakages." - }, - { - "left": "issue:43596", - "right": "issue:45800", - "accept": false, - "reason": "ZeRO-3 Bert IndexError and torch 2.4.1/transformers 5.8.0 incompatibility are separate issues." + "reason": "Whisper model execution failure is unrelated to table-question-answering crashes." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "Detokenization issue and table-question-answering crash do not share a concrete code path." + "reason": "Tokenizer detokenization bug versus table QA crash; unrelated." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Different components and symptoms; no duplicate evidence." + "reason": "Whisper runtime failure and pydantic config regression are different bugs." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Qwen3.5 chat crash and streaming inference API failure are not the same issue." - }, - { - "left": "issue:44292", - "right": "issue:45901", - "accept": false, - "reason": "Model loading with init_empty_weights is unrelated to table-question-answering crashes." - }, - { - "left": "issue:39290", - "right": "issue:45823", - "accept": false, - "reason": "Gemma3/vLLM sliding-window config issue is unrelated to Gemma4 PLE device mismatch." + "reason": "Same model family, but different entrypoints and symptoms: chat crash versus streaming chat/completions failure." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Streaming chat/completions failure and MPS histogram int error are unrelated." + "reason": "Different subsystems and failure modes: streaming inference versus MPS histogram support." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Tied lm_head weights in Mistral3 is unrelated to a corrupted Ovis2 checkpoint." + "reason": "Weight-tying bug in Mistral3 is unrelated to a corrupted checkpoint report." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Dynamo grouped_mm tracing failure and MPS histogram op error are different backend problems." + "reason": "Different backend failures: torch dynamo/grouped_mm tracing versus MPS histogram op support." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Both are dtype/backend-adjacent, but the model, trigger, and code path differ." - }, - { - "left": "issue:44841", - "right": "issue:45084", - "accept": false, - "reason": "Processor failure for Voxtral is unrelated to template-node compilation." + "reason": "Both involve MoE or mixed precision, but the concrete failures and models differ." }, { - "left": "issue:36010", - "right": "issue:39290", + "left": "issue:44805", + "right": "issue:44918", "accept": false, - "reason": "ImportError for GenerationMixin and Gemma3 sliding_window_pattern are unrelated." + "reason": "Both are runtime errors, but one is a mask-shape IndexError and the other is an embedding-unpacking failure." }, { - "left": "issue:45800", - "right": "issue:45823", + "left": "issue:44387", + "right": "issue:45005", "accept": false, - "reason": "Torch version incompatibility and Gemma4 device mismatch are different bugs." + "reason": "Quantization OOM regression versus tied-weights issues; different code paths." }, { - "left": "issue:41093", - "right": "issue:41762", + "left": "issue:43541", + "right": "issue:45161", "accept": false, - "reason": "Both are IndexErrors, but they involve different models and different failing tensors." + "reason": "Different models and problems: Mixtral tracing error versus GPT-OSS tensor-parallel-only failure." }, { - "left": "issue:43296", - "right": "issue:43366", + "left": "issue:43828", + "right": "issue:45464", "accept": false, - "reason": "PaddleOCR-VL load failure and GGUF gpt-oss support are different issues." + "reason": "Dtype mismatch under autocast is unrelated to the Qwen3.5 streaming API failure." }, { - "left": "issue:43572", + "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is unrelated to the pydantic PretrainedConfig regression." + "reason": "Unexpected _is_hf_initialized argument and a PretrainedConfig pydantic-field regression are not the same bug." }, { - "left": "issue:44387", - "right": "issue:45005", + "left": "issue:43366", + "right": "issue:45464", "accept": false, - "reason": "Int4 memory regression and tied-weights translation-model issue are different concrete bugs." + "reason": "GGUF gpt-oss support request is unrelated to the Qwen3.5 streaming API bug." }, { - "left": "issue:44805", - "right": "issue:44918", + "left": "issue:43531", + "right": "issue:45070", "accept": false, - "reason": "Both are runtime errors, but one is a mask/tensor shape mismatch and the other is an embedding unpacking failure." + "reason": "Sliding-window model issue versus pydantic PretrainedConfig regression; different root causes." }, { - "left": "issue:43541", - "right": "issue:45161", + "left": "issue:36010", + "right": "issue:39290", "accept": false, - "reason": "Dynamo tracing grouped_mm failure is not the same as GPT-OSS MoE tensor-parallel failure." + "reason": "GenerationMixin import failure and Gemma3 sliding_window_pattern are unrelated." }, { - "left": "issue:43828", - "right": "issue:45464", + "left": "issue:45800", + "right": "issue:45823", "accept": false, - "reason": "Autocast dtype mismatch and streaming chat/completions failure are unrelated." + "reason": "Torch/transformers compatibility issue versus Gemma4 forward device mismatch." }, { - "left": "issue:44291", - "right": "issue:45070", + "left": "issue:41093", + "right": "issue:41762", "accept": false, - "reason": "_is_hf_initialized init-empty-weights TypeError is different from the pydantic model field regression." + "reason": "Both are Zero-3-related IndexErrors, but the reported shapes and model contexts do not indicate the same bug." }, { - "left": "issue:43366", - "right": "issue:45464", + "left": "issue:43296", + "right": "issue:43366", "accept": false, - "reason": "GGUF architecture support and Qwen3.5 streaming failure are unrelated." + "reason": "PaddleOCR-VL load failure and GGUF gpt-oss support request are different model-specific issues." }, { - "left": "issue:43531", + "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window issue is not the same as the PretrainedConfig pydantic regression." + "reason": "Missing pad_token_idx in StableLmConfig and a pydantic PretrainedConfig regression are different config bugs." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet and stays conservative. The chosen canonical issue (45920) is plausibly the broadest v5 AutoTokenizer wrong-token-ID report, and the only accepted duplicate pair (45812, 45920) is supported by their shared regression class. The rejected soft pairs are phrased cautiously and do not overclaim duplication beyond the evidence shown." + "accept": false, + "feedback": "The pairwise judgments are mostly conservative and grounded in the titles, but the top-level issue reasoning overstates the packet: it refers to an \"accepted duplicate family,\" which is not established in the provided data. I\u2019d want the global-anchor/canonical-issue claims phrased more cautiously and only in terms visible in the packet (broadest v5 tokenizer regression, multiple model families, etc.)." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -916,7 +925,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:36331", "issue:38175", @@ -1001,238 +1010,238 @@ "issue:43866|issue:44403", "issue:38175|issue:43054" ], - "prepared_review_unit_hash": "7926aa37cc33a8172bab9da902688c8bef98e712ae63ffab4a3c2070ecbf74a1", + "prepared_review_unit_hash": "db9132f1972d069a9d611394def93a6cf6aaaae2dcf439f7658a73da63139c23", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 33, - "serialized_chars": 36893, - "estimated_input_tokens": 9224, - "estimated_eval_tokens": 18704 + "serialized_chars": 36897, + "estimated_input_tokens": 9225, + "estimated_eval_tokens": 18706 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are unrelated symptomatically despite overlapping subsystems. The only clear duplicate cluster is the Transformers 5.x non-persistent-buffer regression, centered on the more specific from_pretrained corruption report.", - "confidence": 0.86, - "canonical_issue_reason": "issue:43950 is the most specific and complete report of the non-persistent-buffer regression, directly naming the corrupting from_pretrained path and the transformers 5.x scope.", + "summary": "Mostly a heterogeneous set of unrelated closed issues. The only clear duplicate thread is the Transformers 5.x non-persistent buffer corruption regression around from_pretrained / buffer materialization.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43950 is the clearest root report of the non-persistent buffer corruption regression, with the most direct repro and wording; issue:43644 is the earlier broader report and issue:44534 is a rephrased follow-up.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 is the best representative issue: it is precise, regression-focused, and describes the concrete failure mode more clearly than the broader duplicate reports.", + "best_issue_reason": "issue:43950 is the best representative duplicate target because it describes the concrete failure mode precisely and anchors the confirmed duplicate chain in this cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Different models and failure modes: image preprocessor/model dtype mismatch vs video feature/token count mismatch." + "reason": "Different models and different failure modes (dtype mismatch vs video chat template token/feature mismatch)." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "One is about recent versions breaking remote_code models; the other is a specific torch 2.4.1 vs transformers 5.8.0 incompatibility." + "reason": "One is a remote_code regression, the other is a torch/transformers version compatibility problem; not the same bug." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Different concrete errors: grouped_mm tracing failure vs missing Float8 storage class." + "reason": "Unrelated runtime errors in different code paths (grouped_mm tracing vs missing Float8 storage)." }, { "left": "issue:44479", "right": "issue:46294", "accept": false, - "reason": "Both are video-related, but one is a v5.3.0 regression across Qwen VL models and the other is a chat-template feature/token mismatch on Qwen3.5-9B." + "reason": "Both mention Qwen video, but one is a broad v5.3.0 regression and the other is a specific apply_chat_template mismatch; not clearly the same bug." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are IndexErrors, but they involve different shapes, loaders, and likely different root causes." + "reason": "Both are IndexErrors under DeepSpeed, but the concrete shapes and trigger differ." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Same model family, but the reports describe different symptoms and code paths; not enough evidence of the same underlying bug." + "reason": "Same model family, but one reports degraded embeddings and the other nonsensical pipeline output; too vague to be the same concrete bug." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Different problems: config unwrapping for multimodal AutoModelForCausalLM vs uninitialized rotary embedding memory after materialization." + "reason": "Different issues: composite config unwrapping vs uninitialized rotary embedding memory after materialization." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "Unrelated issues: missing TensorFlow backend handling vs version incompatibility between torch and transformers." + "reason": "TensorBoard/TensorFlow backend handling is unrelated to the torch 2.4.1/transformers 5.8.0 incompatibility report." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "Different stack traces and subsystems: FSDP weight shape error vs TRL SFT embedding unpacking failure." + "reason": "FSDP 2-D weight error and Qwen3.5 SFT embedding unpacking are different failures." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Both are model mismatch-ish, but one is a multimodal dtype issue and the other is a trainer/input embedding failure." + "reason": "Different models and subsystems; image preprocessor dtype mismatch is unrelated to TRL SFT embedding unpacking." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "Different models and symptoms: GPT-OSS perplexity anomaly vs GLM5-specific issue." + "reason": "Completely different models and symptoms (perplexity vs GLM5-related report)." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "Related custom-code loading area, but one is a trust_remote_code cache collision for same leaf names and the other is wrong module loading after save_pretrained; insufficient evidence of one bug." + "reason": "Both involve custom-code loading, but one is a leaf-directory cache collision and the other is wrong module loading after save_pretrained; not clearly the same bug." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom-model initialization in notebooks is unrelated to flash-attn fallback selection logic." + "reason": "Custom model init in Jupyter and flash-attn2 fallback gating are unrelated." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Tied-weight bug in Mistral3 is not the same as the broader translation-model tied-weights regression in v5." + "reason": "Both mention tied weights, but they concern different model families and different manifestations." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Qwen3 MoE FP8 failure is unrelated to flash-attn fallback gating." + "reason": "FineGrainedFP8 failure and flash-attn2 fallback checks are unrelated." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache / linear_attention crash is a different code path from tied-weight loading issues." + "reason": "PagedAttentionCache group-type crash is unrelated to translation-model tied-weight issues." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization offloading behavior is unrelated to tied-weight initialization regressions." + "reason": "Quantization/offloading behavior does not match the tied-weights regression." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint-loading failure is not the same bug as translation-model tied weights." + "reason": "NemotronH checkpoint-loading failure is unrelated to translation-model weight tying." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A DSA indexer algorithm discussion is unrelated to a docs-only v5 command removal issue." + "reason": "A DSA indexer bug and a docs typo are unrelated." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different models and failures: Qwen3.5 embedding unpacking vs GraniteMoEHybrid invalid method call." + "reason": "Different model families and different error mechanisms." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "Typo in cache strategies is unrelated to documentation still mentioning a removed run command." + "reason": "A cache-strategy typo and a removed CLI docs mention are unrelated." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Different models and defects: detokenization regression vs NVFP4 runtime error." + "reason": "Detokenization regression and Qwen NVFP4 runtime error are different bugs." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "No shared underlying code path; one is Qwen-3 NVFP4 runtime failure, the other is GraniteMoEHybrid method dispatch." + "reason": "Different model-specific runtime failures." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Both report the same Transformers v5 regression where non-persistent buffers are filled/corrupted with junk during loading." + "reason": "Both report the same non-persistent-buffer corruption regression in Transformers 5.x; 44534 is a reworded follow-up to 43950." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Docs example/model-processor mismatch and quantization failure is broader and different from the text-embedding quality complaint." + "reason": "SigLIP2 docs example problems and SigLIP2 embedding quality regression are related only at a broad model level, not the same concrete bug." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Same bug: Transformers 5.x corrupts non-persistent buffers on load; 43950 is the more precise report." + "reason": "Same regression: non-persistent buffers get corrupted/junk-filled during loading/materialization in Transformers 5.x." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different failure modes: autocast dtype mismatch vs RLHF NaN explosion from 3D position_ids and SDPA fallback." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE and Qwen3.5 RLHF NaN explosion are different issues." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Different APIs and traces: custom trainer compute_loss signature change vs DeepSpeed ZeRO-3 model loading index error." + "reason": "Trainer API keyword mismatch and DeepSpeed ZeRO-3 load error are unrelated." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Bitsandbytes __new__ argument incompatibility is unrelated to missing Float8 storage objects." + "reason": "Different quantization-related incompatibilities with different symptoms." }, { "left": "issue:43950", "right": "issue:45902", "accept": false, - "reason": "One is a load-time non-persistent-buffer corruption regression; the other is uninitialized memory in Qwen3_5MoE rotary embeddings after materialization." + "reason": "Both mention buffer/materialization issues, but 45902 is a specific uninitialized rotary embedding case and not the same regression." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "Model quality/output regression is not the same as a version-dependent tokenization change." + "reason": "Model embedding quality and tokenization changes across versions are different problems." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Checkpoint corruption is unrelated to noisy transformer-loading logs." + "reason": "Checkpoint corruption and noisy loading are unrelated." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both concern siglip2, but one is zero probabilities and the other is poor embeddings; not enough evidence of the same concrete bug." + "reason": "Both are SigLIP2 complaints, but zero probabilities and poor embeddings are not specific enough to treat as the same bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative on the soft-edge pairs. The Transformers 5.x non-persistent-buffer regression cluster is correctly identified, with issue:43950 as the most specific canonical report and issue:44534/43644 as supporting duplicates. I did not see any overstatement of evidence in the remaining rejections." + "feedback": "The summary is grounded in the packet and the soft-edge verdicts are conservative overall. The accepted duplicate thread around 43644/43950/44534 is well supported by the issue titles, and the other pair rejections avoid over-claiming sameness. Minor caution: a couple of phrases like \"most direct repro\" go a bit beyond what the packet explicitly shows, but not enough to invalidate the cluster assessment." }, "evaluator_used": true, "retried": false, @@ -1251,7 +1260,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:41553", "issue:41762", @@ -1303,8 +1312,8 @@ ], "soft_pairs": [ "issue:41553|issue:43883", - "issue:45663|issue:45902", "issue:44910|issue:45859", + "issue:45663|issue:45902", "issue:42947|issue:44387", "issue:43054|issue:43493", "issue:41553|issue:43881", @@ -1314,13 +1323,13 @@ "issue:43643|issue:45020", "issue:42947|issue:43856", "issue:41762|issue:42915", - "issue:43782|issue:44960", "issue:45685|issue:45823", - "issue:44991|issue:45399", + "issue:43782|issue:44960", "issue:43824|issue:43872", - "issue:42491|issue:44368", + "issue:44991|issue:45399", "issue:43881|issue:43883", "issue:45020|issue:45823", + "issue:42491|issue:44368", "issue:44661|issue:45020", "issue:43828|issue:45823", "issue:45081|issue:45399", @@ -1334,226 +1343,226 @@ "issue:43792|issue:44220", "issue:43257|issue:45305" ], - "prepared_review_unit_hash": "597ad2ea6bcf9666910d19f55a9584e91c47b3cc90a3d116156cd54ca5fcd282", + "prepared_review_unit_hash": "173a96ffbeddc66583a1e2fd4f903d1559bf57ceb1133086007f453a9906736e", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36794, - "estimated_input_tokens": 9199, - "estimated_eval_tokens": 18654 + "serialized_chars": 36798, + "estimated_input_tokens": 9200, + "estimated_eval_tokens": 18656 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of Transformers issues with only a few likely duplicate families: remote_code loading regressions, v5 pipeline-doc cleanup, and the `_is_hf_initialized` / `init_empty_weights` TypeError regression. The rest are model-specific and should stay separate.", - "confidence": 0.82, - "canonical_issue_reason": "Issue 45020 is the broadest umbrella among the likely duplicate families: it captures the general recent `remote_code` loading regression that 43643 appears to exemplify. It is the best single issue to anchor the cluster, even though the cluster as a whole is mixed.", + "summary": "Mostly a grab bag of distinct v5 regression reports around model loading, tokenizers, MoE, quantization, and docs. Only two soft pairs look like true duplicates: the stale v5 docs/pipeline references and the `_is_hf_initialized` init_empty_weights/Int8Params TypeError.", + "confidence": 0.57, + "canonical_issue_reason": "issue:45020 is the broadest umbrella report for recent Transformers v5 loading/remote-code regressions, making it the best representative of the cluster\u2019s shared theme.", "canonical_pr_reason": null, - "best_issue_reason": "45020 is the best representative issue because it is the most general, user-facing regression among the soft matches and is broad enough to subsume at least one specific loading-failure report. It is a better anchor than the more narrow, symptom-level reports.", + "best_issue_reason": "issue:45020 is the most central and general issue here; the rest are narrower, model-specific failures or unrelated regressions.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Different bugs: Voxtral AutoTokenizer error messaging vs Molmo missing `all_tied_weights_keys`." + "reason": "Different models and failure modes: Voxtral has a bad AutoTokenizer error message, Molmo is a missing `all_tied_weights_keys` attribute." }, { - "left": "issue:45663", - "right": "issue:45902", + "left": "issue:44910", + "right": "issue:45859", "accept": false, - "reason": "Different models and code paths: Gemma4 FSDP2 shared KV state handling vs Qwen3.5 MoE rotary embedding materialization." + "reason": "Both involve Qwen3.5/MoE, but one is a FlashAttention illegal memory access from `position_ids`, the other is a CPU offload compatibility bug in rotary embedding." }, { - "left": "issue:44910", - "right": "issue:45859", + "left": "issue:45663", + "right": "issue:45902", "accept": false, - "reason": "Both are Qwen3.5-related, but one is a flash-attn position-id crash and the other is CPU-offload incompatibility in rotary embedding; not the same bug." + "reason": "Distinct bugs: Gemma4 FSDP2 `shared_kv_states` reconstruction vs Qwen3_5Moe `inv_freq` uninitialized memory after `to_empty(cuda)`." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Unrelated regressions: LoRA gradient checkpointing vs int4 CUDA reserved-memory growth/OOM." + "reason": "Different regressions: ineffective gradient checkpointing with LoRA versus extra reserved memory/OOM under int4 quantization." }, { "left": "issue:43054", "right": "issue:43493", - "accept": true, - "reason": "Likely the same SigLIP2 implementation discrepancy: poor text embeddings are a symptom of HF diverging from the JAX reference behavior." + "accept": false, + "reason": "Related model family, but not the same bug: one reports worse SigLIP2 text embeddings, the other a HF-vs-JAX implementation discrepancy." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Different models and failures; no shared underlying bug is evident." + "reason": "Separate loading problems for different models; no clear shared code-path or symptom beyond generic load failure." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "Generic loading noise/warnings vs a concrete load failure for a specific model are not the same issue." + "reason": "One is cosmetic loading noise, the other is an actual model load failure for ScandiBERT." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Tokenizer detokenization regression is unrelated to the Python 3.13 `flash_attn` import key error." + "reason": "Unrelated subsystems: tokenizer detokenization bug versus `flash_attn` import-time `KeyError`." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "One is noisy loading output, the other is a specific tokenizer/model load failure." + "reason": "Generic warning/noise issue is not the same as the specific `cjvt/sleng-bert` loading failure." }, { "left": "issue:43643", "right": "issue:45020", - "accept": true, - "reason": "43643 is a specific instance of the broader `remote_code` loading regression described by 45020." + "accept": false, + "reason": "Both involve `trust_remote_code`, but one is missing fields in `AutoConfig`, while the other is a broader remote-code breakage report." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Different training concerns: ineffective gradient checkpointing vs memory inefficiency in Qwen3 MoE training." + "reason": "Different problems: LoRA gradient checkpointing versus Qwen3 MoE training memory usage." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Different failures: Gemma3 ZeRO-3 loading index error vs Qwen3 MoE with FineGrainedFP8Config." - }, - { - "left": "issue:43782", - "right": "issue:44960", - "accept": false, - "reason": "Different models and symptoms; no concrete shared code-path problem is apparent." + "reason": "Different models and errors: Gemma3 ZeRO-3 index error versus Qwen3 MOE FineGrainedFP8 failure." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS histogram support error vs Gemma4 device-mapping mismatch are unrelated." + "reason": "MPS histogram int limitation in MoE is unrelated to Gemma4 PLE device mismatch in forward." }, { - "left": "issue:44991", - "right": "issue:45399", + "left": "issue:43782", + "right": "issue:44960", "accept": false, - "reason": "Tokenizer load failure for one model vs flash-attn2 fallback gating; different subsystems and bugs." + "reason": "Different model families and failure types; one is a Qwen3VL `weight_only` load error, the other is an unspecified GLM5 issue." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "ImportError for a missing class vs bitsandbytes `_is_hf_initialized` TypeError are separate issues." + "reason": "Import error for `Qwen2_5_VLForConditionalGeneration` is unrelated to the bitsandbytes `_is_hf_initialized` compatibility problem." }, { - "left": "issue:42491", - "right": "issue:44368", + "left": "issue:44991", + "right": "issue:45399", "accept": false, - "reason": "LoRA compatibility across HF versions is not the same bug as the `tie_word_embeddings` warning on Qwen3.5." + "reason": "Tokenizer load failure for a specific model versus FlashAttention2 fallback being blocked by unrelated checks." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different model-loading failures with different root causes." + "reason": "Both are loading failures, but for different models and root causes; no evidence they are the same bug." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote_code regression vs Gemma4 PLE device mismatch; different code paths." + "reason": "Remote-code loading regressions are not the same as Gemma4 PLE device placement bugs." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Same broad Qwen3.5 LoRA area, but one is a hf4.x\u2192hf5.x model incompatibility and the other is a tie-word-embeddings warning." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "Tokenizer-mapping failure in add-new-model-like is distinct from the remote_code loading regression." + "reason": "`add-new-model-like` failing in `TOKENIZER_MAPPING_NAMES` is a different registry bug than remote-code loading breakage." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Autocast dtype mismatch and Gemma4 device mismatch are separate runtime bugs." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE and Gemma4 PLE device mismatch are different code paths." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Mistral regex patch crash is unrelated to flash-attn2 fallback gating." + "reason": "Mistral regex patch AttributeError is unrelated to FlashAttention2 fallback gating." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer performance regressions, but they affect different conversion directions and do not clearly share one underlying bug." + "reason": "Both are tokenizer performance regressions, but they affect different APIs and likely different hot paths (`convert_ids_to_tokens` vs `convert_tokens_to_ids`)." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Remote_code loading regression is not the same as the Kimi-K2.5 tokenizer codec-handling regression." + "reason": "`trust_remote_code` loading breakage is not the same as the Kimi-K2.5 tokenizer codec regression." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "MobileLLM key error and flash-attn fallback gating are unrelated." + "reason": "Different areas: MobileLLM loading key error versus FlashAttention2 fallback selection." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "lm_head tying bug in Mistral3 is unrelated to non-persistent buffer corruption during from_pretrained." + "reason": "`lm_head` tying bug and non-persistent buffer corruption are separate state-management regressions." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "These are the same v5 docs cleanup problem: both say removed pipeline task names are still referenced in documentation." + "reason": "These describe the same docs regression: stale v5 documentation still referencing removed pipeline tasks like summarization/translation/text2text-generation." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Both report the same `_is_hf_initialized` / `init_empty_weights` regression when constructing/loading models in transformers 5.x." + "reason": "Same underlying compatibility bug: `init_empty_weights` / bitsandbytes passes `_is_hf_initialized` into `Int8Params.__new__`, causing the same TypeError." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Qwen3 MoE FP8 training failure and the `_is_hf_initialized` TypeError are different issues." + "reason": "Different issues: Qwen3 MOE FineGrainedFP8 failure versus bitsandbytes `_is_hf_initialized` TypeError." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "Too little evidence that the whisper runtime failure and `_torch_extract_fbank_features()` issue are the same underlying bug." + "reason": "Insufficient evidence that the Whisper load failure and `_torch_extract_fbank_features()` issue are the same bug; could be related but not clearly identical." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed and DeepSpeed gradient averaging are separate training bugs." + "reason": "Different DeepSpeed-related problems: Qwen3 MOE weight conversion on load versus gradient averaging under GAS with `model_accepts_loss_kwargs=True`." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted soft pairs are all reasonably supported by the issue titles, especially the remote_code regression, docs cleanup, and `_is_hf_initialized`/`init_empty_weights` family. The cluster is correctly characterized as mostly heterogeneous." + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The two accepted pairs are well supported as same-bug/doc-duplicate cases, and the rejected pairs avoid overclaiming shared root cause." }, "evaluator_used": true, "retried": false, @@ -1572,7 +1581,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:24643", "issue:30064", @@ -1588,6 +1597,7 @@ "issue:43531", "issue:43550", "issue:43632", + "issue:43638", "issue:43645", "issue:43646", "issue:43824", @@ -1633,8 +1643,8 @@ "issue:44479|issue:44560", "issue:44863|issue:46402", "issue:41553|issue:44488", - "issue:44530|issue:44863", "issue:24643|issue:30064", + "issue:44530|issue:44863", "issue:45325|issue:45902", "issue:43366|issue:43828", "issue:45663|issue:45823", @@ -1643,217 +1653,248 @@ "issue:43950|issue:45834", "issue:44560|issue:45685", "issue:43824|issue:43957", - "issue:43550|issue:44910", "issue:42915|issue:43957", - "issue:43645|issue:43646", "issue:43366|issue:44589", + "issue:43550|issue:44910", + "issue:43645|issue:43646", "issue:44560|issue:44918", "issue:42617|issue:43366", "issue:36010|issue:42915", "issue:44661|issue:45823", "issue:44560|issue:44805", "issue:43632|issue:45834", - "issue:38175|issue:43994" + "issue:38175|issue:43994", + "issue:43638|issue:44661" ], - "prepared_review_unit_hash": "9a022a7c7c0ed17e25e4d76e4d4e14566114e009fead79c689752f628d7013bf", + "prepared_review_unit_hash": "acaa79f4964227d1522ad4e4eb03c600d87eaaf1e19d2c290994db6c1ca86393", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 36559, - "estimated_input_tokens": 9140, - "estimated_eval_tokens": 18536 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37482, + "estimated_input_tokens": 9371, + "estimated_eval_tokens": 18998 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of issue reports; only a few soft pairs look like true duplicates. The strongest duplicate subclusters are the custom model init regression, the SigLIP2 wrong-output reports, the mask/index shape-mismatch errors, and the Qwen3.5 video regression reports.", - "confidence": 0.9, - "canonical_issue_reason": "No single canonical issue for the whole set; the items span several unrelated bugs, with only a few small duplicate subclusters.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers regressions across loading, tokenization, attention, and training. Two likely duplicate mini-clusters stand out: the Transformers 5.0.0 custom model initialization breakage, and the SigLIP2 zero-probability/nonsensical-output regression.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43646 is the clearest broad duplicate target for the Transformers 5.0.0 custom model initialization regression; issue:43645 looks like a narrower notebook-specific repro of the same breakage.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43994 is the strongest representative of an actual duplicate pair: it is clearer and broader than issue:38175 while describing the same SigLIP2 base-patch16-224 bad-output problem.", + "best_issue_reason": "issue:43994 is the strongest standalone issue to keep for the SigLIP2 cluster: it has a concrete model-level repro and matches issue:38175\u2019s zero-probability / nonsensical-output symptoms.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Different failures: Trainer API kwarg mismatch vs tensor mask shape mismatch; no shared code path." + "reason": "Different bugs: a custom trainer API kwarg mismatch vs a tensor/mask shape mismatch." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Both mention gradient checkpointing, but one is LoRA ineffectiveness and the other is a missing flag on a trust_remote_code model; not the same bug." + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA checkpointing behavior and the other is a missing capability flag." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Same broad area (DeepSpeed ZeRO-3 loading), but different Qwen model families and different missing-weight patterns." + "reason": "Both are ZeRO-3 loading failures, but they affect different Qwen model families and report different concrete failures." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Unrelated tokenizer-load problems affecting different models; one is an error-message issue, the other is an actual loading failure." + "reason": "Tokenizer loading failures, but for different models and different failure modes." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different models and different load failures; too little evidence of a shared underlying bug." + "reason": "Different model-loading regressions on different architectures; not the same underlying bug." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "Different subsystems and symptoms: RLHF/SDPA gradient instability vs flash_attention attribute crash." + "reason": "One is a Qwen3.5 RLHF gradient explosion from 3D position_ids/SDPA behavior; the other is a flash_attention.py None-handling crash." }, { "left": "issue:41093", "right": "issue:44805", - "accept": true, - "reason": "Near-identical IndexError about mask length not matching indexed tensor length; same underlying shape-mismatch bug." + "accept": false, + "reason": "Both are mask/index shape errors, but the reported shapes and contexts differ enough to indicate separate bugs." }, { "left": "issue:44479", "right": "issue:44560", - "accept": true, - "reason": "Both describe the same v5.3.0 Qwen3/VL video-input regression; 44560 looks like a specific reproduction of the broader failure." + "accept": false, + "reason": "Both involve video on Qwen variants, but one is a broad video-input regression and the other is a StopIteration in a different embedding path." }, { "left": "issue:44863", "right": "issue:46402", "accept": false, - "reason": "Different model-loading issues for different architectures; no clear common bug." + "reason": "Separate loading problems for different model families and different code paths." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Different tokenizer/model-loading failures with different models and symptoms." + "reason": "Both are model/tokenizer loading issues, but they involve unrelated models and symptoms." }, { - "left": "issue:44530", - "right": "issue:44863", + "left": "issue:24643", + "right": "issue:30064", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is unrelated to NemotronH checkpoint loading." + "reason": "Training with DeepSpeed weight-shape error vs image processor failing on void segmentation maps; unrelated." }, { - "left": "issue:24643", - "right": "issue:30064", + "left": "issue:44530", + "right": "issue:44863", "accept": false, - "reason": "Completely different domains: DeepSpeed training weight-shape error vs image processor void-segmentation bug." + "reason": "PagedAttentionCache invalid group-type crash vs NemotronH checkpoint loading failure; different code paths." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Different regressions: Qwen2.5-VL rope scaling vs Qwen3.5-MoE uninitialized rotary embedding memory." + "reason": "Different Qwen3.5-related bugs: one is temporal position_id scaling, the other is uninitialized memory after meta-to-empty materialization." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Different features and failures: GGUF gpt-oss support vs autocast dtype mismatch on Phi-tiny-MoE." + "reason": "GGUF gpt-oss support vs Phi MoE autocast dtype mismatch; unrelated." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Same model family, but different bugs: FSDP2 shared_kv_states KeyError vs device_map auto mismatch." + "reason": "Same model family, but different bugs: per-layer shared_kv_states rebuild vs device_map device mismatch." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Different Qwen/gguf support vs sliding_window behavior; not the same bug." + "reason": "GGUF architecture support issue vs Qwen3-MoE sliding_window behavior; unrelated." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "AudioFlamingo3 batch leak is unrelated to Whisper empty-transcription generation failure." + "reason": "AudioFlamingo3 track leakage vs Whisper empty-transcription handling; different bugs." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Different bugs: non-persistent buffer corruption in from_pretrained vs Kosmos2.5 long OCR index error." + "reason": "Non-persistent buffer corruption regression vs Kosmos2.5 long-OCR index error; different code paths." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Different subsystems: Qwen3-vl video StopIteration vs MoE MPS histogram not implemented." + "reason": "Qwen3-vl video StopIteration vs MoE MPS histogram kernel not implemented for Int." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Both are model-loading regressions, but one is an import/export issue and the other is meta-device initialization; not the same bug." + "reason": "Missing import symbol for a specific class vs meta-device model loading regression." }, { - "left": "issue:43550", - "right": "issue:44910", + "left": "issue:42915", + "right": "issue:43957", "accept": false, - "reason": "Different models and failures: torch.compile/SDPA issue vs FlashAttention illegal-memory-access on Qwen3.5." + "reason": "GenerationMixin import failure vs meta-device loading breakage; unrelated failures." }, { - "left": "issue:42915", - "right": "issue:43957", + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF support request vs Float8 storage lookup error; not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44910", "accept": false, - "reason": "FineGrainedFP8Config failure is unrelated to meta-device loading regressions." + "reason": "Bamba torch.compile/SDPA failure vs Qwen3.5 Flash Attention illegal-memory-access bug from 3D position_ids." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same custom model initialization regression in Transformers 5.0.0; one report is just a notebook-specific variant." + "reason": "Same Transformers 5.0.0 regression: custom model initialization breaks, with 43645 just being a notebook-specific repro of the broader issue." }, { - "left": "issue:43366", - "right": "issue:44589", + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Different failures: video StopIteration in Qwen3-vl-embedding vs TRL SFT unpacking of Qwen3.5 embeddings." + }, + { + "left": "issue:42617", + "right": "issue:43366", "accept": false, - "reason": "Unrelated support/loading request vs float8 storage lookup error." + "reason": "3d_parallel.py execution failure vs GGUF model architecture support; unrelated." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "Missing GenerationMixin import vs FineGrainedFP8Config failure in Qwen3Moe; different problems." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer mapping-name bug vs Gemma4 PLE device mismatch; unrelated." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Video StopIteration and mask/tensor shape mismatch are unrelated failures." + "reason": "Different symptoms and contexts: video StopIteration vs mask shape mismatch." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "Different regressions: _is_hf_initialized flag breakage vs Kosmos2.5 OCR index error." + "reason": "_is_hf_initialized regression vs Kosmos2.5 OCR indexing bug; unrelated." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Same SigLIP2 base-patch16-224 incorrect output bug; 38175's zero probabilities match 43994's nonsensical results symptomatically." + "reason": "Same underlying SigLIP2-base-patch16-224 inference regression: one report shows zero probabilities, the other nonsensical outputs from AutoModel/pipeline." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed ZeRO-3/Bert index error vs tokenizer mapping-name failure; unrelated." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays appropriately conservative overall. The rejected soft pairs are explained with clear title-level differences, and the accepted pairs are plausible duplicates without overclaiming beyond the available evidence. The heterogeneous-set conclusion is supported by the item titles, and no PR-mergeability issue applies here since this is an issue-only packet." + "accept": false, + "feedback": "The summary is mostly conservative, but one key claim is overstated: issue:43994 does not have downstream references in the packet (inbound_references is 0), so that part of the best-issue rationale is unsupported. Keep the 43994/38175 pairing as a likely duplicate cluster, but ground it only in the shared SigLIP2 regression and repro symptoms. The rest of the soft-edge verdicts look appropriately cautious." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1869,7 +1910,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:38175", "issue:39692", @@ -1881,7 +1922,6 @@ "issue:43404", "issue:43454", "issue:43596", - "issue:43638", "issue:43644", "issue:43645", "issue:43708", @@ -1922,7 +1962,6 @@ ], "soft_pairs": [ "issue:44877|issue:45030", - "issue:43638|issue:44661", "issue:43596|issue:44661", "issue:44661|issue:44805", "issue:43317|issue:44756", @@ -1931,8 +1970,8 @@ "issue:42491|issue:43720", "issue:41720|issue:42915", "issue:44960|issue:44977", - "issue:44451|issue:44488", "issue:42915|issue:44756", + "issue:44451|issue:44488", "issue:45313|issue:45357", "issue:42491|issue:45276", "issue:43708|issue:44075", @@ -1940,246 +1979,240 @@ "issue:43644|issue:43645", "issue:39692|issue:43994", "issue:45488|issue:45812", + "issue:43856|issue:43873", "issue:43854|issue:43931", "issue:43257|issue:43854", - "issue:43856|issue:43873", "issue:43856|issue:44387", - "issue:42491|issue:43931", - "issue:44960|issue:46050", "issue:45362|issue:45406", + "issue:44960|issue:46050", "issue:42491|issue:45292", + "issue:42491|issue:43931", "issue:44910|issue:45588", "issue:44164|issue:44291", "issue:45276|issue:45335", - "issue:41720|issue:44155", - "issue:43404|issue:43454" + "issue:43404|issue:43454", + "issue:41720|issue:44155" ], - "prepared_review_unit_hash": "d420d19c4837f760a11bc9f490dd9d951083e57d93747370660e62267f2f3854", + "prepared_review_unit_hash": "d27bfa787da9b42891026160a0bd4e3a2a46cc7621621cd9b672437f3bd59da2", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37374, - "estimated_input_tokens": 9344, - "estimated_eval_tokens": 18944 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36446, + "estimated_input_tokens": 9112, + "estimated_eval_tokens": 18480 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of model-loading, quantization, and generation regressions, with only a few clear duplicate families. The strongest matches are around non-persistent buffers being junk-filled, resize_token_embeddings not propagating to output embeddings, and missing lm_head weight tying in multimodal models. The ZeRO-3 BERT, Qwen3, and similar pairs are better treated as related but distinct bugs rather than true duplicates.", - "confidence": 0.82, - "canonical_issue_reason": "`resize_token_embeddings does not effect to output_embeddings` is the cleanest generic representative of the resize-token-embedding regression family and captures the shared failure more broadly than the model-specific variant.", + "summary": "This cluster is mostly a collection of unrelated bug reports across different models and subsystems, with a few clear duplicate subclusters around config-validation load failures, non-persistent-buffer initialization, and resize_token_embeddings regressions. The BERT-loading pair issue:44451 vs issue:44488 is not strong enough to merge on the available evidence.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45276 is the best canonical for the resize_token_embeddings subcluster: it states the core bug most generally, and it aligns with the related Gemma/T5Gemma reports about embeddings not updating after resizing.", "canonical_pr_reason": null, - "best_issue_reason": "Best canonical issue for the cluster because it states the common bug in a model-agnostic way and is the most reusable anchor for the resize/update-embeddings family.", + "best_issue_reason": "issue:45488 is the strongest standalone report overall: it describes a broad, silent tokenizer regression that breaks a major model family and has clear user impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Both are config-loading regressions, but they affect different models and failure surfaces." - }, - { - "left": "issue:43638", - "right": "issue:44661", - "accept": false, - "reason": "ZeRO-3 BERT init IndexError and tokenizer-mapping helper failure are different code paths." + "reason": "Related loading failures, but the packet does not establish the same root cause; this is too speculative to merge as duplicates." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Same as above: deepspeed zero3 BERT init failure is unrelated to add-new-model-like/tokenizer mapping." + "reason": "Different bugs: a DeepSpeed/BertModel zero-sized tensor error versus add-new-model-like failing on TOKENIZER_MAPPING_NAMES." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer-mapping failure and mask/tensor shape IndexError are unrelated." + "reason": "No shared code path is evident; tokenizer-mapping registration failure and tensor mask shape mismatch are different failures." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Both involve loading/offload behavior, but one is a dequantized-model offload failure and the other is mmap OOM mitigation." + "reason": "Different problems: dequantized-model offload loading failure versus a Strix Halo mmap/OOM workaround." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Both mention SigLIP2, but one is zero probabilities and the other is a docs/example mismatch plus quantization failure." + "reason": "Same model family, but different concrete failures: zero probabilities at inference versus docs/example model-processor and quantization errors." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Same underlying regression: Transformers 5 fills non-persistent buffers with junk." + "reason": "Both describe the same Transformers v5 non-persistent-buffer initialization regression where buffers get filled with junk." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "LoRA/Qwen3 MoE compatibility is a different problem from BitNet packed-weight loading." + "reason": "Different failure modes: qwen3_moe LoRA incompatibility across versions versus BitNet packed-weight unpacking during accelerate loading." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Both are Qwen3-family issues, but the concrete failures differ: CUDA assert vs FP8 config failure." + "reason": "Both involve Qwen3, but one is auto device mapping cuda asserts and the other is FineGrainedFP8Config loading; not the same bug." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Both mention Qwen3.5/flash-attention, but the actual failure modes differ." + "reason": "Different models and symptoms: GLM5 versus Qwen3.5 flash-attention generation failure." }, { - "left": "issue:44451", - "right": "issue:44488", + "left": "issue:42915", + "right": "issue:44756", "accept": false, - "reason": "Both are BERT model-loading failures, but there is no evidence they are the same underlying bug." + "reason": "Unrelated: FineGrainedFP8/Qwen3Moe issue versus mmap/OOM on Strix Halo." }, { - "left": "issue:42915", - "right": "issue:44756", + "left": "issue:44451", + "right": "issue:44488", "accept": false, - "reason": "Qwen3 MoE FP8 failure and mmap/OOM avoidance are unrelated." + "reason": "Both are model-loading failures for BERT-like models, but the packet does not establish a shared root cause or the same failing code path." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Both are Qwen3.5-related, but one is ZeRO-3 weight loading and the other is save_pretrained visual-key regression." + "reason": "Different code paths: DeepSpeed ZeRO-3 language_model loading versus incorrect visual-encoder keys in save_pretrained." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "LoRA/Qwen3 MoE compatibility is unrelated to resize_token_embeddings behavior." + "reason": "Different bugs: qwen3_moe LoRA/version incompatibility versus resize_token_embeddings not propagating to embeddings." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer resume/max_steps logic and SGD argument handling are unrelated." + "reason": "Trainer resume/max_steps calculation bug is unrelated to SGD args not being used." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading regressions, but for different models and likely different root causes." + "reason": "Both are loading failures, but one is a model load regression and the other is tokenizer loading for a specific repo; no clear same root cause." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Non-persistent buffer junk-fill and Jupyter custom-model initialization are unrelated." + "reason": "Different regressions: junk-filled non-persistent buffers versus Jupyter custom-model initialization failure." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both involve SigLIP2, but one is a docs example issue and the other is incorrect AutoModel/pipeline output." + "reason": "Both mention SigLIP2, but one is docs/example breakage and the other is incorrect inference outputs; not the same concrete bug." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions in v5, but the affected models and tokenizer changes differ." + "reason": "Different tokenizer regressions: LlamaTokenizer pre-tokenizer override versus Granite token-id mis-generation." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Different areas: Qwen3 MoE training memory efficiency versus quantization offloading behavior." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different model-loading failures for different models; no shared bug evidence." + "reason": "Different model-loading failures with different models and error types." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to GLM-4.7 Flash test loading." - }, - { - "left": "issue:43856", - "right": "issue:43873", - "accept": false, - "reason": "Both are memory/offload related, but one is training memory inefficiency and the other is quantization/offload behavior." + "reason": "Different bugs: accelerate+deepspeed MOE weight conversion versus GLM-4.7-Flash loading in tests." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Both mention memory usage, but one is Qwen3 MoE training memory and the other is int4 quantization reserved-memory/OOM regression." + "reason": "Training-memory inefficiency and int4 CUDA reserved-memory OOM are related only broadly, not the same bug." }, { - "left": "issue:42491", - "right": "issue:43931", + "left": "issue:45362", + "right": "issue:45406", "accept": false, - "reason": "Different model families and different failure modes." + "reason": "Different model/processor crashes: Qwen3.5 chat crash versus Gemma4Processor missing _tokenizer." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "GLM5 discussion and sinq quantization save/load regression are unrelated." + "reason": "GLM5 issue versus quantized model broken after save/load are unrelated." }, { - "left": "issue:45362", - "right": "issue:45406", + "left": "issue:42491", + "right": "issue:45292", "accept": false, - "reason": "Qwen3.5 chat crash and Gemma4Processor missing _tokenizer are unrelated." + "reason": "Different bugs: qwen3_moe version incompatibility versus resize_token_embeddings not updating output embeddings." }, { "left": "issue:42491", - "right": "issue:45292", + "right": "issue:43931", "accept": false, - "reason": "Qwen3 MoE compatibility and embedding-resize behavior are unrelated." + "reason": "Different failures: qwen3_moe branch incompatibility versus Qwen3-VL weight-shape mismatch." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Different flash-attention failures with different root causes." + "reason": "Both touch flash attention, but one is packed-sequence misinterpretation and the other is an s_aux=None AttributeError." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Save/from_pretrained extra_state handling and init_empty_weights argument handling are distinct bugs." + "reason": "Different serialization/loading regressions: extra_state handling versus init_empty_weights unexpected argument." }, { "left": "issue:45276", "right": "issue:45335", "accept": true, - "reason": "Same concrete regression: resize_token_embeddings does not propagate to model-specific output embeddings/tied modules." + "reason": "Same resize_token_embeddings regression: both report target embeddings not being updated in related model variants." }, { - "left": "issue:41720", - "right": "issue:44155", + "left": "issue:43404", + "right": "issue:43454", "accept": false, - "reason": "Qwen3 auto device mapping assert and AudioFlamingo3 batched inference token leak are unrelated." + "reason": "Both are lm_head tying problems, but they affect different model classes and are not clearly the same concrete fix." }, { - "left": "issue:43404", - "right": "issue:43454", - "accept": true, - "reason": "Same underlying bug class: lm_head weights are not tied in multimodal generation models, leading to broken generation." + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Unrelated: cuda assert during auto device mapping versus AudioFlamingo3 batched inference token/embedding leakage." } ] }, "evaluator_result": { "accept": false, - "feedback": "The output is mostly conservative, but the summary overstates the evidence. It claims there were true duplicates/near-duplicates around the ZeRO-3 BERT init issues, yet the actual verdicts reject those pairs as unrelated. That mismatch makes the summary not fully grounded in the packet. The accepted duplicates around non-persistent buffers and resize_token_embeddings look plausible, but the overall cluster characterization should be toned down." + "feedback": "One soft-edge verdict is too strong for the evidence in the packet: issue:44451 vs issue:44488 are both model-loading failures, but the packet does not establish that they share the same root cause. The acceptance claim (\u201csame load regression affecting different BERT-like models\u201d) goes beyond the titles/metadata, so this is not conservative enough. The rest of the summary is mostly grounded, but this overstates one duplicate pair." }, "evaluator_used": true, "retried": true, @@ -2198,7 +2231,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:36010", "issue:39404", @@ -2211,6 +2244,7 @@ "issue:43645", "issue:43646", "issue:43708", + "issue:43716", "issue:43828", "issue:43854", "issue:43873", @@ -2224,7 +2258,6 @@ "issue:44451", "issue:44479", "issue:44488", - "issue:44530", "issue:44589", "issue:44683", "issue:44756", @@ -2232,7 +2265,6 @@ "issue:44912", "issue:44928", "issue:44960", - "issue:45084", "issue:45216", "issue:45276", "issue:45292", @@ -2254,282 +2286,282 @@ "issue:39692|issue:44960", "issue:44402|issue:44488", "issue:42947|issue:45704", - "issue:43873|issue:44387", "issue:36010|issue:39404", + "issue:43873|issue:44387", "issue:45663|issue:45759", "issue:44291|issue:44756", "issue:45663|issue:45859", "issue:44155|issue:45823", - "issue:44387|issue:44756", "issue:44292|issue:44912", + "issue:44387|issue:44756", "issue:43335|issue:45663", "issue:43957|issue:44756", "issue:42947|issue:43381", "issue:45276|issue:45292", "issue:44164|issue:44756", - "issue:43381|issue:44928", "issue:43854|issue:45216", + "issue:43381|issue:44928", "issue:44479|issue:44877", "issue:45684|issue:45698", "issue:45588|issue:45663", "issue:43646|issue:43708", "issue:43645|issue:44164", - "issue:45812|issue:46315", "issue:45292|issue:45335", + "issue:45812|issue:46315", "issue:43828|issue:43957", - "issue:44402|issue:44451", "issue:43425|issue:43828", + "issue:44402|issue:44451", "issue:44589|issue:44756", "issue:39404|issue:45701", "issue:44683|issue:45823", "issue:43611|issue:43645", "issue:43425|issue:45823", - "issue:44530|issue:45084" + "issue:43425|issue:43716" ], - "prepared_review_unit_hash": "2679b7ee30c27a3601f1a7375fe208d3fda6c3deae526efcdc1894ef38b18045", + "prepared_review_unit_hash": "27560eb294d5eed38b7f9336c7602e9072714cf7dce32354d4191b49c52ccb2d", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 35, - "serialized_chars": 38225, - "estimated_input_tokens": 9557, - "estimated_eval_tokens": 19370 + "serialized_chars": 37650, + "estimated_input_tokens": 9413, + "estimated_eval_tokens": 19082 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous and does not form a broad duplicate set. The only clear duplicate-like pair is the resize_token_embeddings/output_embeddings regression; most other pairs are merely same-subsystem or same-version similarities and should stay separate.", - "confidence": 0.93, - "canonical_issue_reason": "issue:45292 is the cleaner, more general statement of the resize_token_embeddings/output_embeddings regression; issue:45276 looks like the Gemma4-specific manifestation of the same bug.", + "summary": "This cluster is mostly a set of unrelated Transformers issues grouped by loose textual similarity. The only clear duplicate family is the resize_token_embeddings regression affecting output embeddings / decoder embeddings in Gemma4/T5Gemma variants.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 45292 is the best canonical issue because it states the core regression in the most general form: resize_token_embeddings does not update output embeddings. The neighboring reports are narrower model-specific manifestations of the same behavior.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 best represents the shared bug because it states the core failure succinctly and broadly, without extra model-specific noise.", + "best_issue_reason": "Issue 45292 is the strongest global issue target because it is broad, concise, and describes the shared underlying bug rather than a single model variant.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Different failures: generic Torch 2.10 incompatibility vs meta-device loading breakage on some models." + "reason": "Both are version/runtime breakages, but one is Torch compatibility and the other is meta-device model loading; they do not describe the same bug." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "Unrelated topics: SigLIP2 docs example errors vs a GLM5 model issue." + "reason": "Completely different topics: SigLIP2 documentation errors vs. a GLM5 model-loading issue." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Both are model-loading problems, but the reported symptoms differ and there is no clear shared bug." + "reason": "Both are model-loading complaints, but the model names and failure modes differ; not enough evidence of the same underlying bug." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Different code paths: PEFT/gradient checkpointing behavior vs a T5 RMSNorm memory leak." + "reason": "Gradient checkpointing with LoRA and an apex RMSNorm memory leak are unrelated bugs." }, { - "left": "issue:43873", - "right": "issue:44387", + "left": "issue:36010", + "right": "issue:39404", "accept": false, - "reason": "Both involve quantization, but one is offloading behavior and the other is reserved-memory/OOM regression." + "reason": "ImportError for GenerationMixin and Whisper pipeline return_language are different failures with no shared code-path indicated." }, { - "left": "issue:36010", - "right": "issue:39404", + "left": "issue:43873", + "right": "issue:44387", "accept": false, - "reason": "Import path breakage vs Whisper pipeline return_language regression; not the same bug." + "reason": "Both mention quantization, but one is offloading behavior and the other is increased reserved memory/OOM under int4; too broad to treat as the same bug." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Different model families and failures: Gemma4 FSDP attention KeyError vs Qwen config unwrapping in from_config." + "reason": "Gemma4 FSDP2 shared_kv_states and Qwen multimodal config unwrapping are different model-integration problems." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "Unexpected init_empty_weights argument vs mmap/OOM on Strix Halo; no shared code-path evidence." + "reason": "Unexpected _is_hf_initialized during init_empty_weights is unrelated to disabling mmap to avoid OOM." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Both are Qwen/Gemma model issues, but the failures are distinct and not mergeable as one fix." + "reason": "Different models and failure modes: Gemma4 FSDP2 KeyError vs Qwen3_5Moe CPU offload incompatibility." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "Batched inference token leak across tracks vs Gemma4 device_map forward mismatch; unrelated." + "reason": "AudioFlamingo3 batch leakage and Gemma4 device mismatch are unrelated issues." }, { - "left": "issue:44387", - "right": "issue:44756", + "left": "issue:44292", + "right": "issue:44912", "accept": false, - "reason": "Different regressions: int4 CUDA reserved memory/OOM vs mmap-related OOM mitigation." + "reason": "Both involve quantization, but NVFP4 loading failure and MXFP4 fallback to bf16 are not the same bug." }, { - "left": "issue:44292", - "right": "issue:44912", + "left": "issue:44387", + "right": "issue:44756", "accept": false, - "reason": "Both are quantization/load issues, but for different models and formats; too broad to treat as one bug." + "reason": "Int4 quantization OOM and Strix Halo mmap OOM are separate memory issues." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "SwitchTransformers sparse-layer construction bug vs Gemma4 FSDP attention KeyError; unrelated." + "reason": "SwitchTransformers sparse-layer creation and Gemma4 FSDP2 attention-state rebuilding are unrelated." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device loading breakage vs mmap OOM workaround; different failure modes." + "reason": "Meta-device loading breakage is not the same as Strix Halo mmap/OOM behavior." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but one is ineffective with PEFT LoRA and the other is eval-mode restriction." + "reason": "Both mention gradient checkpointing, but one is ineffective with PEFT LoRA and the other is incompatibility in eval mode; different bugs." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same underlying resize_token_embeddings regression: output embeddings are not updated; 45276 is the Gemma4-specific report and 45292 is the generic one." + "reason": "Same underlying regression: resize_token_embeddings does not propagate updated sizes to model output embeddings; 45276 is a model-specific manifestation." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "save/from_pretrained extra_state handling vs mmap OOM on Strix Halo; unrelated." + "reason": "save/from_pretrained extra_state handling and Strix Halo mmap OOM are unrelated." }, { - "left": "issue:43381", - "right": "issue:44928", + "left": "issue:43854", + "right": "issue:45216", "accept": false, - "reason": "Gradient checkpointing eval-mode issue vs RLHF NaN explosion from 3D position_ids; different bugs." + "reason": "Loading GLM-4.7-Flash in tests and Qwen3.5 save_pretrained regression are different issues." }, { - "left": "issue:43854", - "right": "issue:45216", + "left": "issue:43381", + "right": "issue:44928", "accept": false, - "reason": "Model-loading test failure for GLM-4.7-Flash vs Qwen3.5 save_pretrained regression; not the same issue." + "reason": "Gradient checkpointing in eval mode and RLHF NaN from 3D position_ids / SDPA fallback are unrelated." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video-input regression for Qwen VL models vs strict-config loading failure for granite_speech; unrelated." + "reason": "Video-input regression in Qwen VL models and strict config blocking granite_speech are different problems." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Read-only permission propagation in save_pretrained vs wrong custom module loaded after save_pretrained; different bugs." + "reason": "Read-only permissions propagation during save_pretrained and loading the wrong custom module after save_pretrained are distinct save/load issues." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Flash-attention crash on s_aux=None vs Gemma4 FSDP shared_kv_states error; different code paths." + "reason": "Flash-attention sinkless-model crash and Gemma4 FSDP2 shared_kv_states KeyError are unrelated." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization vs Trainer max_steps resume logic; no duplicate signal." + "reason": "Custom model initialization in notebooks and Trainer max_steps calculation on checkpoint resume are unrelated." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook custom-model init regression vs extra_state save/load handling; unrelated." - }, - { - "left": "issue:45812", - "right": "issue:46315", - "accept": false, - "reason": "Tokenizer correctness regression vs tokenizer performance regression; different problems." + "reason": "Notebook custom-model initialization and extra_state save/load handling are different code paths." }, { "left": "issue:45292", "right": "issue:45335", + "accept": true, + "reason": "Same resize_token_embeddings regression, with 45335 showing the decoder.embed_tokens variant of the same missing propagation." + }, + { + "left": "issue:45812", + "right": "issue:46315", "accept": false, - "reason": "Same resize_token_embeddings theme, but 45335 is T5Gemma decoder.embed_tokens-specific and not clearly the same bug." + "reason": "Wrong token IDs from AutoTokenizer and slower convert_tokens_to_ids are different tokenizer regressions." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch on Phi MoE vs meta-device loading breakage; unrelated." + "reason": "Autocast dtype mismatch for Phi-tiny-MoE and meta-device loading breakage are unrelated." }, { - "left": "issue:44402", - "right": "issue:44451", + "left": "issue:43425", + "right": "issue:43828", "accept": false, - "reason": "Token-vocab mismatch in one model vs load failure for another model; too little evidence of a shared bug." + "reason": "Torch version incompatibility and autocast dtype mismatch are not the same bug." }, { - "left": "issue:43425", - "right": "issue:43828", + "left": "issue:44402", + "right": "issue:44451", "accept": false, - "reason": "Torch version compatibility issue vs model-specific autocast dtype mismatch; not the same underlying bug." + "reason": "Both are model-loading failures, but they concern different models and no common failure mode is shown." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup error vs mmap/OOM issue; different root causes." + "reason": "Float8 storage lookup error and mmap OOM on Strix Halo are unrelated." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline return_language regression vs broad tokenization changes across versions; not the same concrete bug." + "reason": "Whisper pipeline return_language regression and tokenizer-version behavior change are different issues." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 vs Gemma4 device mismatch during forward; unrelated." + "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are unrelated." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both are v5 custom-model regressions, but one is about base_model_prefix loading and the other about notebook-defined model initialization; not clearly the same bug." + "reason": "Base_model_prefix loading failure and notebook custom-model initialization are separate custom-model breakages." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "General Torch 2.10 incompatibility vs Gemma4 forward device mismatch; unrelated." + "reason": "Torch 2.10 incompatibility and Gemma4 PLE device mismatch do not share the same underlying bug." }, { - "left": "issue:44530", - "right": "issue:45084", + "left": "issue:43425", + "right": "issue:43716", "accept": false, - "reason": "Qwen3.5 PagedAttentionCache crash vs a template-node compilation TypeError; different failures." + "reason": "Torch compatibility and Mistral-3 image-preprocessor dtype mismatch are different failures." } ] }, "evaluator_result": { "accept": true, - "feedback": "The assessment is conservative and well grounded in the packet. The only accepted pair, issue:45276 vs issue:45292, is a plausible same-bug duplicate family, and the other pair rejections stay within the evidence available from titles/body summaries. The cluster summary does not overstate duplication and correctly characterizes the set as mostly heterogeneous." + "feedback": "The cluster summary is well grounded in the packet, and the soft-edge decisions are conservative. The only clearly supported duplicate family is the resize_token_embeddings/output_embeddings regression across 45276, 45292, and 45335; the other rejected pairs do not have enough evidence of a shared underlying bug." }, "evaluator_used": true, "retried": false, @@ -2548,7 +2580,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:32090", "issue:36032", @@ -2600,7 +2632,7 @@ "issue:46402" ], "soft_pairs": [ - "issue:43425|issue:43716", + "issue:44530|issue:45084", "issue:43906|issue:44530", "issue:39404|issue:44912", "issue:44265|issue:44898", @@ -2613,246 +2645,246 @@ "issue:43299|issue:43931", "issue:43994|issue:44079", "issue:36032|issue:43452", - "issue:44671|issue:44991", "issue:44164|issue:45759", + "issue:44671|issue:44991", "issue:43596|issue:44683", "issue:44991|issue:45081", "issue:43856|issue:44756", "issue:44164|issue:44479", "issue:39404|issue:44671", "issue:44737|issue:45859", - "issue:43582|issue:43632", "issue:44488|issue:45081", + "issue:43582|issue:43632", "issue:44661|issue:44683", "issue:44488|issue:44987", - "issue:32090|issue:39404", "issue:43611|issue:43644", + "issue:32090|issue:39404", "issue:46285|issue:46402", - "issue:40990|issue:43421", "issue:43716|issue:45084", + "issue:40990|issue:43421", "issue:36032|issue:42915", "issue:43643|issue:44530" ], - "prepared_review_unit_hash": "cf79c039b13933bb70acc2c6c81f979cd23d7fd190e12d8686e54725ad317d4a", + "prepared_review_unit_hash": "23fca553bd9a56c734302a0a661de0566b842ed484b0b614b4ad4a5635c04fd1", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37409, - "estimated_input_tokens": 9353, - "estimated_eval_tokens": 18962 + "serialized_chars": 37406, + "estimated_input_tokens": 9352, + "estimated_eval_tokens": 18960 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is not a duplicate set; it\u2019s a loose bucket of unrelated Transformers v5 regressions and model/tokenizer-specific bugs. The closest thing to a representative issue is a broad loading regression, but the soft pairs should all be rejected.", - "confidence": 0.79, - "canonical_issue_reason": "If one issue must represent the cluster, 43611 is the broadest and most central user-facing regression: v5 model loading breaks for a general class of models with `base_model_prefix`.", + "summary": "This set is a grab-bag of mostly unrelated Transformers issues, with several broad v5 regressions and many model-specific loading or runtime bugs. None of the soft pairs look like the same underlying defect closely enough to merge.", + "confidence": 0.93, + "canonical_issue_reason": "#43611 is the best canonical representative because it describes a broad, user-facing Transformers v5 loading regression that can subsume a family of similar breakages; the other issues are narrower, model-specific, or unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "43611 is the best single representative because it is concrete, version-linked, and broadly applicable compared with the more model-specific or docs-only reports.", + "best_issue_reason": "#43611 is the strongest issue to anchor the cluster: it is broad, clearly regression-oriented, and about a core loading path rather than a niche edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43425", - "right": "issue:43716", + "left": "issue:44530", + "right": "issue:45084", "accept": false, - "reason": "Different bugs: Torch 2.10 incompatibility vs image preprocessor/model dtype mismatch." + "reason": "Different failures: a Qwen3.5 PagedAttentionCache crash versus a tokenizer/template compilation TypeError." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "Different code paths: a reproduction of another issue vs a Qwen3.5 PagedAttentionCache crash." + "reason": "Unrelated code paths and symptoms; one is a reproduction of an older issue, the other is a cache/group-type crash on Qwen3.5." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline `return_language` regression is unrelated to gpt-oss MXFP4 quantization fallback." + "reason": "Whisper pipeline language-return regression and GPT-OSS MXFP4 quantization fallback are unrelated bugs." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export/torch_compilable_check failure is unrelated to Perceiver non-default resolution interpolation." + "reason": "torch.export with torch_compilable_check is a different export-path bug than Perceiver image-size handling." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Both are regressions, but one is video input handling and the other is a mask shape IndexError." + "reason": "Different model families and failures: video-input regression versus a mask/index shape mismatch." }, { "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "No shared bug: unrelated reproduction issue vs mistral_common `continue_final_message` behavior." + "reason": "One is a reproduction of an older issue; the other is a mistral_common tokenizer-specific continue_final_message bug." }, { "left": "issue:46050", "right": "issue:46326", "accept": false, - "reason": "Quantized save/load corruption is unrelated to tokenizer message-continuation handling." + "reason": "Quantized save/load corruption is not the same as the mistral_common final-message handling issue." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Different failures: missing Float8 storage vs template-node compilation error." + "reason": "Float8 storage lookup failure and template-node compilation errors are different defects." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both touch docs, but they refer to different removed/changed pipeline behavior." + "reason": "Both are docs-oriented, but they concern different pipeline deprecations and different documentation fixes." }, { "left": "issue:43425", "right": "issue:44918", "accept": false, - "reason": "Environment compatibility issue vs TRL SFT embedding unpacking bug." + "reason": "Torch version incompatibility and Qwen3.5/TRL embedding unpacking are separate issues." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Same model family, but different failures: loading broken vs checkpoint shape mismatch." + "reason": "Both involve Qwen3-VL, but one is a v5 loading regression for MoE models and the other is a specific weight-shape mismatch." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "SigLIP inference nonsense and ModelOutput key assignment are unrelated." + "reason": "SigLIP2 bad inference results and ModelOutput key assignment with None are unrelated." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer method-name conflict is unrelated to gguf file loading failures." + "reason": "A T5 tokenizer method conflict is not the same bug as gguf_file breaking AutoTokenizer/AutoModel loading." }, { - "left": "issue:44671", - "right": "issue:44991", + "left": "issue:44164", + "right": "issue:45759", "accept": false, - "reason": "Different models and symptoms: CamemBERT MLM outputs vs est-roberta tokenizer loading." + "reason": "save/from_pretrained extra_state handling and composite Qwen config unwrapping are different loading/saving paths." }, { - "left": "issue:44164", - "right": "issue:45759", + "left": "issue:44671", + "right": "issue:44991", "accept": false, - "reason": "save/from_pretrained extra_state handling is unrelated to `text_config` unwrapping in composite configs." + "reason": "CamemBERT MLM prediction regression and EMBEDDIA tokenizer loading failure are different model-specific bugs." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "DeepSpeed ZeRO3 init error is unrelated to compiled flex_attention on newer torch." + "reason": "DeepSpeed ZeRO-3/BertModel initialization and compiled flex_attention on newer torch versions are unrelated." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizer loading, but they are different models with different root causes." + "reason": "Different tokenizer problems: EMBEDDIA load failure versus Mistral regex patch crashing on backend_tokenizer." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Memory inefficiency in Qwen3 MoE training is not the same as Strix Halo mmap OOM avoidance." + "reason": "MoE training memory inefficiency and Strix Halo mmap OOM are different memory issues with different causes." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "Save/load extra_state bug is unrelated to the Qwen VL video-input regression." + "reason": "Extra_state save/load handling has nothing to do with the Qwen video-input regression." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline bug and CamemBERT masked-LM regression are unrelated." + "reason": "Whisper pipeline behavior and CamemBERT masked-LM predictions are separate model regressions." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "Both mention device behavior, but one is CPU tensor creation and the other is CPU offload compatibility." + "reason": "Both involve device/offload behavior, but the XLNet CPU arange bug and Qwen3_5MoE rotary embedding incompatibility are distinct." }, { - "left": "issue:43582", - "right": "issue:43632", + "left": "issue:44488", + "right": "issue:45081", "accept": false, - "reason": "Apple Silicon warmup TypeError is unrelated to `_is_hf_initialized` breakage." + "reason": "Loading cjvt/sleng-bert and Mistral regex patching are different tokenizer-loading problems." }, { - "left": "issue:44488", - "right": "issue:45081", + "left": "issue:43582", + "right": "issue:43632", "accept": false, - "reason": "Both are tokenizer-load failures, but for different models and different failure mechanisms." + "reason": "Apple Silicon warmup TypeError and the _is_hf_initialized v5 regression are unrelated." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Model-registration logic and flex_attention compilation are unrelated." + "reason": "Tokenizer mapping-name failure and compiled flex_attention are on different subsystems entirely." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Different model-loading regressions with no shared underlying code-path." + "reason": "They are both loading failures, but for different models and different failure modes." }, { - "left": "issue:32090", - "right": "issue:39404", + "left": "issue:43611", + "right": "issue:43644", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is unrelated to Whisper `return_language` behavior." + "reason": "Both are v5 regressions, but one is about base_model_prefix loading and the other about non-persistent buffers being filled with junk." }, { - "left": "issue:43611", - "right": "issue:43644", + "left": "issue:32090", + "right": "issue:39404", "accept": false, - "reason": "Both are v5 regressions, but one is model-loading logic and the other is non-persistent buffer initialization." + "reason": "Trainer broadcast NoneType and Whisper return_language pipeline behavior are unrelated." }, { "left": "issue:46285", "right": "issue:46402", "accept": false, - "reason": "Different model families and failures: CLIP structure inconsistency vs Sam3 text-encoder weights not loading." + "reason": "CLIP text-model structure inconsistency and Sam3 text-encoder weights not loading are different model-specific issues." }, { - "left": "issue:40990", - "right": "issue:43421", + "left": "issue:43716", + "right": "issue:45084", "accept": false, - "reason": "Perplexity quality issue is unrelated to runtime special-token/post-processor updates." + "reason": "Mistral-3 dtype mismatch during preprocessing/model use is unrelated to the template-node compilation error." }, { - "left": "issue:43716", - "right": "issue:45084", + "left": "issue:40990", + "right": "issue:43421", "accept": false, - "reason": "Dtype mismatch in Mistral-3 vision setup is unrelated to template-node compilation failure." + "reason": "Perplexity quality on gpt-oss and tokenizer post-processor updates after runtime special-token changes are different bugs." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer method conflict is unrelated to Qwen3Moe FP8 config failure." + "reason": "T5 tokenizer method-name conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "AutoConfig remote-code field loss is unrelated to Qwen3.5 linear_attention cache crash." + "reason": "AutoConfig trust_remote_code field loss and Qwen3.5 PagedAttentionCache crashes are different code paths and symptoms." } ] }, @@ -2874,7 +2906,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:34567", "issue:36032", @@ -2910,6 +2942,7 @@ "issue:44530", "issue:44534", "issue:44610", + "issue:44683", "issue:44811", "issue:44849", "issue:44987", @@ -2927,259 +2960,278 @@ "soft_pairs": [ "issue:34567|issue:40444", "issue:40444|issue:46326", - "issue:44811|issue:45584", - "issue:43541|issue:45084", "issue:44451|issue:44987", - "issue:42915|issue:44530", + "issue:43541|issue:45084", + "issue:44811|issue:45584", "issue:42175|issue:46178", + "issue:42915|issue:44530", "issue:43644|issue:43646", "issue:44206|issue:44220", "issue:43452|issue:43482", - "issue:44464|issue:45987", "issue:43883|issue:44488", - "issue:43317|issue:43482", "issue:36032|issue:43645", - "issue:44849|issue:45325", + "issue:43317|issue:43482", + "issue:44464|issue:45987", "issue:45084|issue:45507", "issue:44987|issue:44991", + "issue:44849|issue:45325", "issue:43746|issue:46326", "issue:43262|issue:43377", "issue:43645|issue:44534", "issue:43646|issue:44534", "issue:43404|issue:44466", "issue:44403|issue:45081", - "issue:43089|issue:45987", - "issue:43883|issue:44991", "issue:42915|issue:45084", - "issue:43322|issue:43540", + "issue:43883|issue:44991", + "issue:43089|issue:45987", "issue:43994|issue:44610", + "issue:43322|issue:43540", "issue:43377|issue:43819", "issue:44451|issue:45081", - "issue:44265|issue:44610", "issue:44987|issue:45081", + "issue:44265|issue:44610", "issue:44991|issue:46315", - "issue:42915|issue:44291" + "issue:42915|issue:44291", + "issue:44220|issue:44683" ], - "prepared_review_unit_hash": "5c5bd67e33086538ef9e23c5fd9d98ae5c4c285e36b4bbef6d1fe8b07c5dd363", + "prepared_review_unit_hash": "e57898660f82826c15f72c66d687e11995b0434f6c55b5fd3b38cbc45bbe74f4", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 34, - "serialized_chars": 37483, - "estimated_input_tokens": 9371, - "estimated_eval_tokens": 18998 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38323, + "estimated_input_tokens": 9581, + "estimated_eval_tokens": 19418 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster looks like a set of unrelated false-positive similarities across Transformers issues; none of the soft pairs appear to describe the same underlying bug/change, so I would not merge any of them.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue is a reliable duplicate representative here: the items span unrelated regressions in training, tokenizers, audio, model loading, generation, and backend/dependency handling.", + "summary": "This cluster is loosely centered on Transformers v5/v5.x regressions around model/tokenizer loading, init, and generation, but the listed issues are largely distinct bugs affecting different code paths and models. I found no soft-edge pair that clearly looks like the same underlying change/bug.", + "confidence": 0.34, + "canonical_issue_reason": "issue:43646 is the broadest representative: it describes a generic Transformers 5.0 custom-model initialization regression, which best matches the cluster\u2019s overall v5 load/init theme.", "canonical_pr_reason": null, - "best_issue_reason": "No best issue stands out as the cluster's true duplicate target; the overlap is mostly superficial topic similarity rather than one shared code-path bug.", + "best_issue_reason": "issue:43646 is the most general and actionable issue in the set, with a clear regression framing and broader scope than the model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Trainer token accounting and Qwen2.5-VL multi-image dataset failure are unrelated bugs." + "reason": "Trainer token accounting vs multimodal iterable-dataset failure; unrelated bugs." }, { "left": "issue:40444", "right": "issue:46326", "accept": false, - "reason": "IterableDataset multimodal loading and mistral_common continue_final_message are different tokenizer/model paths." + "reason": "Different areas and failure modes: multimodal dataset handling vs mistral_common tokenizer messaging." }, { - "left": "issue:44811", - "right": "issue:45584", + "left": "issue:44451", + "right": "issue:44987", "accept": false, - "reason": "Both involve Whisper, but one is batch_decode skip_special_tokens and the other is empty-transcription handling after align_special_tokens." + "reason": "Both are model-load failures, but for different models and likely different tokenizer/model code paths." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Torch Dynamo grouped_mm tracing and non-template-node compilation are distinct failure modes." + "reason": "Torch dynamo grouped_mm tracing error is unrelated to template-node compilation failure." }, { - "left": "issue:44451", - "right": "issue:44987", + "left": "issue:44811", + "right": "issue:45584", "accept": false, - "reason": "Different models and different loading regressions; no shared underlying bug is evident." + "reason": "Both involve Whisper, but one is batch_decode special-token handling and the other is empty-transcription generation after align_special_tokens." }, { - "left": "issue:42915", - "right": "issue:44530", + "left": "issue:42175", + "right": "issue:46178", "accept": false, - "reason": "Qwen3Moe FP8 config failure and Qwen3.5 PagedAttentionCache crash are separate issues." + "reason": "Backend packaging omission is not the same as an unsupported backend-mapping error for tensorflow_text." }, { - "left": "issue:42175", - "right": "issue:46178", + "left": "issue:42915", + "right": "issue:44530", "accept": false, - "reason": "Both are backend-related, but missing tensorflow in the torch extra and missing tensorflow_text in BACKENDS_MAPPING are different dependency bugs." + "reason": "Different models and subsystems: FP8 config failure vs paged attention cache linear_attention crash." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both mention Transformers 5.0.0, but one is non-persistent buffers and the other is custom model initialization." + "reason": "Related to v5 initialization regressions, but one is junk non-persistent buffers and the other is custom model init failure." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Audio feature-extraction bugs, but the unsupported center arg and the _torch_extract_fbank_features issue are not the same code-path." + "reason": "Both are audio-feature related, but they point to different failures and code paths." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "Both are GGUF-related, but one is generic from_pretrained breakage and the other is a Qwen2.5-GGUF/v5 load failure." - }, - { - "left": "issue:44464", - "right": "issue:45987", - "accept": false, - "reason": "Chunked generation inconsistency with compiled forward and StaticCache.get_seq_length type mismatch are different generation bugs." + "reason": "Both mention GGUF loading, but one is a generic gguf_file/from_pretrained breakage and the other is a specific Qwen2.5-GGUF v5 failure." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing all_tied_weights_keys and a model load failure for cjvt/sleng-bert are unrelated." + "reason": "Molmo tied-weights attribute error vs a separate model-loading failure; no shared underlying bug." }, { - "left": "issue:43317", - "right": "issue:43482", + "left": "issue:36032", + "right": "issue:43645", "accept": false, - "reason": "Dequantized model offload with device_map=auto and Qwen2.5-GGUF loading failure are different loader issues." + "reason": "Tokenizer method-name conflict and notebook custom-model init regression are unrelated." }, { - "left": "issue:36032", - "right": "issue:43645", + "left": "issue:43317", + "right": "issue:43482", "accept": false, - "reason": "T5Tokenizer add_special_tokens conflict is unrelated to custom model initialization in notebooks." + "reason": "GPU/CPU offload dequantization failure is not the same as a Qwen2.5 GGUF loading regression." }, { - "left": "issue:44849", - "right": "issue:45325", + "left": "issue:44464", + "right": "issue:45987", "accept": false, - "reason": "Different Qwen model families and different bugs: hidden_states output vs rope-index scaling." + "reason": "Chunked-generation inconsistency and StaticCache seq-length type mismatch are different bugs." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Can't compile non-template nodes and GraniteMoEHybrid invalid method are not the same underlying error." + "reason": "Template compilation TypeError vs GraniteMoEHybrid invalid-method call; different concrete failures." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are v5 load regressions, but they affect different models and different loading surfaces (model vs tokenizer)." + "reason": "Different models and symptoms: one model fails to load, the other tokenizer loading fails for a different repo." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden-states bug vs Qwen2.5-VL RoPE position-id scaling bug; unrelated." }, { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "PEFT local-checkpoint loading and mistral_common continue_final_message are unrelated issues." + "reason": "PEFT adapter loading vs continue_final_message behavior in mistral_common tokenizer; not the same bug." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Sampling-rate defaulting in apply_chat_template and MIMI padding-mask output drift are separate audio bugs." + "reason": "Audio chat-template sampling-rate default and MIMI padding-mask mismatch are different audio bugs." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Notebook custom-model init breakage is different from the non-persistent-buffer serialization bug." + "reason": "Both are v5 regressions, but custom model init in notebooks is distinct from non-persistent buffer junk initialization." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Generic custom model initialization regression does not match the buffer-filling/serialization issue." + "reason": "Same reasoning: different init-related regressions with different observed failures." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "Untied lm_head weights in Mistral3 are unrelated to device-dependent lm_head serialization." + "reason": "Tied lm_head weights missing vs serialization differing by device are related broadly, but not the same concrete bug." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Noisy transformer loading and the Mistral regex patch backend_tokenizer AttributeError are different problems." + "reason": "Logging/noise issue vs Mistral tokenizer regex patch crash; different scopes and failures." }, { - "left": "issue:43089", - "right": "issue:45987", + "left": "issue:42915", + "right": "issue:44291", "accept": false, - "reason": "Generation GPU-sync/dispatch overhead is unrelated to StaticCache.get_seq_length returning a tensor." + "reason": "FP8-config model failure and init_empty_weights argument error are separate initialization/configuration bugs." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "A Molmo model attribute bug and an EMBEDDIA tokenizer load failure are unrelated." - }, - { - "left": "issue:42915", - "right": "issue:44291", - "accept": false, - "reason": "FineGrainedFP8Config failure and init_empty_weights/_is_hf_initialized TypeError are different initialization paths." + "reason": "Molmo tied-weights attribute error is unrelated to tokenizer loading for EMBEDDIA/est-roberta." }, { - "left": "issue:43322", - "right": "issue:43540", + "left": "issue:43089", + "right": "issue:45987", "accept": false, - "reason": "Llava Next load-time segfault and Qwen3OmniMoe video-processing ValueError are different multimodal bugs." + "reason": "Generation performance overhead is unrelated to StaticCache seq-length contract issues." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "Siglip2 nonsensical outputs and OmDet-Turbo 640px/224px mismatch are unrelated model issues." + "reason": "Nonsensical outputs from one model and processor/model input-size mismatch in another are distinct problems." }, { - "left": "issue:43377", - "right": "issue:43819", + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next segfault and Qwen3OmniMoe video-processing ValueError are different model-loading/processing failures." + }, + { + "left": "issue:43377", + "right": "issue:43819", "accept": false, - "reason": "MIMI batched-vs-single padding-mask mismatch and DAC.from_latents STE mismatch are different implementation bugs." + "reason": "Both are audio reconstruction/comparison bugs, but they involve different models and different missing mechanisms (padding mask vs STE)." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Loading ScandiBERT and patching Mistral regex/tokenizer internals are unrelated." + "reason": "Generic model load failure and Mistral regex-patch tokenizer crash do not share the same root cause." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different failure surfaces: loading a specific model repo vs tokenizer regex patch crash." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export with torch_compilable_check and OmDet-Turbo input-size mismatch are different failures." + "reason": "torch.export failure with torch_compilable_check is unrelated to OmDet processor image-size mismatch." }, { - "left": "issue:44987", + "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "Model loading failure and slow convert_tokens_to_ids on slow tokenizers are not the same bug." + "reason": "Tokenizer load failure and slow convert_tokens_to_ids performance regression are different bugs." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Different error classes and code paths despite both being v5-era regressions." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio fbank extraction issue and compiled flex_attention torch>=2.9 failure are unrelated." } ] }, @@ -3201,7 +3253,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:33666", "issue:35707", @@ -3216,6 +3268,7 @@ "issue:43421", "issue:43425", "issue:43482", + "issue:43525", "issue:43531", "issue:43611", "issue:43632", @@ -3223,7 +3276,6 @@ "issue:43792", "issue:44062", "issue:44079", - "issue:44220", "issue:44291", "issue:44403", "issue:44462", @@ -3253,13 +3305,12 @@ ], "soft_pairs": [ "issue:45216|issue:45357", - "issue:44220|issue:44683", - "issue:36032|issue:43742", "issue:44062|issue:44987", + "issue:36032|issue:43742", "issue:46209|issue:46326", "issue:43632|issue:44756", - "issue:44479|issue:45325", "issue:42915|issue:44589", + "issue:44479|issue:45325", "issue:36032|issue:41720", "issue:43404|issue:45127", "issue:43531|issue:44877", @@ -3268,228 +3319,229 @@ "issue:43296|issue:43531", "issue:43425|issue:44403", "issue:40990|issue:44811", - "issue:44079|issue:45663", "issue:43792|issue:44589", "issue:43632|issue:44534", - "issue:33666|issue:40444", + "issue:44079|issue:45663", "issue:44403|issue:44991", + "issue:33666|issue:40444", "issue:43421|issue:44568", "issue:44683|issue:45020", "issue:43381|issue:45923", - "issue:35707|issue:40444", "issue:44291|issue:44589", - "issue:44829|issue:45479", + "issue:35707|issue:40444", "issue:44462|issue:45812", - "issue:43482|issue:45538" + "issue:43482|issue:45538", + "issue:44829|issue:45479", + "issue:43482|issue:43525" ], - "prepared_review_unit_hash": "97c1b7b497b137178096c281af5dd0bd5bf6bd6aca464db80ae296f65ba7e101", + "prepared_review_unit_hash": "4d2d30ce9149dbf23e51fb608a3aa96dbff1191552b5d42fff0a9d2456de563f", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35772, - "estimated_input_tokens": 8943, - "estimated_eval_tokens": 18142 + "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 issue reports that only share broad theme words like \"v5 regression\" or \"tokenizer\". I\u2019d accept only two soft matches as likely same underlying bugs: Qwen3.5 save_pretrained regressions, and AutoTokenizer/tokenizer.json loading regressions. The rest are too model- or code-path-specific to merge.", - "confidence": 0.86, - "canonical_issue_reason": "issue:45020 is the best umbrella representative: it captures the broad, recurring class of recent transformers v5 loading regressions around remote_code, which is the closest common thread in this mostly heterogeneous set.", + "summary": "Mostly unrelated issues were grouped together. I only accept two true duplicate pairs: the Qwen3.5 save_pretrained regression and the AutoTokenizer/tokenizer.json Granite regression. The rest are adjacent by subsystem or model family, but not the same underlying bug.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 45812 is the broadest, highest-signal representative of a real duplicate set: it describes the tokenizer regression at the root level, while 44462 reads like a narrower symptom. It also has more inbound references.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the most central and general issue in the set, so it best represents the cluster without being overly tied to one model or one narrow failure mode.", + "best_issue_reason": "45812 is the best single issue to keep for this cluster because it captures the clearest cross-cutting bug and is the most likely root report among the few plausible duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Both report Qwen3.5 save_pretrained producing incorrect checkpoints after recent transformers versions; 45357 looks like a specific continuation of the same save_pretrained regression." + "reason": "Same Qwen3.5 save_pretrained regression: both report incorrect checkpoint serialization, with 45357 reading as a later/continued symptom." }, { - "left": "issue:44220", - "right": "issue:44683", + "left": "issue:44062", + "right": "issue:44987", "accept": false, - "reason": "Different subsystems and failures: fbank feature extraction vs compiled flex_attention on new torch versions." + "reason": "Both are tokenizer-related, but the concrete failures differ: AddedToken argument collision vs loading a specific model tokenizer in transformers 5.1." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Unrelated tokenizer construction error vs a model-loading KeyError for MobileLLM-125M." - }, - { - "left": "issue:44062", - "right": "issue:44987", - "accept": false, - "reason": "Both mention loading problems, but the underlying failure modes and affected code paths are different." + "reason": "Different models and different errors; one is a T5 tokenizer method conflict, the other is a key error loading MobileLLM-125M." }, { "left": "issue:46209", "right": "issue:46326", "accept": false, - "reason": "FP8 inference on B200 is unrelated to a mistral_common tokenizer continue_final_message bug." + "reason": "Completely different code paths: FP8 inference on B200 vs mistral_common continue_final_message behavior." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "_is_hf_initialized regression and Strix Halo mmap OOM are unrelated issues." + "reason": "One is an _is_hf_initialized regression, the other is an mmap/OOM workaround on Strix Halo; not the same bug." }, { - "left": "issue:44479", - "right": "issue:45325", + "left": "issue:42915", + "right": "issue:44589", "accept": false, - "reason": "Both touch Qwen VL models, but one is a broad video-input regression while the other is a specific still-image rope/position-id bug." + "reason": "Both involve FP8, but one is a Qwen3Moe FineGrainedFP8Config failure and the other is a missing Float8 storage type; distinct failures." }, { - "left": "issue:42915", - "right": "issue:44589", + "left": "issue:44479", + "right": "issue:45325", "accept": false, - "reason": "Both involve FP8 terminology, but the reported errors and code paths are different." + "reason": "Same model family, but different regressions: video input handling versus still-image temporal position_id scaling." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Tokenizer method-name conflict vs CUDA assert during auto device mapping are unrelated." + "reason": "No shared underlying bug; tokenizer method conflict vs auto device mapping cuda assert." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both mention tied weights, but one is a missing lm_head tie in a model class and the other is a LoRA/extended-vocab collapse issue." + "reason": "Both mention weights/tie_word_embeddings, but one is a missing lm_head tie in a model class and the other is LoRA merge collapse; not the same fix." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Sliding-window handling in Qwen3-MoE is unrelated to strict config loading for granite_speech." + "reason": "Different issues in different components: Qwen3-MoE sliding_window behavior vs strict config loading for granite_speech." }, { "left": "issue:40444", "right": "issue:45325", "accept": false, - "reason": "Both are Qwen2.5-VL issues, but one is multi-image IterableDataset finetuning and the other is a rope-index temporal-position bug." + "reason": "Same Qwen2.5-VL area, but one is iterable-dataset multi-image finetuning failure and the other is a rope/position-id scaling regression." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "Generic v5 model-loading breakage and junk non-persistent buffers are distinct regressions." + "reason": "Unrelated regressions: base_model_prefix loading broke in v5, while non-persistent buffers are being filled with junk." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM is unrelated to Qwen3-MoE sliding_window behavior." + "reason": "Different model families and different failure modes; PaddleOCR-VL loading vs Qwen3-MoE sliding_window." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch 2.10 compatibility and loading noise are unrelated." + "reason": "Torch version incompatibility is not the same as noisy load logs; these are separate problems." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "Perplexity regression on GPT-OSS is unrelated to Whisper batch_decode special-token handling." - }, - { - "left": "issue:44079", - "right": "issue:45663", - "accept": false, - "reason": "ModelOutput key assignment and Gemma4 FSDP2 KeyError are different bugs." + "reason": "No shared root cause: perplexity issue on gpt-oss-20b vs Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper run failure and Float8 storage lookup error are unrelated." + "reason": "Whisper model runtime failure and Float8 storage lookup failure are unrelated." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "Different v5 regressions with different symptoms and code paths." + "reason": "Different v5 regressions affecting different internal mechanisms; not one underlying bug." }, { - "left": "issue:33666", - "right": "issue:40444", + "left": "issue:44079", + "right": "issue:45663", "accept": false, - "reason": "Both are VL training-related, but multi-GPU training and multi-image IterableDataset failures are not the same bug." + "reason": "ModelOutput key assignment bug and Gemma-4 FSDP2 shared_kv_states KeyError are distinct issues." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Noise during loading and a tokenizer-loading failure for a specific model are unrelated." + "reason": "One concerns load-time noise, the other tokenizer loading for est-roberta; not the same defect." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both are Qwen-VL training issues, but one is multi-GPU training and the other is iterable-dataset multi-image finetuning; different bugs." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Both involve special tokens, but one is runtime post-processor updating and the other is missing BOS/EOS addition; not clearly the same bug." + "reason": "Both are special-token related, but one is runtime post-processor updates and the other is add_special_tokens not inserting BOS/EOS for a specific tokenizer." }, { "left": "issue:44683", "right": "issue:45020", "accept": false, - "reason": "Compiled flex_attention on torch >= 2.9 is unrelated to remote_code loading regressions." + "reason": "Different layers entirely: compiled flex_attention on torch >=2.9 vs remote_code loading regressions." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction and a missing supports_gradient_checkpointing flag are different issues." - }, - { - "left": "issue:35707", - "right": "issue:40444", - "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL multi-image finetuning." + "reason": "Both mention gradient checkpointing, but one is eval-mode misuse and the other is a missing capability flag on a remote-code model." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "Unexpected _is_hf_initialized argument and missing Float8 storage type are distinct TypeErrors." + "reason": "Different failures: init_empty_weights gets an unexpected argument, while this one cannot find a Float8 storage object." }, { - "left": "issue:44829", - "right": "issue:45479", + "left": "issue:35707", + "right": "issue:40444", "accept": false, - "reason": "Degenerate training from flash_attention_3 and zero-loss single-label classification are different defects." + "reason": "Different generation/training scenarios and different failure modes; not the same underlying bug." }, { "left": "issue:44462", "right": "issue:45812", "accept": true, - "reason": "Both point to AutoTokenizer/tokenizer.json loading being wrong, with 45812 describing the downstream symptom of that same regression on Granite models." + "reason": "Same tokenizer regression family: 44462 says AutoTokenizer ignores repository tokenizer.json, and 45812 reports the resulting wrong token IDs for Granite models." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "Qwen2.5-GGUF loading in transformers v5 is unrelated to CLIPTokenizer's model_max_length setting." + "reason": "Unrelated issues: GGUF loading with transformers v5 vs CLIPTokenizer model_max_length semantics." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both involve degenerate training, but the root causes differ: flash_attention_3 behavior vs a loss/label setup issue." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "No meaningful overlap beyond being load-time failures; Qwen2.5-GGUF loading and Llama4Config pad_token_id are separate." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative about the cluster\u2019s heterogeneity. The two accepted soft pairs both look plausibly like the same underlying bug class and are mergeable into a single fix path, while the remaining rejections are appropriately cautious and do not overclaim equivalence." + "feedback": "The summary is grounded in the packet and stays mostly conservative. The two accepted pairs are plausibly the same underlying regressions based on their titles, while the rejected pairs are generally framed as adjacent-but-distinct bugs rather than definite duplicates. No clear overstatement stands out from the available evidence." }, "evaluator_used": true, "retried": false, @@ -3508,7 +3560,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:40444", "issue:41553", @@ -3521,7 +3573,6 @@ "issue:43377", "issue:43454", "issue:43482", - "issue:43525", "issue:43531", "issue:43582", "issue:43606", @@ -3561,21 +3612,20 @@ ], "soft_pairs": [ "issue:43317|issue:43856", - "issue:43482|issue:43525", "issue:43818|issue:46402", "issue:42175|issue:45542", "issue:41553|issue:43582", "issue:44079|issue:44534", "issue:43818|issue:45357", - "issue:43377|issue:45491", "issue:43531|issue:44291", "issue:41720|issue:45081", + "issue:43377|issue:45491", "issue:40444|issue:45381", "issue:43644|issue:44079", "issue:44265|issue:45072", "issue:43749|issue:44079", - "issue:43611|issue:43646", "issue:45325|issue:45381", + "issue:43611|issue:43646", "issue:43632|issue:45072", "issue:42175|issue:43296", "issue:43792|issue:43957", @@ -3588,224 +3638,221 @@ "issue:44291|issue:44877", "issue:43873|issue:45701", "issue:43606|issue:44265", - "issue:43262|issue:43746", - "issue:43296|issue:45800" + "issue:43296|issue:45800", + "issue:43262|issue:43746" ], - "prepared_review_unit_hash": "3b43a34ff5573901c426b0edc319220c3771408bdccb4e9f6df5a08719a6a2cd", + "prepared_review_unit_hash": "2e44c74017f62985395dfa0eba00dc985c726cfadf891b538a2b1aabeac13413", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36753, - "estimated_input_tokens": 9189, - "estimated_eval_tokens": 18634 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35884, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a grab-bag of unrelated issues, with a loose concentration around Transformers v5 model-loading/init regressions. None of the soft pairs look like true duplicates with the same concrete fix path.", - "confidence": 0.71, - "canonical_issue_reason": "Issue 43632 is the most central low-level regression in the group: a v5 initialization flag break that could plausibly underlie several other loading/init failures, making it the best umbrella issue candidate.", + "summary": "Most pairs are related only at a subsystem level and should stay separate. The only clear duplicate-family matches are the Transformers v5 model-initialization regressions (43611/43646) and, more tentatively, the Qwen2.5-VL position-id regressions (45325/45381). No PRs are present.", + "confidence": 0.69, + "canonical_issue_reason": "issue:43646 is the broadest representative of the Transformers v5 initialization regression family; it captures the underlying custom-model init breakage better than the narrower base_model_prefix symptom.", "canonical_pr_reason": null, - "best_issue_reason": "It is broad enough to represent the shared v5 init/load regression theme, but still concrete and code-path specific rather than a vague symptom report.", + "best_issue_reason": "issue:43646 is the most general and reusable issue report among the accepted duplicate candidates, making it the best single issue to anchor triage.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Different problems: device-map/offload loading failure vs MoE training memory inefficiency. Same broad area, not the same bug." - }, - { - "left": "issue:43482", - "right": "issue:43525", - "accept": false, - "reason": "Unrelated failures: GGUF loading under v5 vs missing pad_token_id on Llama4Config." + "reason": "Different bugs: one is load/offload failure with dequantized models, the other is Qwen3 MoE training memory usage." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Different models and failures: Video-LLaVA tower/weight-sharing bug vs Sam3 text encoder weights not loading." + "reason": "Unrelated model families and failure modes: Video-LLaVA weight sharing vs Sam3 text-encoder loading." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both mention TensorFlow/backends, but one is an install-extra dependency gap and the other is a runtime undefined tf backend error; not the same concrete bug." + "reason": "Both involve TensorFlow availability, but one is packaging extras and the other is an undefined backend error; not the same bug." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "AutoTokenizer Voxtral error message vs Apple Silicon TypeError in caching_allocator_warmup are unrelated." + "reason": "Completely different areas: tokenizer error message vs Apple Silicon warmup TypeError." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment bug and non-persistent buffer junk initialization are distinct code paths." + "reason": "One is ModelOutput key assignment, the other is non-persistent buffer serialization; different code paths." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Different multimodal regressions: Video-LLaVA tower issue vs Qwen3.5 visual encoder save_pretrained key regression." - }, - { - "left": "issue:43377", - "right": "issue:45491", - "accept": false, - "reason": "Both are batching-related, but one is MIMI padding-mask output divergence and the other is Gemma3 NaN embeddings from sliding-window/all-padding windows." + "reason": "Different product areas: Video-LLaVA architecture bug vs Qwen3.5 save_pretrained visual-key regression." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior vs init_empty_weights/_is_hf_initialized TypeError are different failures." + "reason": "Qwen3-MoE sliding-window behavior and init_empty_weights/_is_hf_initialized are separate regressions." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device-map cuda assert and Mistral tokenizer regex patch crash are unrelated." + "reason": "Qwen3 device-mapping crash and Mistral tokenizer patch crash are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both are batching/padding-related, but MIMI encoder padding-mask mismatch and Gemma3 NaNs are different bugs." }, { "left": "issue:40444", "right": "issue:45381", "accept": false, - "reason": "Both are Qwen2.5-VL multimodal, but one is iterable-dataset multi-image training failure and the other is video-input position-id regression." + "reason": "Both touch Qwen2.5-VL, but multiple-images-per-prompt finetuning failure is not the same as video position-id regression." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent buffer serialization junk vs ModelOutput key handling are separate issues." + "reason": "Tied-weight serialization inconsistency and ModelOutput key handling are different issues." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export/torch_compilable_check failure and bfloat16 dtype mismatch are unrelated." + "reason": "torch.export/compiler check failure and bfloat16 dtype mismatch are unrelated." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading breakage is not the same as ModelOutput key assignment." - }, - { - "left": "issue:43611", - "right": "issue:43646", - "accept": false, - "reason": "Both are Transformers v5 regressions, but one is base_model_prefix loading and the other is custom model initialization; too broad to merge as the same bug." + "reason": "FSDP CPU RAM efficient loading is a loading-path bug, not a ModelOutput-key issue." }, { "left": "issue:45325", "right": "issue:45381", - "accept": false, - "reason": "Both involve Qwen2.5-VL position IDs, but one is still-image temporal scaling and the other is video-input vision_position_ids; related but not identical." + "accept": true, + "reason": "Same Qwen2.5-VL visual position-id bug family; both report incorrect rope/vision position ids in the video/image path." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": true, + "reason": "Same Transformers v5 model-init/load regression; base_model_prefix failures and custom initialization failures are the same underlying breakage." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "Initialization-flag regression vs inference dtype mismatch are unrelated." + "reason": "_is_hf_initialized handling in v5 and dtype mismatch in inference are different regressions." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "Install/back-end packaging issue vs vLLM model-loading failure for PaddleOCR-VL; not the same bug." + "reason": "Package/backend installation issue vs vLLM model-load failure; not duplicates." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper run failure and meta-device loading breakage are different loading/runtime issues." + "reason": "Whisper runtime failure and meta-device loading breakage are distinct problems." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Different vision-model behaviors: SigLIP2 nonsensical outputs vs Perceiver non-default resolution failure." + "reason": "Different vision-model symptoms: SigLIP2 nonsense outputs vs Perceiver non-default resolution failure." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Both are Qwen/OCR-related position/index problems, but one is rope-index scaling and the other is long OCR input index error; not clearly the same bug." + "reason": "Both mention OCR/video-like inputs, but Qwen2.5-VL position ids and Kosmos2.5 long OCR index errors are not the same bug." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "AyaVision lm_head tying bug and v5 serialization inconsistency are not the same concrete failure." + "reason": "Both involve lm_head/tied weights, but one is a broken tie in model config and the other is device-dependent serialization inconsistency." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3MoE FP8 config failure and Qwen2.5-GGUF loading failure are unrelated." + "reason": "Different model-loading failures: Qwen3Moe FP8 config vs Qwen2.5 GGUF with transformers v5." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Checkpoint resume key mismatch vs Qwen2.5-VL rope-index regression are unrelated." + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL position-id computation." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Qwen multimodal video-input regression and Gemma4 PLE device mismatch are different bugs." + "reason": "Both are device/offload-related, but Qwen video regression and Gemma4 PLE forward mismatch are separate bugs." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "init_empty_weights argument mismatch vs strict config loading are separate issues." + "reason": "Unexpected _is_hf_initialized argument and strict granite_speech config loading are different issues." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization/offload behavior vs tokenization changes are unrelated." + "reason": "Quantization offload behavior and tokenization changes across versions are unrelated." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "CPU offload device mismatch and torch.export failures are different code paths." + "reason": "CPU offload device mismatch and torch.export/compiler-check failure are different code paths." }, { - "left": "issue:43262", - "right": "issue:43746", + "left": "issue:43296", + "right": "issue:45800", "accept": false, - "reason": "Audio chat-template sampling-rate default and PEFT checkpoint loading are unrelated." + "reason": "PaddleOCR-VL/vLLM load failure and torch/transformers version incompatibility are not the same bug." }, { - "left": "issue:43296", - "right": "issue:45800", + "left": "issue:43262", + "right": "issue:43746", "accept": false, - "reason": "PaddleOCR-VL load failure in vLLM vs torch/Transformers version incompatibility are different problems." + "reason": "Audio chat-template sampling-rate default and GraniteSpeech PEFT checkpoint loading are unrelated." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the soft-edge calls are conservative overall. The two accepted duplicate-family pairs are reasonably supported by the issue titles, and the many rejections avoid overstating similarity. No PRs are present, so the null PR fields are appropriate." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -3820,7 +3867,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:34634", "issue:35707", @@ -3880,8 +3927,8 @@ "issue:43606|issue:43819", "issue:43454|issue:43550", "issue:39692|issue:40444", - "issue:46305|issue:46315", "issue:43881|issue:44488", + "issue:46305|issue:46315", "issue:43377|issue:44610", "issue:43454|issue:45072", "issue:43257|issue:45923", @@ -3894,215 +3941,222 @@ "issue:41720|issue:43322", "issue:45325|issue:46315", "issue:43749|issue:44534", - "issue:41720|issue:44977", + "issue:44991|issue:46291", "issue:42617|issue:42915", + "issue:41720|issue:44977", "issue:44466|issue:45357", - "issue:44991|issue:46291", "issue:43526|issue:43883", - "issue:40990|issue:45701" + "issue:40990|issue:45701", + "issue:43881|issue:44991" ], - "prepared_review_unit_hash": "d3604920a2a1d1f71668eee8403b9447e156e5d72901516bf4d52445922f0a96", + "prepared_review_unit_hash": "6e157776a4ad77e721deb1c5b1a02fa5219f3331e063296532503f0eb32a4833", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35931, - "estimated_input_tokens": 8983, - "estimated_eval_tokens": 18222 + "soft_pair_count": 30, + "serialized_chars": 36233, + "estimated_input_tokens": 9059, + "estimated_eval_tokens": 18374 }, "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 set is mostly a grab-bag of unrelated Transformers regressions, so the broad tokenizer regression in issue 45920 is the best single representative issue.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 45920 is the strongest canonical issue here: it describes a concrete v5 tokenizer regression, affects multiple model families, and is broad enough to represent the most plausible duplicate cluster in this set.", + "summary": "The cluster is mostly a grab-bag of model-specific regression reports, but the closest common thread is loading/offloading/quantization behavior. The best umbrella issue is the generic offloading + quantization report.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43873 is the broadest and most reusable issue in the set: it captures the shared offloading/quantization failure mode without being tied to one model, and it has enough discussion to serve as an umbrella.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45920 is the best standalone issue because it is specific, reproducible, and broad in scope without being vague; the other issues are either much narrower or clearly about different failure modes.", + "best_issue_reason": "issue:43873 is the strongest global representative because it is open, broadly phrased, and closest to the recurring loading/offload regressions that show up in nearby issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "Different bugs: high perplexity on gpt-oss/WikiText vs Whisper empty-transcription generation failure after align_special_tokens." + "reason": "Different symptoms and domains: perplexity on GPT-OSS vs Whisper empty-transcription generation." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Bark voice_preset loading issue vs generation behavior with inputs_embeds/past_key_values; unrelated code paths." + "reason": "Unrelated models and bugs: Bark voice preset vs progressive generation with inputs_embeds/past_key_values." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "One is a generation API issue, the other is a device-map/cudaErrorAssert failure on Qwen3; not the same defect." + "reason": "Different code paths: generation with cached embeddings vs Qwen3 auto device mapping crash." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "Both are tokenizer regressions, but one is missing BOS/EOS insertion for mDeBERTa while the other is wrong token IDs across many models; different concrete bugs." + "reason": "Both tokenizer-related, but one is special-token insertion for a specific tokenizer and the other is wrong token IDs across many models; not the same concrete bug." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes on NER/padding vs torch.export failures with torch_compilable_check; different subsystems and failure modes." + "reason": "Different subsystems: LayoutLMv2 tokenizer padding/NER crash vs torch.export failure on torch_compilable_check." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Both mention device_map=auto, but one is a Qwen3 cuda assert and the other is a Gemma4 PLE forward device mismatch; not the same code path." + "reason": "Both mention device_map=auto, but they are different models and different failure modes (assert vs PLE forward mismatch)." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "Bark CPU offload device mismatch vs DAC.from_latents/forward mismatch from missing STE; unrelated models and bugs." + "reason": "Both are device-related, but one is CPU offload mismatch for Bark and the other is DAC latent/STE mismatch." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "AyaVision weight-tying/serialization issue vs Bamba torch.compile SDPA failure; different components and fixes." + "reason": "Unrelated bugs: weight tying in AyaVision vs torch.compile/SDPA on Bamba." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 docs/model-processor mismatch and quantization failure vs Qwen2.5-VL finetuning with multi-image iterable dataset; not the same issue." + "reason": "Different problems: SigLIP2 docs/model-processor mismatch vs Qwen2.5-VL finetuning with multiple images." }, { - "left": "issue:46305", - "right": "issue:46315", + "left": "issue:43881", + "right": "issue:44488", "accept": false, - "reason": "Sliding-window mask off-by-one vs slow convert_tokens_to_ids regression; different functions and symptoms." + "reason": "Both are load failures, but for different models and likely different causes; not enough to be the same bug." }, { - "left": "issue:43881", - "right": "issue:44488", + "left": "issue:46305", + "right": "issue:46315", "accept": false, - "reason": "Both are model loading failures, but glm-4v-9b and cjvt/sleng-bert fail for different reasons and code paths." + "reason": "Completely different areas: streaming window masking vs tokenizer conversion performance." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI padding-mask discrepancy vs OmDet-Turbo processor producing the wrong image size; different preprocessing/model bugs." + "reason": "Different model components: MIMI encoder padding-mask correctness vs OmDet-Turbo processor image size mismatch." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "lm_head weight tying bug vs dtype mismatch in SwitchTransformers/TimmWrapperModel; not the same underlying problem." + "reason": "Different issues: tied weights/garbage generation vs bfloat16 dtype mismatch." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "Qwen3 MoE weights not converted during accelerate/deepspeed loading vs missing supports_gradient_checkpointing flag on Nemotron remote-code variant; unrelated fixes." + "reason": "Different bugs: Qwen3 MoE weight conversion with accelerate/deepspeed vs missing gradient-checkpointing flag." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "StableLmConfig missing pad_token_idx vs composite Qwen config not unwrapping text_config; different config bugs." + "reason": "Different configuration problems: missing pad_token_idx vs from_config text_config unwrapping for composite multimodal configs." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "device_map auto failing to load a dequantized model on gpu+cpu offload is not the same concrete bug as generic offloading not working as expected with quantization." + "reason": "Related offloading/quantization theme, but the reports are still distinct: dequantized model loading on gpu+cpu offload vs generic offloading not working as expected with quantization." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Both involve vision models and shapes, but MIMI padding behavior and Perceiver non-default resolution handling are different defects." + "reason": "Different model families and failures: MIMI batching/padding vs Perceiver image resolution/interpolation." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Perplexity regression vs gradient checkpointing ineffectiveness with PEFT LoRA; unrelated training/inference issues." + "reason": "Unrelated: perplexity regression vs gradient checkpointing ineffectiveness with PEFT LoRA." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Llava Next loading segfault vs quantized-model save_pretrained/from_pretrained breakage; different model families and failure points." + "reason": "Different bug classes: Llava Next load-time segfault vs sinq quantization save/load corruption." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "Weight tying/serialization in AyaVision vs packed-weight unpacking during accelerate loading for BitNet; not the same bug." + "reason": "Different code paths: lm_head weight tying vs packed-weight unpacking during accelerate loading." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Qwen3 auto device-map cuda assert and Llava Next load segfault are distinct model-loading failures with different root causes." + "reason": "Both are loading/runtime failures, but for different models and root causes." }, { "left": "issue:45325", "right": "issue:46315", "accept": false, - "reason": "Qwen2.5-VL rope position scaling bug is unrelated to convert_tokens_to_ids performance regression." + "reason": "Different areas entirely: Qwen2.5-VL rope indexing vs tokenizer conversion speed." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU RAM efficient loading breakage vs non-persistent buffers being filled with junk; different state-loading problems." + "reason": "Both involve loading/state handling, but one is FSDP CPU-efficient loading and the other is non-persistent buffer initialization; not the same bug." }, { - "left": "issue:41720", - "right": "issue:44977", + "left": "issue:44991", + "right": "issue:46291", "accept": false, - "reason": "Both are Qwen-family issues, but auto device mapping cudaErrorAssert and flash-attention generation errors are not the same concrete bug." + "reason": "Unrelated: tokenizer loading regression vs kernels dependency requiring version/revision." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel.py runtime failure vs Qwen3Moe FineGrainedFP8Config failure; unrelated scripts and failure modes." + "reason": "Different model/runtime problems: 3d_parallel.py execution vs Qwen3Moe FineGrainedFP8Config failure." }, { - "left": "issue:44466", - "right": "issue:45357", + "left": "issue:41720", + "right": "issue:44977", "accept": false, - "reason": "Both are serialization-related, but one concerns lm_head.weight tying while the other concerns incorrect visual encoder keys in save_pretrained." + "reason": "Both touch Qwen3-family inference, but one is device-map crash and the other is flash-attention generation quality/behavior; not the same concrete bug." }, { - "left": "issue:44991", - "right": "issue:46291", + "left": "issue:44466", + "right": "issue:45357", "accept": false, - "reason": "Tokenizer loading regression for EMBEDDIA/est-roberta vs kernels import failure due to missing version/revision; different packages and errors." + "reason": "Both are save_pretrained serialization issues, but they affect different tensors and different models." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug vs Molmo missing all_tied_weights_keys attribute; no shared underlying change." + "reason": "Unrelated: BeitImageProcessorFast label reduction vs Molmo tied-weights attribute missing." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "High perplexity on a model evaluation benchmark vs a general tokenization-version change report; not the same issue." + "reason": "Different symptom space: perplexity issue vs tokenizer behavior changes across versions." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Different failures: model load crash vs tokenizer load regression for a different repository." } ] }, @@ -4124,7 +4178,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:36010", "issue:36331", @@ -4148,7 +4202,6 @@ "issue:43708", "issue:43720", "issue:43873", - "issue:43881", "issue:44008", "issue:44186", "issue:44387", @@ -4177,244 +4230,225 @@ ], "soft_pairs": [ "issue:44387|issue:45704", - "issue:43881|issue:44991", "issue:36331|issue:42617", "issue:43065|issue:46402", "issue:44186|issue:45072", - "issue:43577|issue:43708", "issue:44008|issue:45823", + "issue:43577|issue:43708", "issue:36331|issue:42915", - "issue:43450|issue:45701", "issue:44877|issue:44991", + "issue:43450|issue:45701", "issue:43873|issue:44756", - "issue:43404|issue:45357", "issue:43550|issue:43720", + "issue:43404|issue:45357", "issue:42175|issue:46291", - "issue:41720|issue:42617", "issue:44423|issue:45406", + "issue:41720|issue:42617", "issue:43550|issue:43606", "issue:44610|issue:44898", "issue:44186|issue:44610", + "issue:42915|issue:43317", "issue:43295|issue:45701", "issue:45632|issue:45684", - "issue:42915|issue:43317", - "issue:36010|issue:42617", - "issue:41720|issue:45161", "issue:43425|issue:46291", - "issue:44671|issue:44743", + "issue:41720|issue:45161", + "issue:36010|issue:42617", "issue:43525|issue:46291", + "issue:44671|issue:44743", "issue:45081|issue:45812", "issue:43531|issue:43572", "issue:43316|issue:44913" ], - "prepared_review_unit_hash": "118de5ff74138fc3522ea63ba7957c2270fb7780de2b2f28a11db22700426355", + "prepared_review_unit_hash": "e57a32d3b6ab5f0bd3309d26800b2d2c94fcb84652066ca0a15c9ba4594e3b82", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36770, - "estimated_input_tokens": 9193, - "estimated_eval_tokens": 18642 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35944, + "estimated_input_tokens": 8986, + "estimated_eval_tokens": 18228 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: these are mostly distinct Transformers bugs across unrelated models, configs, tokenizers, offloading, and trainer behaviors. None of the soft edge pairs look like the same underlying bug/change, so I would not collapse any of them as duplicates.", - "confidence": 0.89, - "canonical_issue_reason": "issue:43873 is the best hub-style issue here: it is open, has substantial discussion/inbound references, and covers a broad, recurring Transformers failure mode around quantization/offloading. It is the closest thing to a central representative, even though the cluster is too diverse to treat as one duplicate set.", + "summary": "This cluster is heterogeneous: the soft pairs are mostly same-subsystem but not the same bug/change, so I reject them all. The broadest and most representative issue here is the open quantization/offload report, but there is no real duplicate canonical across the set.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 43873 is the broadest, most active, and most referenced report in this set, so it is the best anchor if one must be chosen; however, the cluster does not contain a true duplicate group.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the strongest global issue candidate because it has the highest signal of active impact (discussion + inbound references) and a broader symptom surface than the more narrowly scoped reports.", + "best_issue_reason": "Issue 43873 has the strongest combination of scope (offloading + quantization), discussion, and external references, making it the most suitable representative issue among otherwise unrelated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Both mention memory/OOM-related problems, but one is int4 quantization reserve growth and the other is a T5 apex RMSNorm leak; different bugs and code paths." - }, - { - "left": "issue:43881", - "right": "issue:44991", - "accept": false, - "reason": "Different failures in different models: glm-4v-9b loading vs tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Both involve memory symptoms, but one is int4 quantization OOM and the other is an apex RMSNorm leak; different code paths and fixes." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "A custom Trainer signature mismatch is unrelated to a 3d_parallel.py runtime failure." + "reason": "Trainer compute_loss signature breakage is unrelated to not being able to run 3d_parallel.py." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "Both are Sam3-related, but one is a dummy Conv2d in the pixel decoder and the other is missing text encoder weights; different components." + "reason": "Both are Sam3-related, but one is a dummy Conv2d in the pixel decoder and the other is missing text-encoder weights." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "LayoutLMv2 tokenizer padding/NER crashes and bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." - }, - { - "left": "issue:43577", - "right": "issue:43708", - "accept": false, - "reason": "BLIP2 dtype propagation and Trainer resume/max_steps calculation are different subsystems and different bugs." + "reason": "Tokenizer padding/NER crash and bfloat16 dtype mismatch are unrelated." }, { "left": "issue:44008", "right": "issue:45823", "accept": false, - "reason": "Gemma3n forward variable collision and Gemma4 PLE device mismatch are separate model bugs." + "reason": "Different models and different failures: variable-name collision in Gemma3n vs device mismatch in Gemma4 PLE." }, { - "left": "issue:36331", - "right": "issue:42915", + "left": "issue:43577", + "right": "issue:43708", "accept": false, - "reason": "Unexpected Trainer keyword argument vs Qwen3Moe FineGrainedFP8 loading failure are unrelated." + "reason": "BLIP2 dtype loading bug and Trainer resume/max_steps regression are separate problems." }, { - "left": "issue:43450", - "right": "issue:45701", + "left": "issue:36331", + "right": "issue:42915", "accept": false, - "reason": "Video processor batch-shape bug and a tokenizer-version regression are different areas and symptoms." + "reason": "Unexpected compute_loss kwarg and Qwen3 MoE FP8 failure do not share the same underlying bug." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for granite_speech is unrelated to tokenizer loading failure for est-roberta." + "reason": "Both are loading issues, but one is strict config handling and the other is tokenizer loading for a specific model." }, { - "left": "issue:43873", - "right": "issue:44756", + "left": "issue:43450", + "right": "issue:45701", "accept": false, - "reason": "Quantization/offloading not working and disabling mmap to avoid Strix Halo OOM are both memory-adjacent, but not the same concrete bug." + "reason": "Video processor batching shape bug is unrelated to tokenization changing across versions." }, { - "left": "issue:43404", - "right": "issue:45357", + "left": "issue:43873", + "right": "issue:44756", "accept": false, - "reason": "Mistral3 lm_head tying and Qwen3.5 save_pretrained visual key corruption are distinct model-specific regressions." + "reason": "Both involve memory/offload, but one is quantization offloading and the other is mmap/OOM on Strix Halo." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "torch.compile with SDPA on Bamba-9B-v2 and packed-weight unpacking during accelerate loading are different failure modes." + "reason": "torch.compile/SDPA failure and accelerate packed-weight loading failure are different code paths." }, { - "left": "issue:42175", - "right": "issue:46291", + "left": "issue:43404", + "right": "issue:45357", "accept": false, - "reason": "Install extras missing TensorFlow backend vs kernels v0.15 requiring version/revision are unrelated dependency issues." + "reason": "Mistral lm_head tying and Qwen3.5 visual encoder key saving are unrelated serialization/model issues." }, { - "left": "issue:41720", - "right": "issue:42617", + "left": "issue:42175", + "right": "issue:46291", "accept": false, - "reason": "Qwen3 device-mapping cuda assert and inability to run 3d_parallel.py are not the same bug." + "reason": "Packaging the torch extra and kernels version/revision requirements are unrelated dependency issues." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both involve serve/multimodal processors, but one crashes on 'str'.to and the other on missing _tokenizer; different root causes." - }, - { - "left": "issue:43550", - "right": "issue:43606", - "accept": false, - "reason": "Both are device/runtime mismatches, but Bamba torch.compile+SDPA and Bark-small CPU offload mismatch are separate bugs." + "reason": "Both affect serving, but one is continuous-batching with multimodal inputs and the other is a Gemma4Processor attribute error." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Processor output size mismatch for OmDet-Turbo and Perceiver interpolation failure are different image-processing problems." + "reason": "Both are image-resolution mismatches, but they affect different models and processor/model contracts." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer crashes on NER/batched padding and OmDet processor size mismatch are unrelated input-processing bugs." + "reason": "Tokenizer crash on NER/batched padding is unrelated to OmDet-Turbo image size mismatch." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3 MoE FP8 failure and dequantized model offload failure are different bugs." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor.tokenizer regression in v4.57.5 and generic version-dependent tokenization changes are too broad and not the same concrete defect." + "reason": "Regression around processor.tokenizer/images is not the same as a general tokenization change across versions." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collisions for trust_remote_code and read-only permission propagation in save_pretrained are different filesystem issues." + "reason": "Cache-path collision and read-only permission propagation are separate save/load issues." }, { - "left": "issue:42915", - "right": "issue:43317", + "left": "issue:43425", + "right": "issue:46291", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8 failure and dequantized model offload loading failure are different quantization/loading paths." + "reason": "Torch 2.10 incompatibility and kernels import/version requirements are unrelated." }, { - "left": "issue:36010", - "right": "issue:42617", + "left": "issue:41720", + "right": "issue:45161", "accept": false, - "reason": "GenerationMixin import failure is unrelated to a 3d_parallel execution problem." + "reason": "cudaErrorAssert on auto device mapping is not the same as TP failing for GPT-OSS MoE." }, { - "left": "issue:41720", - "right": "issue:45161", + "left": "issue:36010", + "right": "issue:42617", "accept": false, - "reason": "Qwen3 auto device mapping assert and GPT-OSS MoE only-TP failure are both parallelism-related, but not the same underlying bug." + "reason": "GenerationMixin import error and 3d_parallel.py runtime failure are unrelated." }, { - "left": "issue:43425", + "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Torch 2.10 incompatibility and kernels v0.15 import requirements are different dependency breakages." + "reason": "Missing pad_token_id in Llama4Config is unrelated to kernels import/version handling." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked LM regression and Qwen3.5 recurrent-state reset with cache are unrelated model bugs." - }, - { - "left": "issue:43525", - "right": "issue:46291", - "accept": false, - "reason": "Missing pad_token_id in Llama4Config is not the same as the kernels version/revision import issue." + "reason": "CamemBERT masked-LM regression and Qwen3.5 recurrent-state reset are different model-specific bugs." }, { "left": "issue:45081", "right": "issue:45812", "accept": false, - "reason": "Mistral regex patch/backend_tokenizer crash and Granite tokenizer ID regression are separate tokenizer bugs." + "reason": "Mistral regex patch crash and Granite token-ID regression involve different tokenizer paths and failures." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig missing pad_token_idx are different config regressions." + "reason": "Qwen3-MoE sliding_window behavior and StableLm pad_token_idx regression are unrelated config issues." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct reload default are distinct config serialization/API issues." + "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct reload regression are different config bugs." } ] }, @@ -4436,7 +4470,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:29159", "issue:30064", @@ -4464,7 +4498,6 @@ "issue:43782", "issue:43873", "issue:43881", - "issue:44077", "issue:44186", "issue:44361", "issue:44451", @@ -4473,6 +4506,7 @@ "issue:44514", "issue:44530", "issue:44545", + "issue:44589", "issue:44625", "issue:44991", "issue:45059", @@ -4498,231 +4532,231 @@ "issue:33357|issue:35707", "issue:43388|issue:45059", "issue:30064|issue:34634", - "issue:41720|issue:43746", "issue:43550|issue:43653", - "issue:43873|issue:45704", "issue:43606|issue:43720", "issue:45406|issue:45464", + "issue:43873|issue:45704", + "issue:41720|issue:43746", "issue:44530|issue:46291", "issue:43421|issue:45701", "issue:45357|issue:46402", - "issue:34634|issue:41720", "issue:44492|issue:44509", "issue:43606|issue:43653", "issue:43531|issue:45759", + "issue:34634|issue:41720", "issue:42490|issue:43329", "issue:34567|issue:35707", - "issue:44625|issue:45479", "issue:44514|issue:44545", + "issue:44625|issue:45479", "issue:44451|issue:44991", - "issue:44077|issue:44625" + "issue:43531|issue:44589" ], - "prepared_review_unit_hash": "fcb06e7d3d71e91959da319787851d2be2285184097be68bdb43037c383aebdf", + "prepared_review_unit_hash": "7011f91bff7b82282dc749d02e763fe07497cdd0c91f0d241a4d63124e68e88f", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36234, - "estimated_input_tokens": 9059, - "estimated_eval_tokens": 18374 + "serialized_chars": 36231, + "estimated_input_tokens": 9058, + "estimated_eval_tokens": 18372 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly unrelated. The Qwen2_5_VLProcessor pair (44514/44545) is thematically very close, but the packet does not support a deterministic duplicate call, so I would not accept it as a true duplicate based on this evidence alone.", - "confidence": 0.81, - "canonical_issue_reason": "Issue 44514 is the most specific and representative report in the cluster for the Qwen2_5_VLProcessor batched-input/padding=False crash, but the evidence is not strong enough to treat it as a confirmed duplicate pair.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues. The only clear duplicate is the Qwen2.5-VL `apply_chat_template` crash on batched input with `padding=False` (44514/44545).", + "confidence": 0.64, + "canonical_issue_reason": "Issue 44514 is the earliest and most explicit report of the only confirmed duplicate bug in the set, with a concrete reproducer and clear failure condition.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44514 is the strongest issue candidate because it is concrete, narrowly scoped, and closest to the only near-match in the set.", + "best_issue_reason": "44514 is the strongest representative because it is a precise, reproducible bug report and anchors the only accepted duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Different model families and failure modes; not the same bug." + "reason": "Both involve Qwen multimodal paths, but one is a video-input ValueError and the other is a `from_pretrained`/`weight_only` load error in a different class and code path." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Qwen3Moe FP8 loading failure vs Llava Next segfault; unrelated code paths." + "reason": "Different models and failure modes: FineGrainedFP8 loading for Qwen3-MoE vs a Llava Next segmentation fault." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Both are tokenizer crashes, but on different models with different concrete failure conditions." + "reason": "Tokenizer bugs, but for different tokenizers, tasks, and error mechanisms; not the same underlying issue." }, { "left": "issue:43421", "right": "issue:45812", "accept": false, - "reason": "Runtime special-token/post-processor behavior vs a tokenizer ID regression; different underlying issues." + "reason": "Both are tokenizer-related, but one is runtime post-processor synchronization and the other is a Granite token-id regression." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer API keyword mismatch vs device-mapping CUDA assert; unrelated." + "reason": "Unrelated: trainer loss signature mismatch vs Qwen3 auto device-mapping CUDA assert." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "Image processor label reduction bug vs model loading failure; not the same problem." + "reason": "ImageProcessor label reduction bug and glm-4v model loading failure are separate problems." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "General slow/fast tokenizer inconsistency vs a specific T5 tokenizer load error; not one concrete bug." + "reason": "Both mention tokenizers, but one is slow/fast behavioral inconsistency and the other is a T5Tokenizer load-time attribute conflict." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS CLIP bus error vs generation with inputs_embeds/past_key_values; different failure surfaces." + "reason": "MacOS CLIP bus error and progressive generation with `inputs_embeds`/`past_key_values` are different code paths." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Metrics batching bug vs SAM3 text/bbox behavior; unrelated subsystems." + "reason": "Metrics label-truncation bug vs SAM3 text/bounding-box behavior; no shared concrete defect." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Void segmentation map processing vs BarkProcessor voice_preset behavior; different code paths." - }, - { - "left": "issue:41720", - "right": "issue:43746", - "accept": false, - "reason": "Device-mapping assert vs PEFT/local-checkpoint loading; not the same bug." + "reason": "Void segmentation map processing and BarkProcessor `voice_preset` are unrelated image/audio processor issues." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "torch.compile/SDPA failure vs tokenizer special-token registration issue; unrelated." - }, - { - "left": "issue:43873", - "right": "issue:45704", - "accept": false, - "reason": "Quantization/offloading behavior vs T5 apex RMSNorm memory leak; different failure classes." + "reason": "`torch.compile`/SDPA failure in Bamba is unrelated to BigBirdTokenizer special-token registration." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "CPU offload mismatch in Bark vs BitNet packed-weight handling during accelerate loading; different code paths." + "reason": "Both touch loading/offload behavior, but they are different models and different bugs (CPU offload mismatch vs unpacking packed weights)." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Gemma4Processor missing tokenizer attribute vs Qwen3.5 streaming chat/completions failure; unrelated." + "reason": "Serving crash from missing `_tokenizer` and streaming inference failure on Qwen3.5 are distinct runtime issues." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior and a T5 RMSNorm memory leak are not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Auto device mapping CUDA assert vs PEFT local checkpoint loading; different failure surfaces." }, { "left": "issue:44530", "right": "issue:46291", "accept": false, - "reason": "PagedAttentionCache crash vs kernels version/revision import error; not the same bug." + "reason": "PagedAttention cache group-type crash and `evaluate` import/version requirement are unrelated." }, { "left": "issue:43421", "right": "issue:45701", "accept": false, - "reason": "Specific runtime tokenizer post-processing bug vs broad version-change tokenization report; not one concrete issue." + "reason": "Runtime special-token updates and version-dependent tokenization are too broad and not the same concrete defect." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Different multimodal checkpoint-loading problems affecting different components." - }, - { - "left": "issue:34634", - "right": "issue:41720", - "accept": false, - "reason": "BarkProcessor voice_preset bug vs Qwen3 device-mapping assert; unrelated." + "reason": "Both involve Qwen/Sam3 multimodal weight handling, but one is incorrect save_pretrained keys and the other is missing text-encoder weights on load." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "Typo in cache strategies vs removed-docs task mismatch; different scope and fix." + "reason": "A typo in cache-strategy docs and obsolete pipeline-task documentation are unrelated documentation issues." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "Bark CPU-offload mismatch vs BigBirdTokenizer mask-token decode issue; unrelated." + "reason": "Device-mismatch on CPU offload is not the same underlying bug as a tokenizer mask-token registration problem." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Sliding-window Qwen3-MoE issue vs composite-config unwrapping for Qwen multimodal models; different concrete problems." + "reason": "Qwen3-MoE sliding-window behavior and composite config `text_config` unwrapping are separate model/config issues." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor `voice_preset` and Qwen3 auto device mapping are unrelated." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Trailing-slash import warning vs undefined video-branch helpers in multimodal token counting; unrelated." + "reason": "Trailing-slash import warning and undefined video-branch variables in multimodal token counting are different bugs." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token-count tracking bug vs generation with inputs_embeds/past_key_values; different areas." - }, - { - "left": "issue:44625", - "right": "issue:45479", - "accept": false, - "reason": "Qwen3.5 num_labels propagation vs degenerate single-label classification loss; different model/config issues." + "reason": "TrainerState token counters and progressive generation with embeddings/past KV are unrelated." }, { "left": "issue:44514", "right": "issue:44545", + "accept": true, + "reason": "Same processor, same method, same batched `padding=False` crash; this is a direct duplicate." + }, + { + "left": "issue:44625", + "right": "issue:45479", "accept": false, - "reason": "These titles are very similar, but the packet marks the pair as not deterministic; without stronger body/diff evidence I would not accept them as a confirmed duplicate." + "reason": "Qwen3.5 config propagation and a general zero-loss classification issue are not the same defect." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Different tokenizer-loading failures for different models; no shared concrete code path." + "reason": "Different models and tokenizer load failures; not a duplicate." }, { - "left": "issue:44077", - "right": "issue:44625", + "left": "issue:43531", + "right": "issue:44589", "accept": false, - "reason": "patchtsmixer post_init policy issue vs Qwen3.5 num_labels propagation; unrelated." + "reason": "Qwen3-MoE sliding-window behavior and missing `Float8_e4m3fnStorage` are unrelated." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is mostly conservative, but it overstates the evidence for the only positive pair. In the packet, the 44514/44545 pair is still marked `deterministic_accept: false`, so calling it a clear duplicate and accepting it goes beyond the provided evidence. The rest of the pair judgments look conservative, but this one positive verdict is not grounded enough to accept the cluster as written." + "accept": true, + "feedback": "The summary is conservative and grounded in the packet. The only accepted duplicate pair, issue:44514 / issue:44545, is strongly supported by the nearly identical titles and shared failure mode, and the rejected soft edges are explained cautiously without overclaiming. No clear overstatement stands out." }, "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": false, + "retried": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -4737,7 +4771,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:30064", "issue:36010", @@ -4759,12 +4793,13 @@ "issue:43927", "issue:43994", "issue:44062", + "issue:44077", "issue:44186", "issue:44220", "issue:44368", "issue:44451", - "issue:44589", "issue:44610", + "issue:44625", "issue:44671", "issue:44843", "issue:44898", @@ -4788,255 +4823,252 @@ "issue:46294" ], "soft_pairs": [ - "issue:45759|issue:45859", - "issue:43531|issue:44589", "issue:44898|issue:45072", + "issue:44077|issue:44625", + "issue:45759|issue:45859", "issue:43421|issue:43927", "issue:45588|issue:45759", - "issue:44368|issue:45357", "issue:44062|issue:44451", + "issue:44368|issue:45357", "issue:44987|issue:45042", + "issue:44918|issue:44936", "issue:43782|issue:45381", "issue:39692|issue:44610", - "issue:44918|issue:44936", "issue:43746|issue:46050", - "issue:45704|issue:45923", "issue:42886|issue:44843", + "issue:45704|issue:45923", "issue:43381|issue:43844", "issue:43421|issue:43653", "issue:45663|issue:45923", "issue:45704|issue:45859", "issue:41720|issue:45704", "issue:43531|issue:43645", - "issue:43421|issue:44186", "issue:43994|issue:44671", + "issue:43421|issue:44186", "issue:43746|issue:43873", "issue:45081|issue:45356", "issue:39692|issue:43873", "issue:43540|issue:46294", + "issue:43596|issue:44220", "issue:42915|issue:43746", "issue:45356|issue:45907", - "issue:43596|issue:44220", "issue:30064|issue:36010" ], - "prepared_review_unit_hash": "83ce9842732a8668f7c49422752a108bb1336032f3b5e59e6ee23a9654308127", + "prepared_review_unit_hash": "a0e70ac03b9be15595bee2d41bc7c934deab096870a253bcc7bafafe56b011f6", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36505, - "estimated_input_tokens": 9127, - "estimated_eval_tokens": 18510 + "serialized_chars": 37086, + "estimated_input_tokens": 9272, + "estimated_eval_tokens": 18800 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers issues across tokenizer, multimodal, training, and quantization paths. Only the offline tokenizer/cache regression pair looks like the same underlying bug.", - "confidence": 0.86, - "canonical_issue_reason": "issue:44843 is the clearest root-cause report: it names the concrete failing path (`_patch_mistral_regex` calling `model_info()` during offline loading). issue:42886 is the broader symptom report, so 44843 is the better duplicate anchor.", + "summary": "This cluster is mostly a grab-bag of unrelated Transformers bugs and feature requests. The soft pairs are mostly thematic lookalikes (tokenizers, Qwen, quantization, training), but they do not appear to be the same underlying defect or change.", + "confidence": 0.89, + "canonical_issue_reason": "No clear canonical issue: the items span unrelated subsystems and failure modes, so none is a strong duplicate representative for the whole cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44843 is the best issue to keep as the representative because it ties the user-visible offline failure to a specific code path and fixable root cause.", + "best_issue_reason": "No single best issue emerges because the cluster is heterogeneous and the soft-paired items are not true duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45759", - "right": "issue:45859", + "left": "issue:44898", + "right": "issue:45072", "accept": false, - "reason": "Different bugs: one is `from_config` / composite multimodal config unwrapping, the other is CPU offload incompatibility in a MoE rotary embedding." + "reason": "Both are vision-related, but one is Perceiver resolution handling and the other is a bfloat16 dtype mismatch in different models/code paths." }, { - "left": "issue:43531", - "right": "issue:44589", + "left": "issue:44077", + "right": "issue:44625", "accept": false, - "reason": "Unrelated failures: Qwen3-MoE sliding window handling vs a float8 storage import/type error." + "reason": "Unrelated problems: patchtsmixer post_init policy vs Qwen3.5 num_labels propagation." }, { - "left": "issue:44898", - "right": "issue:45072", + "left": "issue:45759", + "right": "issue:45859", "accept": false, - "reason": "Both mention model/runtime mismatches, but they affect different models and different code paths (Perceiver resize vs dtype mismatch in SwitchTransformers/TimmWrapperModel)." + "reason": "Both involve Qwen configs, but one is composite config unwrapping and the other is CPU-offload incompatibility in a rotary embedding path." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Tokenizer runtime special-token update bug is not the same as DiaConfig losing custom token IDs after save/load." + "reason": "Token ID / special-token handling is broadly related, but one is runtime post-processor updates and the other is save/load config persistence." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "Different subsystems and failures: flash-attention sinkless-model crash vs config unwrapping for Qwen multimodal models." + "reason": "Different code paths: flash-attention sink handling vs Qwen multimodal config unwrapping." }, { - "left": "issue:44368", - "right": "issue:45357", + "left": "issue:44062", + "right": "issue:44451", "accept": false, - "reason": "A warning about `tie_word_embeddings` is unrelated to incorrect visual encoder keys being saved." + "reason": "AddedToken argument duplication is a tokenizer construction bug; loading ScandiBERT is a separate model-loading failure." }, { - "left": "issue:44062", - "right": "issue:44451", + "left": "issue:44368", + "right": "issue:45357", "accept": false, - "reason": "`AddedToken` keyword collision is a different tokenizer construction bug from the broader ScandiBERT load failure." + "reason": "One is a warning about tie_word_embeddings, the other is incorrect visual-encoder key serialization." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Loading a specific model failing is unrelated to PIL image processors unnecessarily requiring torchvision." + "reason": "Loading a specific model versus PIL image processors wrongly requiring torchvision; different failures." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Trainer/SFT embedding unpacking and evaluate-after-train are distinct trainer lifecycle bugs." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Different multimodal problems: Qwen3VL weight-only loading vs Qwen2.5-VL video position IDs." + "reason": "Both are Qwen vision issues, but one is weight_only loading and the other is incorrect video position IDs." }, { "left": "issue:39692", "right": "issue:44610", "accept": false, - "reason": "Both involve processor/model mismatches, but for different models and different symptoms; not the same underlying bug." + "reason": "Documentation example errors for SigLIP2 are unrelated to OmDet-Turbo processor resolution mismatch." }, { - "left": "issue:44918", - "right": "issue:44936", + "left": "issue:43746", + "right": "issue:46050", "accept": false, - "reason": "One is Qwen3.5 embedding unpacking with TRL SFT, the other is evaluate-after-train behavior; no concrete shared code path." + "reason": "PEFT local checkpoint loading and quantized save/load corruption are separate persistence paths." }, { - "left": "issue:43746", - "right": "issue:46050", + "left": "issue:42886", + "right": "issue:44843", "accept": false, - "reason": "PEFT-adapter checkpoint loading and sinq quantized model save/load are distinct persistence bugs." + "reason": "Both mention HF_HUB_OFFLINE, but 44843 is specifically _patch_mistral_regex/model_info, not the general cache-load failure in 42886." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "RMSNorm memory leak in T5 is unrelated to a missing gradient-checkpointing flag in Nemotron." - }, - { - "left": "issue:42886", - "right": "issue:44843", - "accept": true, - "reason": "Same offline tokenizer-loading regression: 42886 reports cache loading failing under HF_HUB_OFFLINE, and 44843 identifies the offending `_patch_mistral_regex`/`model_info()` call." + "reason": "A T5 RMSNorm memory leak is unrelated to a missing supports_gradient_checkpointing flag." }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction and ZeRO-3 gradient growth are different training issues." + "reason": "Gradient checkpointing in eval mode and gradient blow-up under ZeRO-3 are different problems." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Both concern token special-token handling, but one is runtime post-processor updates while the other is BigBird mask-token registration/decode output." + "reason": "Both touch special tokens, but one is runtime post-processing updates and the other is BigBirdTokenizer mask-token registration." }, { "left": "issue:45663", "right": "issue:45923", "accept": false, - "reason": "Gemma-4 FSDP2 shared KV-state bug is unrelated to a missing model flag in Nemotron." + "reason": "Gemma4 shared_kv_states rebuild bug is unrelated to a trust_remote_code flag omission." }, { "left": "issue:45704", "right": "issue:45859", "accept": false, - "reason": "T5 RMSNorm memory leak and Qwen3_5Moe CPU offload incompatibility are unrelated." + "reason": "Different failures: apex FusedRMSNorm leak vs Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert is unrelated to T5/apex RMSNorm leakage." + "reason": "Qwen3 auto device mapping CUDA assert and T5 RMSNorm memory leak are unrelated." }, { "left": "issue:43531", "right": "issue:43645", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior and Jupyter custom-model initialization are unrelated." + "reason": "Sliding-window behavior in Qwen3-MoE is separate from Jupyter custom-model initialization in Transformers 5.0." }, { - "left": "issue:43421", - "right": "issue:44186", + "left": "issue:43994", + "right": "issue:44671", "accept": false, - "reason": "Special-token post-processor update bug does not match LayoutLMv2 NER/padding crashes." + "reason": "Both are bad outputs, but they affect different models and likely different preprocessing/model-selection bugs." }, { - "left": "issue:43994", - "right": "issue:44671", + "left": "issue:43421", + "right": "issue:44186", "accept": false, - "reason": "Different model-specific inference issues: SigLIP2 pipeline output vs CamemBERT masked LM predictions." + "reason": "Runtime special-token updates and LayoutLMv2 NER/padding crashes are not the same bug." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "Checkpoint loading with PEFT adapters is unrelated to quantization/offloading behavior." + "reason": "PEFT checkpoint loading and quantization/offloading behavior are related only at a high level, not the same code path." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both touch Mistral/Kimi tokenizer logic, but one is a hard crash on missing `backend_tokenizer` and the other is a codec/warning regression; not enough to call them the same bug." + "reason": "Both involve fix_mistral_regex, but one crashes on backend_tokenizer access while the other is a codec-handling regression in a different tokenizer." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 docs/processor errors are unrelated to quantization offloading behavior." + "reason": "SigLIP2 docs/example mistakes are unrelated to quantization offloading behavior." }, { "left": "issue:43540", "right": "issue:46294", "accept": false, - "reason": "Both involve video inputs, but the models and failure modes differ: Qwen3OmniMoe video processing vs Qwen3.5 chat-template feature/token mismatch." + "reason": "Both are video-input issues in Qwen models, but they fail at different stages with different root causes." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "DeepSpeed zero3/BertModel indexing and audio fbank extraction are unrelated." }, { "left": "issue:42915", "right": "issue:43746", "accept": false, - "reason": "Qwen3Moe fine-grained FP8 config failure is unrelated to GraniteSpeech PEFT checkpoint loading." + "reason": "FineGrainedFP8Config failure in Qwen3Moe is unrelated to GraniteSpeech local checkpoint loading." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Tokenizer codec/fix_mistral_regex regression is unrelated to `list_repo_files` not forwarding `revision`." - }, - { - "left": "issue:43596", - "right": "issue:44220", - "accept": false, - "reason": "Deepspeed ZeRO-3/BertModel index error is unrelated to `_torch_extract_fbank_features()`." + "reason": "Tokenizer regression in Kimi-K2.5 is unrelated to list_repo_files missing revision forwarding." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Image processor void segmentation maps and a GenerationMixin import error are unrelated." + "reason": "Void segmentation map processing and a GenerationMixin import error are completely different issues." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet, and the only accepted soft-edge pair (42886/44843) is supported by the titles and described failure mode. The other pairwise rejections are conservative and do not overstate evidence. No PR-mergeability issue 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 }, @@ -5051,7 +5083,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:30064", "issue:36331", @@ -5073,7 +5105,6 @@ "issue:43716", "issue:43720", "issue:43756", - "issue:43782", "issue:43818", "issue:43819", "issue:43950", @@ -5086,13 +5117,14 @@ "issue:44661", "issue:44671", "issue:44811", + "issue:44821", "issue:44843", "issue:44877", "issue:44898", - "issue:44977", "issue:44991", "issue:45072", "issue:45081", + "issue:45084", "issue:45183", "issue:45381", "issue:45479", @@ -5105,8 +5137,8 @@ "issue:43720|issue:43819", "issue:43653|issue:43720", "issue:44220|issue:44661", - "issue:43540|issue:44008", "issue:43550|issue:45072", + "issue:43540|issue:44008", "issue:43295|issue:45834", "issue:43618|issue:43697", "issue:44062|issue:44488", @@ -5114,219 +5146,231 @@ "issue:45183|issue:45479", "issue:44186|issue:44898", "issue:43322|issue:45685", - "issue:43388|issue:46032", - "issue:43335|issue:43756", - "issue:43531|issue:43643", "issue:43716|issue:44008", "issue:43653|issue:43819", - "issue:43643|issue:44877", - "issue:44843|issue:45081", + "issue:43531|issue:43643", + "issue:43335|issue:43756", + "issue:43388|issue:46032", "issue:43950|issue:45987", + "issue:44843|issue:45081", + "issue:43643|issue:44877", "issue:30064|issue:36331", "issue:44448|issue:44671", "issue:41720|issue:44991", "issue:43540|issue:45381", "issue:43421|issue:44811", "issue:43577|issue:43818", - "issue:43782|issue:44977", + "issue:44821|issue:45084", "issue:39692|issue:45834" ], - "prepared_review_unit_hash": "0cbab9a2adaae7ee823d72ba1208f6f358e058e8a4b1521918b34c92bdc7c499", + "prepared_review_unit_hash": "0ae1db94a7e5ad43226dc81a019702be45f0502d1d5159c79c951129dfd3952f", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35999, - "estimated_input_tokens": 9000, - "estimated_eval_tokens": 18256 + "serialized_chars": 35954, + "estimated_input_tokens": 8989, + "estimated_eval_tokens": 18234 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most items are unrelated regressions across tokenizers, multimodal models, training utilities, and config loading. There is a small Mistral-tokenizer subtheme, but none of the soft pairs look like true duplicates, and there are no PRs in the cluster.", + "summary": "The set is highly heterogeneous: most items are unrelated bugs spanning tokenizers, multimodal processors, model loading, caching, and model-specific forward paths. No soft edge looks like the same underlying duplicate, though a few are loosely related by theme.", "confidence": 0.91, - "canonical_issue_reason": "No single true canonical duplicate stands out. If one issue must represent the cluster, issue 44843 is the closest fit because it is the most detailed report in the only loose shared subtheme (_patch_mistral_regex / Mistral tokenizer loading).", + "canonical_issue_reason": "issue:44843 is the closest representative of the only obvious subcluster: a concrete tokenizer-loading regression with a direct code-path failure and a nearby related Mistral regex issue. Even so, the overall cluster is too mixed to treat it as a strong canonical duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44843 is the best representative only by proximity: it has the clearest root-cause description, multiple inbound references, and aligns with the closest related issue in the set. It is still not a global duplicate anchor for the whole cluster.", + "best_issue_reason": "issue:44843 is the best single issue to anchor the cluster because it is a specific, reproducible regression in core loading logic and is the only item that clearly sits near another issue in the set. It is still not a safe umbrella for the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "Different code paths: BitNet loading/unpacking vs DAC latent reconstruction." + "reason": "Different failures: accelerate loading of packed BitNet weights vs DAC latent reconstruction without STE." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "Tokenizer special-token registration bug vs accelerate loading of packed BitNet weights." + "reason": "Tokenizer special-token registration bug and BitNet packed-weight loading bug are unrelated." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio fbank feature extraction and tokenizer-mapping code are unrelated." + "reason": "Audio feature extraction issue vs add-new-model-like failing inside TOKENIZER_MAPPING_NAMES; different code paths." }, { - "left": "issue:43540", - "right": "issue:44008", + "left": "issue:43550", + "right": "issue:45072", "accept": false, - "reason": "Both are multimodal, but the failures are different and in different models/code paths." + "reason": "Different model/runtime problems: torch.compile+SDPA on Bamba vs bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." }, { - "left": "issue:43550", - "right": "issue:45072", + "left": "issue:43540", + "right": "issue:44008", "accept": false, - "reason": "Torch.compile SDPA issue vs dtype mismatch in a different model family." + "reason": "Video-input processing failure in Qwen3OmniMoe vs Gemma3n variable-name collision; not the same bug." }, { "left": "issue:43295", "right": "issue:45834", "accept": false, - "reason": "Processor/tokenizer regression vs OCR long-input index error." + "reason": "Processor.tokenizer regression and long OCR index error are separate failures in different models." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "Missing CLIPOutput attentions field is unrelated to RTDetr output drift." + "reason": "CLIP attentions serialization issue and RTDetrV2 output divergence are different model behaviors." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken constructor argument conflict vs tokenizer load failure for a different model." + "reason": "AddedToken special-argument duplication vs loading a specific tokenizer model are unrelated." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Different model bugs and unrelated runtime failures." + "reason": "Both are model bugs, but one is compile/SDPA and the other is DAC latent/forward mismatch; not mergeable as one fix." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Input validation/XOR messaging bug vs zero-loss classification bug." + "reason": "Validation error messaging for input_ids/inputs_embeds is unrelated to single-label classification zero-loss behavior." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Tokenizer crash on NER/padding vs Perceiver image-resolution handling." + "reason": "LayoutLMv2 tokenizer padding/NER crash and Perceiver image-classification resolution issue are distinct." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Llava Next segfault and an MPS histogram dtype limitation are unrelated." + "reason": "Llava Next loading crash and MPS histogram int support error are completely different subsystems." }, { - "left": "issue:43388", - "right": "issue:46032", + "left": "issue:43716", + "right": "issue:44008", "accept": false, - "reason": "Metrics gathering truncation vs Mamba2 cache correctness." + "reason": "Both touch multimodal inputs, but one is image-preprocessor dtype mismatch and the other is audio mask attribute collision." }, { - "left": "issue:43335", - "right": "issue:43756", + "left": "issue:43653", + "right": "issue:43819", "accept": false, - "reason": "Sparse-layer config bug vs RoPE-layer-count mismatch; different model settings." + "reason": "Special-token registration in BigBirdTokenizer is unrelated to DAC forward/from_latents mismatch." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Sliding-window MoE behavior and trust_remote_code config loading are different issues." + "reason": "Sliding-window behavior in Qwen3-MoE and missing fields with trust_remote_code are unrelated." }, { - "left": "issue:43716", - "right": "issue:44008", + "left": "issue:43335", + "right": "issue:43756", "accept": false, - "reason": "Preprocessor/model dtype mismatch vs variable-name collision in forward()." + "reason": "SwitchTransformers sparse-layer creation and Smollm3 RoPE-layer dropping are different configuration bugs." }, { - "left": "issue:43653", - "right": "issue:43819", + "left": "issue:43388", + "right": "issue:46032", "accept": false, - "reason": "Tokenizer special-token registration and DAC STE behavior are unrelated." + "reason": "Metrics label trimming bug and Mamba2 cache wrong-results bug are separate code paths." }, { - "left": "issue:43643", - "right": "issue:44877", + "left": "issue:43950", + "right": "issue:45987", "accept": false, - "reason": "Both involve config loading, but one is missing fields from trust_remote_code and the other is strict config rejection." + "reason": "Non-persistent buffer corruption on load is unrelated to StaticCache get_seq_length shape/type contract." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Same helper name, but different concrete bugs: offline model_info call vs missing backend_tokenizer attribute." + "reason": "Same helper name, but different failures: offline model_info call vs backend_tokenizer AttributeError. Not the same underlying bug." }, { - "left": "issue:43950", - "right": "issue:45987", + "left": "issue:43643", + "right": "issue:44877", "accept": false, - "reason": "Non-persistent buffer corruption on load vs StaticCache return-type contract." + "reason": "Missing fields from trust_remote_code config and strict granite_speech config rejection are separate loading issues." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image processor segmentation-map handling vs CustomTrainer.compute_loss signature mismatch." + "reason": "Void segmentation maps in image processor and CustomTrainer compute_loss signature mismatch are unrelated." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Versioned output differences for different models, not one shared bug." + "reason": "Pegasus v4/v5 output differences and CamemBERT masked-LM prediction regressions are model-specific and distinct." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "CUDA assert in device mapping vs tokenizer loading failure." + "reason": "A800 cudaErrorAssert with auto device mapping and tokenizer loading failure for est-roberta are different problems." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Both mention video/Qwen, but the model variants and failure modes differ." + "reason": "Both involve video inputs, but they affect different models and different failures (processing error vs vision_position_ids mismatch)." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Runtime post-processor updates and batch_decode skip_special_tokens are related only at a broad tokenizer level." + "reason": "Runtime special-token/post-processor update bug and Whisper batch_decode skip_special_tokens bug are not the same code path." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "BLIP2 dtype propagation and Video-LLaVA temporal-attention/weight-sharing are different bugs." + "reason": "Blip2 dtype loading mismatch and Video-LLaVA missing temporal attention/shared weights are unrelated." }, { - "left": "issue:43782", - "right": "issue:44977", + "left": "issue:44821", + "right": "issue:45084", "accept": false, - "reason": "Qwen-family issues, but one is weight_only loading and the other is flash-attention generation." + "reason": "AutoImageProcessor URL loading and template-node compilation failure are unrelated." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "Documentation example errors vs long-OCR index error are unrelated." + "reason": "SigLIP2 doc-example errors and Kosmos2.5 long-OCR index error do not share the same underlying bug." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token registration and DAC latent/forward mismatch are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Duplicate of the earlier pair: sliding-window MoE behavior is not the same as trust_remote_code config loading." } ] }, @@ -5348,7 +5392,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:28218", "issue:29159", @@ -5379,14 +5423,13 @@ "issue:44610", "issue:44671", "issue:44805", - "issue:44821", "issue:44857", "issue:44936", + "issue:44977", "issue:45003", "issue:45020", "issue:45072", "issue:45081", - "issue:45084", "issue:45325", "issue:45372", "issue:45406", @@ -5400,238 +5443,241 @@ "issue:46326" ], "soft_pairs": [ - "issue:44821|issue:45084", + "issue:43782|issue:44977", "issue:45584|issue:45834", "issue:44186|issue:44442", "issue:43644|issue:43749", "issue:43606|issue:45072", "issue:44479|issue:45834", - "issue:34634|issue:36010", "issue:43825|issue:43994", - "issue:43577|issue:43582", "issue:44805|issue:44936", + "issue:34634|issue:36010", + "issue:43577|issue:43582", "issue:44485|issue:44509", "issue:44561|issue:45020", - "issue:28218|issue:29159", "issue:45987|issue:46192", - "issue:44610|issue:45072", + "issue:28218|issue:29159", "issue:43720|issue:45072", + "issue:44610|issue:45072", "issue:44671|issue:45701", "issue:43782|issue:45325", "issue:43825|issue:43827", + "issue:43906|issue:46209", "issue:43582|issue:44857", "issue:43749|issue:44291", - "issue:43906|issue:46209", "issue:45081|issue:45406", - "issue:45372|issue:46326", "issue:42175|issue:45081", + "issue:45372|issue:46326", "issue:43262|issue:43819", "issue:34634|issue:36331", "issue:45003|issue:46285" ], - "prepared_review_unit_hash": "782b20ec264f7c2ce3df317e8e6a57a8dbdea1bfda77123511e5b8b76683fb4b", + "prepared_review_unit_hash": "724dd2db3cc1034403828bc4e5e4ad17b0fdb1eb0fa923e567ab491ccb5865e9", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36213, - "estimated_input_tokens": 9054, - "estimated_eval_tokens": 18364 + "serialized_chars": 35678, + "estimated_input_tokens": 8920, + "estimated_eval_tokens": 18096 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is highly heterogeneous; the soft pairs mostly describe different bugs or unrelated regressions, so I would not merge any of them as duplicates. The strongest standalone issue here is the FSDP CPU/RAM-efficient loading regression.", - "confidence": 0.9, - "canonical_issue_reason": "Issue 43749 is the most concrete and high-signal report: it names a specific loading regression, has substantial discussion/inbound references, and is broader in impact than the other candidates.", + "summary": "This set is mostly heterogeneous, with only one plausible duplicate pair in the tokenizer behavior area. The strongest representative issue is the broad tokenization-regression report, while the rest are distinct bugs across unrelated subsystems.", + "confidence": 0.28, + "canonical_issue_reason": "issue:45701 is the best umbrella representative: it is broad, recent, open, and captures the recurring theme of version-driven tokenization changes better than the more specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "43749 is the best standalone issue to anchor the set because it is clearly scoped, actionable, and appears to affect an important loading path.", + "best_issue_reason": "issue:45701 is the most cluster-like issue overall; it has the broadest scope and the highest likelihood of subsuming adjacent tokenizer-regression reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44821", - "right": "issue:45084", + "left": "issue:43782", + "right": "issue:44977", "accept": false, - "reason": "Different failure modes and code paths: URL image processor loading vs a template compilation TypeError." + "reason": "Same model family, but different failure modes: loading with weight_only vs generation under flash-attention." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Unrelated models and symptoms: Whisper empty transcription handling vs Kosmos2.5 OCR index errors." + "reason": "Whisper empty transcription failure and Kosmos2.5 long-OCR index error are unrelated code paths." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "A tokenizer crash on NER/padding is not the same bug as AutoTokenizer failing to load a specific tokenizer class." + "reason": "Tokenizer crash on NER padding/truncation is not the same as AutoTokenizer failing to load a tokenizer class." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Both involve loading-related behavior, but the concrete bugs differ: non-persistent buffers junk initialization vs FSDP CPU RAM-efficient loading." + "reason": "Non-persistent buffer corruption and FSDP RAM-efficient loading are different bugs in different loading paths." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "These are distinct device/dtype mismatch reports affecting different models and execution paths." + "reason": "CPU-offload device mismatch and bfloat16 dtype mismatch are distinct failure modes." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "Different model families and regressions; one is video input handling, the other OCR long-input indexing." + "reason": "Video-input regression for Qwen VL models is unrelated to Kosmos2.5 long OCR indexing." }, { - "left": "issue:34634", - "right": "issue:36010", + "left": "issue:43825", + "right": "issue:43994", "accept": false, - "reason": "BarkProcessor voice_preset behavior and GenerationMixin import errors are unrelated." + "reason": "One is an error-message wording issue; the other is a model-output correctness bug." }, { - "left": "issue:43825", - "right": "issue:43994", + "left": "issue:44805", + "right": "issue:44936", "accept": false, - "reason": "One is an incorrect pipeline error message; the other is a model producing bad outputs through AutoModel/pipeline." + "reason": "Mask-shape IndexError during indexing is not the same as evaluate() failing after train()." }, { - "left": "issue:43577", - "right": "issue:43582", + "left": "issue:34634", + "right": "issue:36010", "accept": false, - "reason": "Mixed precision dtype loading for BLIP2 is unrelated to an Apple Silicon caching allocator TypeError." + "reason": "BarkProcessor voice_preset and GenerationMixin import failure are unrelated issues." }, { - "left": "issue:44805", - "right": "issue:44936", + "left": "issue:43577", + "right": "issue:43582", "accept": false, - "reason": "These describe different failures: tensor-mask shape mismatch vs trainer evaluate-after-train behavior." + "reason": "dtype propagation for BLIP2 and Apple Silicon warmup TypeError are separate bugs." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "RoPE implementation discussion is unrelated to docs still mentioning removed pipeline tasks." + "reason": "GLM-5 RoPE discussion and docs mentioning removed pipeline tasks are not the same underlying defect." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both touch remote_code breakage broadly, but they do not point to the same concrete API removal or code path." - }, - { - "left": "issue:28218", - "right": "issue:29159", - "accept": false, - "reason": "Both mention tokenizer behavior, but the reports are too generic and do not clearly identify the same underlying bug." + "reason": "A specific helper removal breaking trust_remote_code is not the same as the broader remote_code regression report." }, { "left": "issue:45987", "right": "issue:46192", "accept": false, - "reason": "StaticCache return type mismatch and return_tensors being ignored are separate API bugs." + "reason": "StaticCache seq-length contract breakage and return_tensors being ignored are unrelated API bugs." }, { - "left": "issue:44610", + "left": "issue:28218", + "right": "issue:29159", + "accept": true, + "reason": "Both point to tokenizer behavior divergence around added tokens and slow/fast tokenizer inconsistency, plausibly the same underlying whitespace bug." + }, + { + "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Processor image-size mismatch and bfloat16 dtype mismatches are different regressions." + "reason": "Packed-weight unpacking during accelerate loading and bfloat16 dtype mismatches are different problems." }, { - "left": "issue:43720", + "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Packed-weight loading failure in BitNet is a different issue from general dtype mismatch in inference." + "reason": "Processor output size mismatch and dtype mismatch are unrelated failure modes." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "Incorrect CamemBERT MLM predictions and version-dependent tokenization changes are not the same concrete bug." + "reason": "CamemBERT masked-LM prediction regression is not the same concrete bug as the generic tokenization-change report." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Different Qwen-VL model families and different defects: weight_only loading error vs RoPE position-id scaling bug." + "reason": "Qwen VL loading/weight_only error and RoPE temporal scaling are separate bugs." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "They are related to v5 pipeline removals, but one is runtime messaging and the other is stale documentation; not the same bug/change." + "reason": "They share the v5 pipeline-removal theme, but one is runtime messaging and the other is documentation; not the same concrete issue." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated repro of an older issue and an FP8 inference failure on B200 are unrelated." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Apple Silicon allocator TypeError and float16 CUDA loss crashes are unrelated." + "reason": "Apple Silicon allocator warmup and CUDA AMP loss crash are different hardware-specific problems." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "Both relate to loading, but the specific failure modes differ: FSDP RAM-efficient loading vs init_empty_weights argument incompatibility." - }, - { - "left": "issue:43906", - "right": "issue:46209", - "accept": false, - "reason": "An isolated reproduction of an older issue is not the same as a B200 FP8 inference failure." + "reason": "Both involve loading, but FSDP CPU RAM-efficient loading and init_empty_weights argument handling are distinct code paths." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Mistral tokenizer regex patching error and Gemma4 processor missing _tokenizer are different code paths." + "reason": "Mistral regex patch crash and Gemma4Processor missing _tokenizer are different tokenizer/processor bugs." }, { - "left": "issue:45372", - "right": "issue:46326", + "left": "issue:42175", + "right": "issue:45081", "accept": false, - "reason": "Both involve mistral_common, but one is an import-time dependency break and the other is a tokenizer behavior bug." + "reason": "Backend installation extras and tokenizer regex patching are unrelated." }, { - "left": "issue:42175", - "right": "issue:45081", + "left": "issue:45372", + "right": "issue:46326", "accept": false, - "reason": "Package backend installation omission is unrelated to tokenizer regex patch crashes." + "reason": "Missing ReasoningEffort import and continue_final_message behavior are different mistral_common-related defects." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio chat-template sampling-rate defaults and DAC latent decoding mismatch are unrelated." + "reason": "Sampling-rate defaulting and DAC latent/forward mismatch do not share the same bug." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor preset handling and CustomTrainer loss-signature breakage do not share the same underlying bug." + "reason": "BarkProcessor voice_preset and compute_loss signature mismatch are unrelated." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils is unrelated to the CLIP text-model structural inconsistency." + "reason": "Unsafe sys.modules access and CLIPTextModel structural inconsistency are different regressions." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it identifies one plausible tokenizer-related duplicate pair and treats the rest as distinct, which matches the titles. The representative issue choice is also supported by the broad scope of issue:45701. No soft-edge verdict appears to overstate the evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -5646,7 +5692,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:36010", "issue:36331", @@ -5663,7 +5709,6 @@ "issue:43550", "issue:43577", "issue:43638", - "issue:43643", "issue:43645", "issue:43650", "issue:43653", @@ -5694,26 +5739,28 @@ "issue:45459", "issue:45538", "issue:45600", - "issue:45759", - "issue:45803" + "issue:45803", + "issue:46421", + "issue:46424" ], "soft_pairs": [ - "issue:43653|issue:44912", "issue:43526|issue:43697", + "issue:43653|issue:44912", "issue:44060|issue:44368", "issue:43742|issue:45020", - "issue:43526|issue:44554", "issue:45084|issue:45538", - "issue:43329|issue:45600", + "issue:43526|issue:44554", "issue:43650|issue:43906", + "issue:43329|issue:45600", "issue:44589|issue:45042", - "issue:43746|issue:44811", "issue:43452|issue:44991", + "issue:43746|issue:44811", "issue:43638|issue:44936", - "issue:44387|issue:44617", "issue:43550|issue:44912", - "issue:43257|issue:43329", + "issue:44387|issue:44617", "issue:43819|issue:43927", + "issue:46421|issue:46424", + "issue:43257|issue:43329", "issue:43854|issue:44977", "issue:36010|issue:43645", "issue:43720|issue:44610", @@ -5721,222 +5768,224 @@ "issue:39692|issue:44912", "issue:36010|issue:36331", "issue:43531|issue:45042", + "issue:45459|issue:45803", "issue:43454|issue:43577", "issue:42490|issue:43257", - "issue:45459|issue:45803", - "issue:43257|issue:43696", "issue:43262|issue:44811", - "issue:43643|issue:45759" + "issue:43257|issue:43696" ], - "prepared_review_unit_hash": "4d72aae9b12985c1c563303c63e0e4a8561a01393f68c42d03b36b9d83d0340f", + "prepared_review_unit_hash": "782bb380ad668c4c6d035e480ae83294f8ee870085a3970ece9311e96d2c4f24", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36684, - "estimated_input_tokens": 9171, - "estimated_eval_tokens": 18598 + "serialized_chars": 36612, + "estimated_input_tokens": 9153, + "estimated_eval_tokens": 18562 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are unrelated model-loading, quantization, processor, trainer, and config bugs. None of the soft pairs look like the same concrete underlying issue, so all soft edges should be rejected.", - "confidence": 0.74, - "canonical_issue_reason": "issue:45020 is the broadest and most representative report in the set: it describes a general regression in recent transformers versions affecting remote_code models, with substantial detail and discussion. It is a better anchor than the narrower model-specific reports.", + "summary": "This cluster is mostly heterogeneous single-bug reports, not one broad duplicate set. The only strong duplicate-like pair is the Qwen tied-weights warning reports (44060/44368); the rest are too far apart in subsystem and failure mode to merge.", + "confidence": 0.72, + "canonical_issue_reason": "No single canonical issue cleanly covers the whole cluster; if forced, 44060 is the best anchor because it states the most concrete Qwen tied-weights warning bug and appears to match 44368.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is also the strongest standalone issue to keep as the cluster representative because it has the clearest broad impact and the most directly reusable framing for follow-up triage.", + "best_issue_reason": "44060 is the best representative issue because it is the clearest, most actionable description of the only likely duplicate set in the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43653", - "right": "issue:44912", + "left": "issue:43526", + "right": "issue:43697", "accept": false, - "reason": "Different subsystems and failure modes: BigBird tokenizer special-token decoding vs MXFP4 quantization falling back to bf16." + "reason": "Different models and failure modes: BEiT label reduction vs RT-DETR output drift." }, { - "left": "issue:43526", - "right": "issue:43697", + "left": "issue:43653", + "right": "issue:44912", "accept": false, - "reason": "Unrelated bugs: BEiT label reduction in an image processor vs RTDetrV2 output differences." + "reason": "Tokenizer special-token decode bug vs MXFP4 quantization fallback; unrelated code paths." }, { "left": "issue:44060", "right": "issue:44368", - "accept": false, - "reason": "Both mention tied-weights warnings, but they concern different Qwen models and different root causes; too broad to merge." + "accept": true, + "reason": "Both describe the same Qwen tied-weights warning regression; one is the core warning bug and the other is the same warning surfacing during Qwen3.5 LoRA fine-tuning." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "Both involve model loading, but one is a specific key error for MobileLLM while the other is a broad remote_code regression." + "reason": "MobileLLM key error is too specific to be the same bug as a broad remote_code-loading regression." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Compilation error on template nodes is unrelated to CLIPTokenizer model_max_length." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "No shared code path: BEiT processor label handling vs MPS attention correctness when value head dim differs." + "reason": "No shared underlying bug; image processor label reduction and MPS attention correctness are distinct." }, { - "left": "issue:45084", - "right": "issue:45538", + "left": "issue:43650", + "right": "issue:43906", "accept": false, - "reason": "Compile-time node/template error is unrelated to CLIPTokenizer max_length behavior." + "reason": "A vague request for data is not the same as an isolated reproduction of another issue." }, { "left": "issue:43329", "right": "issue:45600", "accept": false, - "reason": "Different areas entirely: multimodal token counting bug vs stale auto_mappings references in CI." + "reason": "Both touch multimodal/model-mapping areas, but they are different concrete bugs with different symptoms." }, { - "left": "issue:43650", - "right": "issue:43906", + "left": "issue:44589", + "right": "issue:45042", "accept": false, - "reason": "One is a nonspecific placeholder issue title; the other is an isolated reproduction of a separate upstream bug." + "reason": "Float8 storage lookup failure is unrelated to PIL backend torchvision requirements." }, { - "left": "issue:44589", - "right": "issue:45042", + "left": "issue:43452", + "right": "issue:44991", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to PIL image processors requiring torchvision." + "reason": "Both involve loading, but one is gguf/from_pretrained breakage and the other is tokenizer loading for a specific model." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT/local checkpoint loading for GraniteSpeech is unrelated to Whisper batch_decode skipping special tokens." + "reason": "Local PEFT checkpoint loading and Whisper batch_decode skip_special_tokens are unrelated." }, { - "left": "issue:43452", - "right": "issue:44991", + "left": "issue:43638", + "right": "issue:44936", "accept": false, - "reason": "Both are tokenizer/model loading complaints, but they target different mechanisms and model families; not the same bug." + "reason": "Zero3/non-pretrained BERT indexing and trainer.evaluate after train are different trainer lifecycle bugs." }, { - "left": "issue:43638", - "right": "issue:44936", + "left": "issue:43550", + "right": "issue:44912", "accept": false, - "reason": "Deepspeed ZeRO3 index error on a non-pretrained BERT model is not the same as trainer.evaluate() failing after train()." + "reason": "torch.compile+SDPA on Bamba is not the same problem as MXFP4 quantization loading." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "General int4 quantization memory regression is not the same as a specific Sam3Video OOM report." + "reason": "Both mention OOM, but one is int4 reserved-memory regression and the other is a model-specific memory exhaustion report." }, { - "left": "issue:43550", - "right": "issue:44912", + "left": "issue:43819", + "right": "issue:43927", "accept": false, - "reason": "torch.compile + SDPA failure in Bamba is unrelated to MXFP4 loading falling back to bf16." + "reason": "DAC from_latents/STE mismatch and DiaConfig token ID persistence are unrelated." }, { - "left": "issue:43257", - "right": "issue:43329", + "left": "issue:46421", + "right": "issue:46424", "accept": false, - "reason": "Qwen3 MoE weight conversion/loading bug is unrelated to multimodal token counting in video processors." + "reason": "Both are cache-related, but one is recompilation from lazy init and the other is max_cache_length semantics; not the same bug." }, { - "left": "issue:43819", - "right": "issue:43927", + "left": "issue:43257", + "right": "issue:43329", "accept": false, - "reason": "DAC latent/STE mismatch is a modeling correctness issue, not the same as DiaConfig losing custom token IDs on save/load." + "reason": "Qwen3 MoE weight conversion and sliding_window behavior are different Qwen3 issues." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Model-loading test failure for GLM-4.7-Flash is unrelated to flash-attention generation issues in Qwen3.5." + "reason": "GLM-4.7-Flash load failure and Qwen3.5 flash-attention generation problems are unrelated." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "Importing GenerationMixin from a moved module is different from notebook-defined custom model initialization breaking in Transformers 5.0." + "reason": "Importing GenerationMixin and Jupyter custom-model initialization are different regressions." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Packed-weight loading for BitNet is unrelated to OmDet-Turbo processor/model input-size mismatch." + "reason": "Packed-weight loading in BitNet and OmDet-Turbo processor size mismatch are unrelated." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 zero probabilities and OmDet-Turbo resolution mismatch are different model-specific bugs." + "reason": "SigLIP2 zero probabilities and OmDet-Turbo input-size mismatch are unrelated." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Documentation example errors for SigLIP2 are not the same as mxfp4 quantization fallback behavior." + "reason": "Documentation example errors and MXFP4 quantization loading are not the same underlying issue." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "ImportError for GenerationMixin and compute_loss signature change are unrelated trainer/API issues." + "reason": "ImportError for GenerationMixin vs unexpected compute_loss keyword are unrelated training/loading bugs." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding window behavior has nothing to do with PIL backend torchvision requirements." + "reason": "Qwen3-MoE sliding_window and PIL backend torchvision requirements are different code paths." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both involve exception handling, but they affect different components and different failure modes." }, { "left": "issue:43454", "right": "issue:43577", "accept": false, - "reason": "AyaVision tied-weight generation bug and Blip2 dtype propagation bug are distinct model initialization problems." + "reason": "AyaVision weight tying and Blip2 dtype handling are separate model-loading defects." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Trailing-slash local path warning is unrelated to Qwen3 MoE accelerate/deepspeed conversion." - }, - { - "left": "issue:45459", - "right": "issue:45803", - "accept": false, - "reason": "Both concern exception handling, but they are different code paths and not the same concrete bug." - }, - { - "left": "issue:43257", - "right": "issue:43696", - "accept": false, - "reason": "Qwen3 MoE weight conversion/loading is unrelated to GPT-oss-20b CUDA OOM." + "reason": "Trailing-slash path warning and Qwen3 MoE weight conversion are unrelated." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sample-rate defaulting is unrelated to Whisper batch_decode skip_special_tokens handling." + "reason": "Audio apply_chat_template sample-rate default and Whisper batch_decode behavior are not the same bug." }, { - "left": "issue:43643", - "right": "issue:45759", + "left": "issue:43257", + "right": "issue:43696", "accept": false, - "reason": "Both are config-loading regressions, but one is about trust_remote_code missing fields and the other about from_config not unwrapping text_config." + "reason": "Qwen3 MoE conversion and GPT-oss OOM are unrelated." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The clustering is grounded in the packet and remains conservative. The single accepted soft edge (44060/44368) is plausible from the titles alone, and the rest are appropriately rejected as cross-subsystem or cross-failure-mode pairs. The summary does not overclaim a broader duplicate set, and the canonical/best-issue choices are presented cautiously rather than as certainty." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -5951,11 +6000,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:24643", "issue:28282", - "issue:29127", "issue:29159", "issue:33357", "issue:34567", @@ -5972,6 +6020,7 @@ "issue:43540", "issue:43582", "issue:43606", + "issue:43643", "issue:43697", "issue:43742", "issue:43749", @@ -5998,10 +6047,12 @@ "issue:45488", "issue:45538", "issue:45701", + "issue:45759", "issue:45859", "issue:45907" ], "soft_pairs": [ + "issue:43643|issue:45759", "issue:43329|issue:44479", "issue:43329|issue:43450", "issue:43606|issue:45859", @@ -6013,12 +6064,12 @@ "issue:42915|issue:43742", "issue:44291|issue:45042", "issue:42915|issue:44821", - "issue:44849|issue:45701", - "issue:43452|issue:43582", "issue:44987|issue:45538", - "issue:43927|issue:44912", + "issue:43452|issue:43582", "issue:43697|issue:45357", + "issue:43927|issue:44912", "issue:42915|issue:45538", + "issue:44849|issue:45701", "issue:44568|issue:45488", "issue:41720|issue:43540", "issue:33357|issue:34567", @@ -6027,240 +6078,239 @@ "issue:43317|issue:45538", "issue:43540|issue:44479", "issue:44062|issue:44991", + "issue:43540|issue:44560", "issue:28282|issue:34689", "issue:45137|issue:45405", - "issue:43540|issue:44560", "issue:24643|issue:29159", "issue:43742|issue:44291", - "issue:44987|issue:45907", - "issue:28282|issue:29127" + "issue:44987|issue:45907" ], - "prepared_review_unit_hash": "28559efeea11e5e7a84f10cf0b5d932221cc997ab8780abfbbd50d6e9636a645", + "prepared_review_unit_hash": "0a5399bd818fa84c0ad5cca7a31f1a8efaa98e203c491883f1ab81843f89335c", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36721, - "estimated_input_tokens": 9181, - "estimated_eval_tokens": 18618 + "serialized_chars": 37325, + "estimated_input_tokens": 9332, + "estimated_eval_tokens": 18920 }, "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 bugs. Only one soft pair looks plausibly duplicate-level: the Qwen video-input regression pair (44479/45381).", - "confidence": 0.58, - "canonical_issue_reason": "issue:45488 is the clearest actionable open issue here: it describes a concrete tokenizer-regression bug with a clear failure mode and broad downstream impact.", + "summary": "Most items are unrelated standalone issues; there is one likely duplicate cluster around the Qwen v5.3.0 video-input regression. No PRs are present.", + "confidence": 0.68, + "canonical_issue_reason": "issue:44479 is the broadest and most representative report of the Qwen video-input regression, naming the release and multiple affected model families.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45488 is the best single issue to anchor duplicates against: specific enough to reproduce, broad enough to represent the tokenizer regression family, and still open.", + "best_issue_reason": "issue:44479 is the best standalone issue to anchor the cluster because it captures the shared regression surface more generally than the model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Different bugs: missing fields with trust_remote_code vs composite Qwen text_config unwrapping." + }, { "left": "issue:43329", "right": "issue:44479", - "accept": false, - "reason": "Both are video-path Qwen bugs, but one is an undefined-name bug in token counting and the other is a video-input regression; different concrete failures." + "accept": true, + "reason": "Same underlying video-input regression in the Qwen VL path; 43329 appears to pinpoint the root helper bug behind the broader 5.3.0 regression." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "Different bugs in the video pipeline: undefined helpers in multimodal token counting vs incorrect batched video processor shape." + "reason": "One is an undefined-variable bug in token counting; the other is a batched shape issue in video processors." }, { "left": "issue:43606", "right": "issue:45859", "accept": false, - "reason": "Both mention CPU offload, but they affect different models and different code paths; not the same bug." + "reason": "Both mention CPU offload, but they affect different models and code paths." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "Unrelated problems: TrainerState token counting vs a perplexity issue on gpt-oss-20b." + "reason": "Trainer token-count tracking is unrelated to perplexity on gpt-oss-20b." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Both are memory/loading related, but FSDP RAM-efficient loading and Strix Halo mmap OOM are distinct issues." + "reason": "Both concern loading/memory, but one is FSDP RAM-efficient loading and the other is mmap OOM on Strix Halo." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Different model-loading failures with different symptoms and causes: weight conversion under accelerate+deepspeed vs a shape mismatch on Qwen3-VL." + "reason": "Both are Qwen loading problems, but one is accelerate/deepspeed weight conversion and the other is a Qwen3-VL shape mismatch." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Separate tokenizer regressions: special-token insertion for mdeberta vs codec/warning behavior for Kimi-K2.5." + "reason": "Distinct tokenizer regressions affecting different models and different behaviors." }, { "left": "issue:29159", "right": "issue:34567", "accept": false, - "reason": "Tokenizer consistency issue vs TrainerState token counter bug; not the same underlying problem." + "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState token counting." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Different load-time failures: Qwen3Moe FP8 config failure vs MobileLLM key error." + "reason": "FP8 fine-grained config failure is unrelated to MobileLLM key errors." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "Completely different subsystems: init_empty_weights argument handling vs PIL image processor torchvision dependency." + "reason": "Init-empty-weights argument handling is unrelated to PIL/torchvision backend requirements." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "No shared bug path; one is Qwen3Moe FP8 loading, the other is AutoImageProcessor-from-URL loading." + "reason": "Different areas: Qwen3Moe FP8 vs AutoImageProcessor URL loading." }, { - "left": "issue:44849", - "right": "issue:45701", + "left": "issue:44987", + "right": "issue:45538", "accept": false, - "reason": "Qwen3.5 hidden-state output bug is unrelated to the broad tokenization-change report." + "reason": "Model-loading failure for one repo is not the same as CLIPTokenizer max-length bug." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "GGUF tokenizer/model loading breakage and Apple Silicon allocator warmup TypeError are unrelated." + "reason": "gguf loading issue is unrelated to AppleSilicon allocator warmup TypeError." }, { - "left": "issue:44987", - "right": "issue:45538", + "left": "issue:43697", + "right": "issue:45357", "accept": false, - "reason": "Different tokenizer issues: loading a specific model fails in 5.1.0+, while CLIPTokenizer has an absurd max-length value." + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained key regression are different bugs." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "DiaConfig save/load token-ID loss and MXFP4 quantization fallback are different bugs." + "reason": "Config save/load token IDs issue is unrelated to MXFP4 quantization fallback." }, { - "left": "issue:43697", - "right": "issue:45357", + "left": "issue:42915", + "right": "issue:45538", "accept": false, - "reason": "Different model families and failures: RTDetrV2 output drift vs Qwen3.5 save_pretrained key mapping." + "reason": "Qwen3Moe FP8 and CLIPTokenizer model_max_length are unrelated." }, { - "left": "issue:42915", - "right": "issue:45538", + "left": "issue:44849", + "right": "issue:45701", "accept": false, - "reason": "No overlap beyond both touching model/tokenizer behavior; the bugs are unrelated." + "reason": "Qwen3.5 hidden-states behavior and generic tokenization changes are not the same bug." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer regressions, but one is special-token insertion and the other is tokenizer.json pre-tokenizer override; distinct code paths." + "reason": "Different tokenizer regressions: special-token insertion vs LlamaTokenizer pre-tokenizer override." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "Different Qwen issues: device-mapping CUDA assert vs video-input ValueError." + "reason": "Different symptoms and models: CUDA assert on A800 vs video ValueError in Qwen3OmniMoe." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS bus error in CLIP model loading is unrelated to TrainerState token tracking." + "reason": "MacOS CLIP bus error is unrelated to TrainerState token accounting." }, { "left": "issue:44479", "right": "issue:45381", "accept": true, - "reason": "Same video-input regression family for Qwen VL models on 5.3.0; 45381 is a specific symptom of the broader bug described in 44479." + "reason": "Same Qwen video-input regression family; 45381 is a model-specific report within the broader 44479 regression." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "TensorFlow backend packaging issue vs import error for a specific model class; not the same bug." + "reason": "TensorFlow backend packaging issue is unrelated to Qwen2.5-VL import failure." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "Dequantized model offload failure is unrelated to CLIPTokenizer max_length behavior." + "reason": "Device-map dequantized-model offload failure is unrelated to CLIPTokenizer max-length behavior." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Both involve Qwen video processing, but one is a generic processing ValueError and the other is a version-specific video regression; not enough evidence of the same bug." + "reason": "Both are video-related, but Qwen3OmniMoe ValueError is not clearly the same bug as the multi-model 5.3.0 regression." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "Tokenizer constructor TypeError and est-roberta tokenizer loading failure are different issues." + "reason": "AddedToken special-argument conflict is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Different video-input failures in different Qwen models; not enough evidence they share one bug." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "General PyTorch import error vs Llama 3.2 Vision loading regression; unrelated." + "reason": "PyTorch-missing import error is unrelated to Llama 3.2 model loading breakage." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "DeepSpeed ZeRO3 deque IndexError is unrelated to the PEFT version bump issue." - }, - { - "left": "issue:43540", - "right": "issue:44560", - "accept": false, - "reason": "Both are Qwen video issues, but the failure modes differ: a processing ValueError vs a StopIteration in embedding/video handling." + "reason": "DeepSpeed ZeRO3 deque error is unrelated to the PEFT version bump issue." }, { "left": "issue:24643", "right": "issue:29159", "accept": false, - "reason": "DeepSpeed weight-dimension runtime error and tokenizer slow/fast inconsistency are unrelated." + "reason": "DeepSpeed training weight-shape error is unrelated to tokenizer slow/fast inconsistency." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "MobileLLM key error and unexpected _is_hf_initialized argument are different model-loading bugs." + "reason": "Both are loading-related, but one is a key error for a model and the other is an init_empty_weights argument regression." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Specific model-loading regression and list_repo_files revision-forwarding bug are unrelated." - }, - { - "left": "issue:28282", - "right": "issue:29127", - "accept": false, - "reason": "Missing PyTorch import and LayoutLMv3 box-content error are unrelated." + "reason": "Repo loading failure and list_repo_files revision forwarding are different Hugging Face Hub issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall and conservative in the soft-edge calls. The only accepted pair, 44479/45381, is plausibly duplicate-level from the packet; the rest are reasonably rejected as distinct failures. The canonical issue choice is also supported by the titles and open state." + "feedback": "Grounded overall. The summary stays conservative, correctly centers issue:44479 as the broad Qwen video-input regression, and the accepted soft edges are plausible from the titles alone. No PRs are present in the packet." }, "evaluator_used": true, "retried": false, @@ -6279,8 +6329,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:27957", + "issue:28282", + "issue:29127", "issue:34567", "issue:36331", "issue:38175", @@ -6305,11 +6358,9 @@ "issue:44291", "issue:44488", "issue:44589", - "issue:44625", "issue:44737", "issue:44792", "issue:44821", - "issue:44877", "issue:44898", "issue:44913", "issue:44964", @@ -6331,22 +6382,24 @@ ], "soft_pairs": [ "issue:43577|issue:45072", + "issue:28282|issue:29127", "issue:42915|issue:43749", - "issue:43257|issue:43756", "issue:38175|issue:44898", "issue:42175|issue:44488", + "issue:43257|issue:43756", "issue:43582|issue:43782", - "issue:43493|issue:43994", "issue:44737|issue:45588", + "issue:43493|issue:43994", "issue:34567|issue:36331", + "issue:44291|issue:44821", "issue:43404|issue:45987", "issue:43526|issue:45479", - "issue:44291|issue:44821", "issue:43756|issue:46399", "issue:44062|issue:45081", "issue:43906|issue:45381", - "issue:44913|issue:46121", + "issue:27957|issue:28282", "issue:43756|issue:45325", + "issue:44913|issue:46121", "issue:44792|issue:44964", "issue:45042|issue:45081", "issue:43452|issue:44488", @@ -6354,199 +6407,198 @@ "issue:45357|issue:45920", "issue:43531|issue:44821", "issue:45405|issue:45600", - "issue:43643|issue:45375", - "issue:44625|issue:44877" + "issue:43643|issue:45375" ], - "prepared_review_unit_hash": "569bcee245ee918ff2745768f38564cd9ce830c0a6c9906ed124b795bb5a5a71", + "prepared_review_unit_hash": "c09a82e0820529e2c6e04409f8c52e95ecb0b4edef0a0dc6cb8c839c1ec42536", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 35279, - "estimated_input_tokens": 8820, - "estimated_eval_tokens": 17896 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36185, + "estimated_input_tokens": 9047, + "estimated_eval_tokens": 18350 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly heterogeneous: many titles are superficially similar because they involve Transformers configs, loading, or dtype issues, but they do not look like the same underlying bug. Only the SigLIP2 pair stands out as a plausible duplicate/root-cause match.", - "confidence": 0.67, - "canonical_issue_reason": "issue:43493 is the strongest root-cause style issue here: it describes the underlying SigLIP2 correctness discrepancy, and it best matches the only convincing duplicate-like pair in the set (issue:43994).", + "summary": "This cluster is mostly heterogeneous. The only strong duplicate candidate is the SigLIP2 pair, where one issue describes the underlying HF-vs-JAX implementation mismatch and the other reports the same bad outputs through AutoModel/pipeline. No PRs are present.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43493 is the best canonical issue because it points to the underlying SigLIP2 implementation discrepancy; issue:43994 appears to be the same bug surfacing as nonsensical inference output.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43493 is the best canonical issue candidate because it is the broadest and most upstream explanation for the SigLIP2 failure reports; the other issues are too model-specific or unrelated.", + "best_issue_reason": "issue:43493 is the strongest representative: it is root-cause oriented, specific to the affected model, and can subsume the user-facing symptoms in issue:43994.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Both are dtype-related, but they affect different model families and different code paths (BLIP2 loading vs SwitchTransformers/TimmWrapper inference)." + "reason": "Both mention dtype mismatches, but they affect different model families and code paths (Blip2/QFormer vs SwitchTransformers/TimmWrapperModel)." }, { - "left": "issue:42915", - "right": "issue:43749", + "left": "issue:28282", + "right": "issue:29127", "accept": false, - "reason": "Different failures: FineGrainedFP8Config on Qwen3Moe vs FSDP CPU RAM efficient loading. Similar training/loading theme only." + "reason": "Unrelated problems: one is a missing-PyTorch import error, the other is a layoutlmv3 validation/error-message clarity issue." }, { - "left": "issue:43257", - "right": "issue:43756", + "left": "issue:42915", + "right": "issue:43749", "accept": false, - "reason": "Different bugs: Qwen3 MOE weight conversion with accelerate+deepspeed vs Smollm3 RoPE-layer retention." + "reason": "Different failures in different loading/training mechanisms; both involve advanced memory/loading settings but not the same defect." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Both concern model behavior, but SigLIP2 zero probabilities and Perceiver image classification are unrelated models and bugs." + "reason": "Same broad model area (vision), but the symptoms and mechanics differ: zero probabilities vs non-default resolution/interpolation failure." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "Backend packaging/install issue vs a specific model-loading failure; no shared underlying bug." + "reason": "One is about backend dependency packaging, the other about loading a specific model; no shared concrete bug." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Both are Qwen/SmolLM-related, but one is a MoE weight conversion/loading issue and the other is a RoPE layer-count/configuration bug." }, { "left": "issue:43582", "right": "issue:43782", "accept": false, - "reason": "Different areas and symptoms: AppleSilicon allocator warmup TypeError vs Qwen3VL weight_only load error." + "reason": "Different subsystems: Apple Silicon allocator warmup TypeError vs Qwen3VL weight_only loading failure." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "One is a CPU-device placement bug in XLNet; the other is a flash-attention sink/s_aux crash. Not the same code path." }, { "left": "issue:43493", "right": "issue:43994", "accept": true, - "reason": "Both point to the same SigLIP2 correctness problem: bad outputs/discrepancy from the expected implementation for the same model family." + "reason": "Both are about SigLIP2 producing incorrect outputs; 43493 frames the implementation discrepancy, while 43994 reports the same underlying bug through AutoModel/pipeline usage." }, { - "left": "issue:44737", - "right": "issue:45588", + "left": "issue:34567", + "right": "issue:36331", "accept": false, - "reason": "XLNet CPU tensor allocation bug is unrelated to flash_attention s_aux=None handling." + "reason": "Both concern trainer behavior, but one is a state counter not updating and the other is an unexpected argument in custom compute_loss." }, { - "left": "issue:34567", - "right": "issue:36331", + "left": "issue:44291", + "right": "issue:44821", "accept": false, - "reason": "Both are training-loop issues, but one is token counting state and the other is a changed compute_loss signature." + "reason": "Different loading failures: unexpected `_is_hf_initialized` argument versus inability to load AutoImageProcessor from URL." }, { "left": "issue:43404", "right": "issue:45987", "accept": false, - "reason": "Different model/runtime problems: tied weights in Mistral3 vs StaticCache return type mismatch." + "reason": "A tied-weight bug in Mistral3 is unrelated to StaticCache returning a tensor instead of an int." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "BEiT label reduction bug and single-label classification zero-loss bug are distinct." - }, - { - "left": "issue:44291", - "right": "issue:44821", - "accept": false, - "reason": "init_empty_weights/_is_hf_initialized argument handling is unrelated to loading AutoImageProcessor from URL." + "reason": "Both are classification/label-related, but one is an image processor label-reduction bug and the other is a degenerate loss issue for single-label classification." }, { "left": "issue:43756", "right": "issue:46399", "accept": false, - "reason": "Both mention RoPE, but one is Smollm3 layer dropping and the other is a Qwen3.5 kernel registration crash." + "reason": "Different rotary-related areas: one is a model architecture/config mismatch, the other is a kernel registration crash." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Tokenizer constructor duplication error and Mistral regex patch backend_tokenizer AttributeError are separate tokenizer bugs." + "reason": "Tokenizer construction failure vs Mistral regex patching backend attribute crash; they are separate bugs." }, { "left": "issue:43906", "right": "issue:45381", "accept": false, - "reason": "A reproduction of issue 38071 is not enough to treat this as the same bug as the Qwen2.5-VL position-id regression." + "reason": "One is just a reproduction of an older issue; the other is a specific Qwen2.5-VL position-id regression. Insufficient evidence they are the same bug." }, { - "left": "issue:44913", - "right": "issue:46121", + "left": "issue:27957", + "right": "issue:28282", "accept": false, - "reason": "One is config reload losing rotary_pct; the other is a TypeError in rope-parameter conversion from JSON-loaded lists." + "reason": "XLMRoberta Flash Attention support and a missing-PyTorch ImportError are unrelated." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Different Qwen/Smollm3 RoPE-related issues with different symptoms and code paths." + "reason": "Both touch positional/rope logic in multimodal models, but they target different models and failure modes." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Different config serialization issues: rotary_pct reload regression vs list-handling TypeError in rope validation." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Janus image-generation test failure and Phi-4 multimodal loading failure are not the same defect." + "reason": "A failed janus test and a Phi-4 multimodal load failure are distinct symptoms on different models." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "PIL backend torchvision dependency regression is unrelated to the Mistral tokenizer regex crash." + "reason": "Dependency requirement regression in image processors vs a tokenizer regex patch crash; not the same defect." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "gguf_file handling for tokenizer/model loading is a different issue from the cjvt/sleng-bert model-load failure." + "reason": "GGUF loading breaks in tokenizers/CausalLM paths, while cjvt/sleng-bert is a separate model-loading issue." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "MobileLLM key mismatch and missing Float8 storage object are separate loading/runtime errors." + "reason": "KeyError loading MobileLLM-125M is unrelated to the Float8 storage-object TypeError." }, { "left": "issue:45357", "right": "issue:45920", "accept": false, - "reason": "Qwen3.5 save_pretrained visual-key regression and AutoTokenizer wrong token IDs are unrelated." + "reason": "One is a save_pretrained visual-key regression for Qwen3.5; the other is a tokenizer ID regression across many models." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." + "reason": "Qwen3-MoE sliding-window behavior is unrelated to AutoImageProcessor URL loading." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "Version pinning to unreleased PEFT and stale auto_mappings references are different repository maintenance issues." - }, - { - "left": "issue:43643", - "right": "issue:45375", - "accept": false, - "reason": "Both are config-field preservation issues, but they involve different configs and different failure modes." - }, - { - "left": "issue:44625", - "right": "issue:44877", - "accept": false, - "reason": "num_labels propagation in Qwen3.5 is unrelated to strict config blocking granite_speech loading." + "reason": "One is a version pin bump for PEFT; the other is a broken auto_mappings reference to removed configs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The assessment is grounded in the packet: most pairs are conservatively rejected as only superficially similar, and the SigLIP2 pair (43493/43994) is a plausible duplicate/root-cause match based on the titles and summaries. No clear overstatement or mergeability issue is evident." + "feedback": "The summary is well grounded in the packet: it stays conservative about the cluster being mostly heterogeneous and only treats issue:43493/issue:43994 as the plausible duplicate pair. The soft-edge rejections are cautious and consistent with the titles/body signals, and no PR-pair mergeability concerns apply because there are no PRs." }, "evaluator_used": true, "retried": false, @@ -6565,17 +6617,16 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:24643", + "issue:27957", "issue:28282", "issue:29127", - "issue:29159", "issue:30064", "issue:32090", "issue:33357", "issue:33666", - "issue:35141", "issue:38175", "issue:41762", "issue:43295", @@ -6596,13 +6647,14 @@ "issue:44493", "issue:44514", "issue:44568", + "issue:44625", "issue:44704", "issue:44743", "issue:44811", "issue:44843", + "issue:44877", "issue:44898", "issue:44936", - "issue:44991", "issue:45005", "issue:45216", "issue:45290", @@ -6617,239 +6669,238 @@ "issue:46202" ], "soft_pairs": [ - "issue:24643|issue:29127", + "issue:44625|issue:44877", "issue:43720|issue:44898", "issue:43697|issue:45216", + "issue:24643|issue:29127", "issue:41762|issue:44936", - "issue:44493|issue:46050", "issue:44514|issue:45381", + "issue:44493|issue:46050", "issue:44466|issue:45005", "issue:43452|issue:43577", "issue:43906|issue:46202", "issue:43452|issue:44843", - "issue:43688|issue:43749", - "issue:28282|issue:32090", "issue:30064|issue:33666", - "issue:28282|issue:33357", + "issue:43688|issue:43749", "issue:43819|issue:44811", "issue:45372|issue:46170", + "issue:28282|issue:32090", "issue:44936|issue:45823", "issue:43295|issue:45356", - "issue:28282|issue:35141", - "issue:43653|issue:44361", + "issue:28282|issue:33357", "issue:44361|issue:44442", + "issue:43653|issue:44361", "issue:44484|issue:44704", "issue:38175|issue:43295", "issue:45542|issue:46178", - "issue:44464|issue:44743", + "issue:24643|issue:27957", "issue:43653|issue:44568", "issue:43577|issue:43697", "issue:44514|issue:45290", - "issue:29159|issue:33357", - "issue:44843|issue:44991" + "issue:44464|issue:44743" ], - "prepared_review_unit_hash": "73d8f2f1b4377ed45bdd725abe1147cf1c977c36f388e9feb160d8c742a17c28", + "prepared_review_unit_hash": "255ddaee4f7c666691da35767c5354b42d633564bcf41a37e8adc4b6d66bb15a", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 37079, - "estimated_input_tokens": 9270, - "estimated_eval_tokens": 18796 + "soft_pair_count": 29, + "serialized_chars": 36684, + "estimated_input_tokens": 9171, + "estimated_eval_tokens": 18598 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only one soft pair looks like the same underlying bug: the tied-weights serialization regression in v5. The rest are either different models/code paths or just broadly related subsystem reports.", - "confidence": 0.74, - "canonical_issue_reason": "issue:45005 is the best canonical issue because it states the broader tied-weights regression, while issue:44466 is a narrower device-dependent manifestation of the same save_pretrained serialization problem.", + "summary": "Mostly a mixed cluster of unrelated Transformers issues across model loading, tokenizers, processors, training, and quantization. Only one soft pair looks like the same underlying regression: the v5 tied-weights/save_pretrained serialization bug.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 45005 is the better canonical because it states the broader tied-weights regression in v5, while 44466 reads like a specific manifestation of that same save/serialization problem.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45005 is the strongest global issue candidate for the tied-weights save/load regression cluster; it is broader and more representative than the device-specific report in issue:44466.", + "best_issue_reason": "45005 is the most representative issue in the only likely duplicate pair: it captures the shared root cause (tied weights breaking in v5) rather than one device-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:24643", - "right": "issue:29127", + "left": "issue:44625", + "right": "issue:44877", "accept": false, - "reason": "Different failures: DeepSpeed training shape error vs LayoutLMV3 box-validation messaging." + "reason": "Different bugs: Qwen3.5 label propagation vs granite_speech strict-config loading." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "Different code paths and models: BitNet accelerate-loading unpacking vs Perceiver interpolation at non-default resolution." + "reason": "Unrelated code paths: BitNet packed-weight loading vs Perceiver image classification interpolation." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "One is RTDetrV2 output drift in v5; the other is a Qwen3.5 save_pretrained regression." + "reason": "Different regressions in different models: RTDetrV2 output changes vs Qwen3.5 checkpoint save_pretrained corruption." }, { - "left": "issue:41762", - "right": "issue:44936", + "left": "issue:24643", + "right": "issue:29127", "accept": false, - "reason": "Different problems: Gemma3 ZeRO-3 load failure vs trainer.evaluate() breaking after train()." + "reason": "DeepSpeed weight-shape crash and LayoutLMv3 error-message clarity are not the same bug." }, { - "left": "issue:44493", - "right": "issue:46050", + "left": "issue:41762", + "right": "issue:44936", "accept": false, - "reason": "Unexpected position-id keys and quantized sinq save/load corruption are not the same bug." + "reason": "Both involve training/loading, but one is Gemma3 ZeRO-3 loading and the other is trainer state after train()\u2014distinct failures." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Both involve Qwen2.5-VL, but the crashes are different: batched padding handling vs wrong vision_position_ids in video input." + "reason": "Same family (Qwen2.5-VL) but different problems: chat-template batching vs video vision_position_ids." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Position-id warnings and sinq save/from_pretrained breakage are different serialization/config issues." }, { "left": "issue:44466", "right": "issue:45005", "accept": true, - "reason": "Both describe the same tied-weights / lm_head serialization regression in v5 save_pretrained, with 45005 as the broader report." + "reason": "Both describe the same tied-weights/v5 serialization regression; 44466 is a concrete device-dependent symptom of the broader 45005 issue." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "gguf_file loading failure and BLIP2 dtype propagation are unrelated." + "reason": "Tokenizer/model-from-pretrained GGUF breakage is unrelated to BLIP2 dtype remaining float32." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "43906 is explicitly a reproduction of another issue; 46202 is a separate Gemma roundtrip report, but the titles alone do not prove the same concrete bug." + "reason": "A reproduction of one issue and a Gemma pipeline roundtrip tokenization bug are not the same underlying defect." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "Different mechanisms: gguf loading vs unconditional model_info() access breaking offline mode." - }, - { - "left": "issue:43688", - "right": "issue:43749", - "accept": false, - "reason": "Auxiliary-loss normalization and FSDP CPU RAM-efficient loading are unrelated." - }, - { - "left": "issue:28282", - "right": "issue:32090", - "accept": false, - "reason": "Missing PyTorch import is unrelated to the Trainer _gpu_broadcast_one NoneType error." + "reason": "Different tokenizer loading failures: GGUF-related breakage vs offline model_info() call in mistral regex patching." }, { "left": "issue:30064", "right": "issue:33666", "accept": false, - "reason": "Void segmentation-map processing and Qwen2-VL multi-GPU training are different bugs." + "reason": "Image-segmentation map handling and Qwen2-VL multi-GPU training are unrelated." }, { - "left": "issue:28282", - "right": "issue:33357", + "left": "issue:43688", + "right": "issue:43749", "accept": false, - "reason": "Dependency import error vs MacOS bus error in a CLIP model; not the same issue." + "reason": "Aux-loss normalization and FSDP CPU RAM-efficient loading affect different subsystems and code paths." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "DAC latent decoding mismatch is unrelated to Whisper batch_decode skip_special_tokens being ignored." + "reason": "DAC latent reconstruction mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:45372", "right": "issue:46170", "accept": false, - "reason": "Gemma 4 processor import failure from mistral_common is unrelated to Windows ROCm torch.distributed dependency issues." + "reason": "Gemma 4 processor import failure and Windows ROCm torch.distributed issues are different environment/import problems." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Missing PyTorch import error and Trainer _gpu_broadcast_one NoneType failure are unrelated." }, { "left": "issue:44936", "right": "issue:45823", "accept": false, - "reason": "Trainer lifecycle failure and Gemma4 device_map='auto' mismatch are different code paths." + "reason": "Trainer evaluate-after-train failure is not the same as Gemma4 device_map=\"auto\" forward device mismatch." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Processor.tokenizer regression and Kimi-K2.5 codec/mistral-regex regression are distinct." + "reason": "Both are tokenizer regressions, but one is processor/tokenizer API breakage and the other is Kimi-K2.5 codec/fix_mistral_regex behavior." }, { - "left": "issue:44464", - "right": "issue:44743", + "left": "issue:44361", + "right": "issue:44442", "accept": false, - "reason": "Chunked generation with compiled forward and recurrent-state reset in modular_qwen3_5 are not the same concrete bug." + "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer loading failure are different tokenizer implementations." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBird mask-token registration/empty decode and MLukeTokenizer task AttributeError are unrelated tokenizer bugs." - }, - { - "left": "issue:44361", - "right": "issue:44442", - "accept": false, - "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer loading failure are different issues." + "reason": "BigBird special-token registration/empty decode is unrelated to MLukeTokenizer task errors." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "Max shard size documentation/question and AutoProcessor cached_file kwargs forwarding are unrelated." + "reason": "save_pretrained shard-size behavior and AutoProcessor cached_file kwargs are different issues." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "Zero probabilities in SigLIP2 and processor.tokenizer regression are unrelated." + "reason": "SigLIP zero probabilities and processor.tokenizer regression are unrelated." }, { "left": "issue:45542", "right": "issue:46178", "accept": false, - "reason": "Both are backend/dependency complaints, but one is about TensorFlow/tensorboard fallback and the other about tensorflow_text not being in BACKENDS_MAPPING." + "reason": "TensorBoard-without-TensorFlow backend error and tensorflow_text backend mapping error are distinct backend-registration problems." + }, + { + "left": "issue:24643", + "right": "issue:27957", + "accept": false, + "reason": "DeepSpeed training weight-shape error and XLMRoberta Flash Attention 2 are different failures." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "BigBird special-token registration and mdeberta BOS/EOS handling are different tokenizer bugs." + "reason": "BigBird mask-token special-token bug is unrelated to mDeBERTa BOS/EOS add_special_tokens regression." }, { - "left": "issue:44514", - "right": "issue:45290", + "left": "issue:43577", + "right": "issue:43697", "accept": false, - "reason": "Both crash in apply_chat_template, but on different inputs and likely different code paths." + "reason": "BLIP2 dtype propagation and RTDetrV2 output drift are unrelated model bugs." }, { - "left": "issue:29159", - "right": "issue:33357", + "left": "issue:44514", + "right": "issue:45290", "accept": false, - "reason": "Slow/fast tokenizer inconsistency is unrelated to the MacOS CLIP bus error." + "reason": "Both hit Qwen2.5-VL chat templating, but one is batched padding=False and the other is assistant tool-call content handling." }, { - "left": "issue:44843", - "right": "issue:44991", + "left": "issue:44464", + "right": "issue:44743", "accept": false, - "reason": "Offline Mistral regex/model_info regression and EMBEDDIA/est-roberta tokenizer loading failure are different problems." + "reason": "Chunked generation with compiled forward and Qwen3.5 recurrent-state reset are different generation/runtime behaviors." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the only accepted soft pair (44466/45005) is conservatively justified by the titles. The other rejections stay appropriately cautious and do not overclaim equivalence from broad subsystem similarity alone." + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft pair, issue:44466 and issue:45005, is a plausible same-bug match based on their titles around tied weights / save_pretrained regression in v5. The other verdicts do not overstate overlap and are reasonably rejected as distinct issues." }, "evaluator_used": true, "retried": false, @@ -6868,16 +6919,16 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:29159", "issue:32090", + "issue:33357", "issue:33666", "issue:34567", "issue:36010", - "issue:36331", "issue:41720", "issue:43262", - "issue:43278", "issue:43295", "issue:43452", "issue:43504", @@ -6894,7 +6945,6 @@ "issue:43749", "issue:43756", "issue:43927", - "issue:43994", "issue:44361", "issue:44451", "issue:44464", @@ -6905,8 +6955,9 @@ "issue:44610", "issue:44625", "issue:44821", - "issue:44849", + "issue:44843", "issue:44977", + "issue:44991", "issue:45440", "issue:45491", "issue:45588", @@ -6915,278 +6966,263 @@ "issue:45823", "issue:46002", "issue:46050", - "issue:46192", "issue:46209", "issue:46402" ], "soft_pairs": [ + "issue:29159|issue:33357", + "issue:44843|issue:44991", "issue:43646|issue:43749", - "issue:43756|issue:46402", "issue:43531|issue:43742", - "issue:43688|issue:45704", + "issue:43756|issue:46402", "issue:43688|issue:45440", - "issue:43653|issue:43927", + "issue:43688|issue:45704", "issue:45491|issue:46002", + "issue:43653|issue:43927", + "issue:43550|issue:43927", "issue:43688|issue:45701", "issue:44464|issue:44977", - "issue:43550|issue:43927", "issue:44479|issue:44625", - "issue:41720|issue:44464", "issue:43643|issue:44561", "issue:43550|issue:44361", - "issue:32090|issue:34567", + "issue:41720|issue:44464", "issue:44977|issue:46209", + "issue:32090|issue:34567", "issue:43577|issue:45588", - "issue:32090|issue:33666", "issue:43577|issue:44554", - "issue:44466|issue:46050", "issue:43504|issue:43720", - "issue:33666|issue:36010", + "issue:44466|issue:46050", + "issue:32090|issue:33666", "issue:43504|issue:43550", "issue:43504|issue:44610", + "issue:33666|issue:36010", "issue:43742|issue:44821", "issue:45823|issue:46050", "issue:43262|issue:43295", - "issue:43720|issue:43927", "issue:43452|issue:44451", - "issue:43606|issue:43927", - "issue:43994|issue:46192", - "issue:44849|issue:46209", - "issue:33666|issue:36331", - "issue:43278|issue:43688" + "issue:43720|issue:43927", + "issue:43606|issue:43927" ], - "prepared_review_unit_hash": "6e53ee60f4ccaa79823ea6b823a245543b96e943bf19caf3600cd655719a202f", + "prepared_review_unit_hash": "ba0f109bd830122856edf37b496a29081f9b34ea64cb20ec149f3a8b2cfd67bb", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37860, - "estimated_input_tokens": 9465, - "estimated_eval_tokens": 19186 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36686, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of unrelated Transformers regressions, deprecations, and model-specific bugs. The pairwise similarity is mostly superficial (shared words like \"broken\", \"load\", \"regression\", or \"v5\"), so I would not merge any of the soft edges as duplicates.", - "confidence": 0.19, - "canonical_issue_reason": "No true canonical duplicate exists here; if forced to anchor the cluster, issue 43646 is the broadest umbrella report about v5 breaking custom model initialization, but the items are too heterogeneous to form one duplicate set.", + "summary": "This is not a true duplicate cluster: the items span many unrelated transformer bugs (tokenizers, model loading, training, multimodal processors, device maps, quantization, etc.). All soft edges look like false positives rather than the same underlying bug/change.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue stands out because the set is heterogeneous and the proposed pairings do not describe the same root cause.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43646 is the most general and triage-friendly umbrella problem among the set, but only as a loose representative, not a real duplicate canonical.", + "best_issue_reason": "If one issue must represent the cluster, issue 44843 is the strongest standalone candidate because it is concrete, reproducible, and clearly scoped to a specific regression (offline tokenizer loading).", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Different subsystems and failure modes: tokenizer slow/fast parity vs a MacOS bus error on a CLIP model." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both involve tokenizer loading, but one is an offline hub access regression and the other is a specific tokenizer failure for a different model family." + }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "Different regressions: custom model init breakage vs FSDP CPU RAM-efficient loading." + "reason": "Distinct problems: custom model initialization breakage vs FSDP CPU RAM efficient loading regression." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding window behavior and MobileLLM key errors are unrelated bugs." }, { "left": "issue:43756", "right": "issue:46402", "accept": false, - "reason": "Unrelated model-specific loading/architecture bugs for different models and components." + "reason": "Different model-specific loading/architecture issues; no shared code-path is evident." }, { - "left": "issue:43531", - "right": "issue:43742", + "left": "issue:43688", + "right": "issue:45440", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is unrelated to MobileLLM key lookup failure." + "reason": "Both are MoE-related, but they concern different implementations and different correctness defects." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Aux-loss normalization bug vs T5/apex RMSNorm memory leak; different code paths." + "reason": "Aux-loss normalization in MoE models is unrelated to a T5 RMSNorm memory leak." }, { - "left": "issue:43688", - "right": "issue:45440", + "left": "issue:45491", + "right": "issue:46002", "accept": false, - "reason": "Both are model issues, but one is auxiliary-loss normalization and the other is a DeepSeekV3MoE implementation divergence." + "reason": "NaNs appear in different places with different triggers: sliding-window attention on GPU vs classification head init/checkout." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Tokenizer special-token registration problem vs config token-ID persistence bug." + "reason": "BigBirdTokenizer special-token registration and DiaConfig token-ID persistence are separate issues." }, { - "left": "issue:45491", - "right": "issue:46002", + "left": "issue:43550", + "right": "issue:43927", "accept": false, - "reason": "Both mention NaNs, but the trigger and subsystem differ: sliding-window attention vs classification head init/checkout." + "reason": "torch.compile + SDPA on Bamba is unrelated to config token-ID loss on save/load." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "General tokenization/version change report is not the same bug as auxiliary-loss normalization." + "reason": "A model-loss normalization bug is not the same as a broad tokenization-version behavior change." }, { "left": "issue:44464", "right": "issue:44977", "accept": false, - "reason": "Different generation failures: compiled-forward inconsistency vs flash-attention regression in Qwen3.5." - }, - { - "left": "issue:43550", - "right": "issue:43927", - "accept": false, - "reason": "Bamba torch.compile/SDPA failure is unrelated to DiaConfig token-ID loss." + "reason": "Chunked generation with compiled forward and flash-attention generation regression are different code paths and symptoms." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression for Qwen models vs num_labels propagation in Qwen3.5 config." - }, - { - "left": "issue:41720", - "right": "issue:44464", - "accept": false, - "reason": "Qwen3 auto-device-map CUDA assert is unrelated to chunked generation inconsistency." + "reason": "Video input regression for Qwen VL models is unrelated to num_labels propagation in Qwen3.5." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Both involve trust_remote_code/v5 compatibility, but they are distinct breakages in different API removals and config loading paths." + "reason": "Both mention trust_remote_code, but one is missing fields in AutoConfig output and the other is removal of an FX helper breaking models." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "Different failures in different models: torch.compile SDPA vs MLukeTokenizer task AttributeError." + "reason": "A compile/SDPA model-forward bug is not the same as an MLukeTokenizer AttributeError." }, { - "left": "issue:32090", - "right": "issue:34567", + "left": "issue:41720", + "right": "issue:44464", "accept": false, - "reason": "Trainer GPU broadcast NoneType error is unrelated to num_input_tokens_seen not updating." + "reason": "Auto device mapping cuda assertions are unrelated to chunked generation discrepancies." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Flash-attention generation regression is not the same as FP8 inference failure on B200." + "reason": "Flash-attention generation failure and FP8 inference on B200 are different hardware/path-specific bugs." }, { - "left": "issue:43577", - "right": "issue:45588", + "left": "issue:32090", + "right": "issue:34567", "accept": false, - "reason": "BLIP2 dtype-loading issue and flash_attention sink-less AttributeError affect different model stacks." + "reason": "Trainer NoneType broadcast error and TrainerState token counter not updating are unrelated Trainer bugs." }, { - "left": "issue:32090", - "right": "issue:33666", + "left": "issue:43577", + "right": "issue:45588", "accept": false, - "reason": "Trainer broadcast error is unrelated to Qwen2-VL multi-GPU training support." + "reason": "Blip2 dtype loading and flash_attention s_aux=None crash are separate issues." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "BLIP2 dtype mismatch is unrelated to the MPS attention correctness bug." + "reason": "A dtype persistence issue does not match an MPS attention correctness bug." }, { - "left": "issue:44466", - "right": "issue:46050", + "left": "issue:43504", + "right": "issue:43720", "accept": false, - "reason": "Serialization of tied weights is unrelated to sinq quantization save/load breakage." + "reason": "Legacy-field model loading and packed-weight unpacking during accelerate loading are different failure causes." }, { - "left": "issue:43504", - "right": "issue:43720", + "left": "issue:44466", + "right": "issue:46050", "accept": false, - "reason": "Beit legacy-field load failure and BitNet packed-weight loading failure are different model-loading bugs." + "reason": "Serialization inconsistency of tied weights is not the same as sinq quantized model reload breakage." }, { - "left": "issue:33666", - "right": "issue:36010", + "left": "issue:32090", + "right": "issue:33666", "accept": false, - "reason": "Qwen2-VL multi-GPU training request is unrelated to GenerationMixin import failure." + "reason": "Trainer broadcast failure and Qwen2-VL multi-GPU training are unrelated." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Different model families and failures: Beit legacy config vs Bamba torch.compile SDPA." + "reason": "These are different model-loading/compile bugs with no shared root cause." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Beit pretrained loading bug is unrelated to OmDet-Turbo processor/model input-size mismatch." + "reason": "Beit legacy-field loading and OmDet-Turbo processor size mismatch are unrelated." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Multi-GPU training for Qwen2-VL has no meaningful overlap with GenerationMixin import failures." }, { "left": "issue:43742", "right": "issue:44821", "accept": false, - "reason": "MobileLLM key error and AutoImageProcessor URL loading are unrelated loader issues." + "reason": "Key error loading one model is not the same as loading AutoImageProcessor from a URL." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Gemma4 device-map forward mismatch is unrelated to sinq quantized save/load corruption." + "reason": "Device-mapped forward mismatch and quantized save/load corruption are different bugs." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio processor sampling-rate default bug is unrelated to processor.tokenizer regression." - }, - { - "left": "issue:43720", - "right": "issue:43927", - "accept": false, - "reason": "BitNet packed-weights loading and DiaConfig token-ID persistence are different mechanisms and models." + "reason": "Audio processor chat-template sampling-rate defaulting and processor.tokenizer/image handling regressions are separate issues." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "gguf_file import/load regression is unrelated to inability to load ScandiBERT." + "reason": "gguf_file-related from_pretrained breakage and inability to load a specific model are not the same defect." }, { - "left": "issue:43606", + "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "CPU-offload device mismatch in bark-small is unrelated to DiaConfig token-ID loss." - }, - { - "left": "issue:43994", - "right": "issue:46192", - "accept": false, - "reason": "Siglip2 bad outputs and return_tensors being ignored are different API/data-flow bugs." - }, - { - "left": "issue:44849", - "right": "issue:46209", - "accept": false, - "reason": "Qwen3.5 hidden-state bug is unrelated to FP8 inference failure on B200." - }, - { - "left": "issue:33666", - "right": "issue:36331", - "accept": false, - "reason": "Multi-GPU training discussion is unrelated to CustomTrainer.compute_loss signature mismatch." + "reason": "Packed-weight loading in BitNet and config token-ID loss in Dia are unrelated." }, { - "left": "issue:43278", - "right": "issue:43688", + "left": "issue:43606", + "right": "issue:43927", "accept": false, - "reason": "Embedding dtype drift during eval is unrelated to auxiliary-loss normalization." + "reason": "CPU offload device mismatch for bark-small is unrelated to DiaConfig token-ID persistence." } ] }, @@ -7208,11 +7244,12 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:28218", "issue:33357", "issue:33666", + "issue:36331", "issue:39692", "issue:43257", "issue:43278", @@ -7224,25 +7261,23 @@ "issue:43525", "issue:43540", "issue:43572", - "issue:43575", "issue:43606", "issue:43646", - "issue:43701", + "issue:43688", "issue:43720", "issue:43824", "issue:43825", "issue:43844", "issue:43927", + "issue:43994", "issue:44291", "issue:44361", "issue:44464", "issue:44466", "issue:44488", - "issue:44493", "issue:44568", "issue:44589", "issue:44625", - "issue:44792", "issue:44843", "issue:44849", "issue:44964", @@ -7252,260 +7287,261 @@ "issue:45357", "issue:45405", "issue:45520", - "issue:45584", "issue:45759", "issue:45812", "issue:45823", "issue:46002", - "issue:46285" + "issue:46192", + "issue:46209", + "issue:46424" ], "soft_pairs": [ + "issue:43994|issue:46192", + "issue:33666|issue:36331", "issue:43525|issue:43572", - "issue:44625|issue:46002", + "issue:44849|issue:46209", + "issue:43278|issue:43688", "issue:43720|issue:44361", + "issue:44625|issue:46002", "issue:43493|issue:43825", "issue:28218|issue:33357", "issue:39692|issue:43493", "issue:44964|issue:45405", "issue:44849|issue:45357", "issue:44568|issue:45812", + "issue:44464|issue:46424", + "issue:45200|issue:45229", "issue:43540|issue:44977", "issue:43493|issue:43844", - "issue:45200|issue:45229", "issue:44488|issue:44843", "issue:39692|issue:43295", "issue:43278|issue:43844", + "issue:43334|issue:43572", "issue:43646|issue:44466", "issue:33357|issue:33666", - "issue:43334|issue:43572", "issue:43257|issue:43299", "issue:43334|issue:45759", - "issue:45229|issue:45823", "issue:44589|issue:45520", "issue:43824|issue:44589", + "issue:45229|issue:45823", "issue:43927|issue:44361", "issue:43504|issue:43606", - "issue:43278|issue:43701", - "issue:43572|issue:44291", - "issue:44464|issue:45584", - "issue:43575|issue:45405", - "issue:44493|issue:46285", - "issue:44792|issue:45405" + "issue:43572|issue:44291" ], - "prepared_review_unit_hash": "e7ac16688954c4d16080573033d0ce3df37eaa3f4f13c695c74e13e5c012efa9", + "prepared_review_unit_hash": "69c312aa1ab65211528509f9ef9734804ac9c78137ec8db2541f773341ad6165", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37097, - "estimated_input_tokens": 9275, - "estimated_eval_tokens": 18806 + "serialized_chars": 37120, + "estimated_input_tokens": 9280, + "estimated_eval_tokens": 18816 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The soft-similarity pairs are mostly false positives: they share broad themes like tokenizers, config fields, or Qwen/Gemma/SigLIP models, but not the same concrete bug or fix path. No duplicates should be merged here.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue: the candidate issues are heterogeneous and do not collapse into one underlying bug cluster.", + "summary": "These issues are a heterogeneous mix of unrelated tokenizer, configuration, model-loading, training, and runtime problems. None of the soft pairs look like the same underlying bug or change, so no duplicates should be merged here.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the set does not form one duplicate cluster and the pairs span distinct subsystems and failure modes.", "canonical_pr_reason": null, - "best_issue_reason": "No single best issue: none of the items is a valid representative duplicate target for the others.", + "best_issue_reason": "No issue is a good global representative because the items are not variants of one bug; even the most similar titles refer to different code paths or models.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "Different problems: SigLIP2 model/pipeline output correctness versus multimodal `return_tensors` handling." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Both involve training, but one is Qwen2-VL multi-GPU training and the other is a `compute_loss` signature break in custom trainers." + }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Both are config-attribute regressions, but they affect different model families and different missing fields (`pad_token_id` vs `pad_token_idx`); not the same bug." + "reason": "Similar symptom class, but different configs and fields: `Llama4Config.pad_token_id` vs `StableLmConfig.pad_token_idx`." }, { - "left": "issue:44625", - "right": "issue:46002", + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-state behavior is unrelated to FP8 inference failure on B200." + }, + { + "left": "issue:43278", + "right": "issue:43688", "accept": false, - "reason": "Different problems: propagated `num_labels` on Qwen3.5 multimodal config versus NaNs in classification heads after checkout/init." + "reason": "Different bugs: embedding dtype drift during eval versus auxiliary-loss normalization in MoE models." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "Unrelated failures: BitNet packed-weight loading versus MLukeTokenizer task-time AttributeError." + "reason": "Unrelated subsystems: BitNet accelerate loading versus tokenizer/task AttributeError for MLuke." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "`num_labels` propagation in Qwen3.5 is not the same as NaNs in classification heads after checkout/init." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 implementation discrepancy and a pipeline task-support error are different code paths and different bugs." + "reason": "SigLIP2 implementation mismatch is unrelated to a pipeline error message about translation tasks." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer spacing bug versus MacOS bus error in CLIP model loading; no shared underlying defect." + "reason": "Tokenizer spacing regression and MacOS bus error are unrelated failures." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Both mention SigLIP2, but one is a doc example/model-processor mismatch and the other is a HF-vs-JAX implementation discrepancy." + "reason": "Both mention SigLIP2, but one is docs-example issues and the other is an implementation discrepancy; not the same bug." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Model-loading failure for Phi-4 multimodal is unrelated to the PEFT minimum-version packaging issue." + "reason": "Model-loading failure for Phi-4 multimodal is unrelated to the PEFT version pin bump." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 regressions: hidden-states behavior versus incorrect save_pretrained key serialization." + "reason": "Different Qwen3.5 regressions: hidden-states output versus incorrect visual encoder keys in save_pretrained." }, { "left": "issue:44568", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer-related, but they concern different models and different failure modes (missing BOS/EOS vs wrong token IDs across Granite)." + "reason": "Both are tokenizer regressions, but they affect different models and different failure modes." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Chunked generation with compiled forward is unrelated to cache length handling for quantized cache." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Gemma4 text-only mm_token_type_ids defaulting and Gemma4 multi-GPU OOM are different issues." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Different Qwen-family issues: video-input processing error versus flash-attention generation bug." + "reason": "Qwen3OmniMoe video-input processing and Qwen3.5 flash-attention generation are unrelated." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 discrepancy is unrelated to gradient explosion with HfDeepSpeedConfig + ZeRO-3." - }, - { - "left": "issue:45200", - "right": "issue:45229", - "accept": false, - "reason": "Gemma4 fine-tuning mask defaults and Gemma4 multi-GPU inference OOM are different problems." + "reason": "SigLIP2 implementation discrepancy versus gradient explosion in ZeRO-3 are not the same code-path problem." }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "A model-specific load failure is not the same as the offline-mode regression in `_patch_mistral_regex`." + "reason": "Model load failure for `cjvt/sleng-bert` is not the same as the offline-mode hub call in `_patch_mistral_regex`." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "Both are multimodal/doc-regression flavored, but one is SigLIP2 docs and the other is a processor/tokenizer API regression." + "reason": "Documentation example errors and a regression in `processor.tokenizer`/image passing are distinct issues." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Embedding dtype drift during evaluate is unrelated to the ZeRO-3 gradient anomaly." + "reason": "Embedding dtype mismatch during evaluation is unrelated to ZeRO-3 gradient inflation." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different missing-field regressions in different models: Qwen3-VL `pad_token_id` vs StableLm `pad_token_idx`." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization breakage and device-dependent lm_head serialization are distinct v5 issues." + "reason": "Custom model initialization breakage is unrelated to device-dependent `lm_head.weight` serialization." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "Completely different areas: CLIP MacOS bus error versus Qwen2-VL multi-GPU training." - }, - { - "left": "issue:43334", - "right": "issue:43572", - "accept": false, - "reason": "Both involve missing token-id-style fields, but in different configs, different names, and different model stacks." + "reason": "MacOS bus error on CLIP loading is unrelated to Qwen2-VL multi-GPU training." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Related Qwen MoE models, but one is accelerate/deepspeed weight conversion and the other is dev loading of Qwen3VL MoE; not the same bug." + "reason": "Both mention Qwen3 MoE loading, but one is accelerate/deepspeed weight conversion and the other is a v5 Qwen3VL loading regression." }, { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Both touch composite model configs, but one is a missing pad token attribute and the other is AutoModelForCausalLM from_config unwrapping text_config." - }, - { - "left": "issue:45229", - "right": "issue:45823", - "accept": false, - "reason": "Both are Gemma4, but one is inference OOM and the other is device mismatch in PLE forward." + "reason": "Missing `pad_token_id` in Qwen3-VL is not the same as failing to unwrap `text_config` in composite multimodal configs." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Different serialization/import failures: missing Float8 storage class versus flash_attn import_utils KeyError." + "reason": "Float8 storage deserialization and `flash_attn` import mapping on Python 3.13 are unrelated." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "Qwen2.5-VL import failure is unrelated to the Float8 storage-object lookup error." + "reason": "Importing a missing Qwen2.5-VL class is unrelated to a Float8 storage lookup failure." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma4, but OOM under multi-GPU inference and device mismatch under `device_map=\"auto\"` are different failures." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "DiaConfig save/load token-ID loss is a different subsystem from MLukeTokenizer task AttributeError." + "reason": "DiaConfig save/load token-ID loss is unrelated to MLuke tokenizer task errors." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field model preset loading and CPU offload device mismatch are unrelated bugs." - }, - { - "left": "issue:43278", - "right": "issue:43701", - "accept": false, - "reason": "BF16/FP32 dtype drift during evaluate is not the same as resume_from_checkpoint key mismatch." + "reason": "Legacy-field preset loading for BEiT and CPU-offload device mismatch for Bark are different bugs." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Both are model-loading regressions, but one is a missing config field and the other is an unexpected `_is_hf_initialized` constructor arg." - }, - { - "left": "issue:44464", - "right": "issue:45584", - "accept": false, - "reason": "Chunked generation with compiled forward and Whisper empty-transcription handling are unrelated generation bugs." - }, - { - "left": "issue:43575", - "right": "issue:45405", - "accept": false, - "reason": "Qwen2-57B tensor-parallel OOM has no connection to the PEFT version bump." - }, - { - "left": "issue:44493", - "right": "issue:46285", - "accept": false, - "reason": "Position-id key warnings across many models and CLIP text-model structure inconsistency are different regressions." - }, - { - "left": "issue:44792", - "right": "issue:45405", - "accept": false, - "reason": "Janus image-generation test failure is unrelated to the PEFT release-version problem." + "reason": "Missing `pad_token_idx` in StableLmConfig and an `init_empty_weights` `_is_hf_initialized` TypeError are different model-loading regressions." } ] }, @@ -7527,7 +7563,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:15354", "issue:28218", @@ -7540,10 +7576,10 @@ "issue:38175", "issue:42490", "issue:43065", - "issue:43232", - "issue:43493", + "issue:43278", "issue:43504", "issue:43540", + "issue:43575", "issue:43582", "issue:43606", "issue:43653", @@ -7554,13 +7590,12 @@ "issue:43957", "issue:43994", "issue:44112", - "issue:44186", "issue:44423", "issue:44493", "issue:44521", "issue:44554", "issue:44734", - "issue:44849", + "issue:44792", "issue:44964", "issue:44977", "issue:45072", @@ -7568,6 +7603,7 @@ "issue:45198", "issue:45200", "issue:45290", + "issue:45405", "issue:45406", "issue:45446", "issue:45520", @@ -7575,11 +7611,15 @@ "issue:45823", "issue:46050", "issue:46121", - "issue:46192", + "issue:46285", "issue:46369" ], "soft_pairs": [ + "issue:43278|issue:43701", + "issue:44493|issue:46285", + "issue:43575|issue:45405", "issue:43504|issue:43653", + "issue:44792|issue:45405", "issue:43606|issue:44112", "issue:43994|issue:44493", "issue:43873|issue:46050", @@ -7588,208 +7628,204 @@ "issue:43065|issue:46369", "issue:29942|issue:30333", "issue:42490|issue:43696", - "issue:43844|issue:44964", "issue:45406|issue:46121", - "issue:29159|issue:32090", + "issue:43844|issue:44964", "issue:28218|issue:34567", - "issue:43582|issue:44977", + "issue:29159|issue:32090", "issue:44554|issue:45446", "issue:45084|issue:45520", + "issue:43582|issue:44977", "issue:15354|issue:38175", "issue:43540|issue:45200", "issue:44521|issue:45290", "issue:45200|issue:45823", "issue:45072|issue:45198", - "issue:43493|issue:43701", - "issue:37428|issue:43957", - "issue:44186|issue:46192", - "issue:43232|issue:44849", - "issue:45823|issue:46192" + "issue:37428|issue:43957" ], - "prepared_review_unit_hash": "686c313acdd81cf9c31023082baba5469c21ab3c23a30c413cdd34d4d295aff0", + "prepared_review_unit_hash": "f3360c0e438f7552f80ad9a387970b0ac18664bb4cc595bd1b5a0bad49d26c90", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 35558, - "estimated_input_tokens": 8890, - "estimated_eval_tokens": 18036 + "serialized_chars": 35581, + "estimated_input_tokens": 8896, + "estimated_eval_tokens": 18048 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose lexical cluster, not a true duplicate set: the items span unrelated bugs in tokenization, model loading, serving, quantization/offload, and attention code paths. None of the soft pairs looks like the same underlying change or defect, so all soft edges should be rejected.", - "confidence": 0.95, - "canonical_issue_reason": "No single issue is a good canonical duplicate target; the apparent similarities are superficial and the underlying bugs differ across subsystems.", + "summary": "This cluster is heterogeneous: it spans tokenizer behavior, multimodal/chat templating, model loading, attention/flash-attn, training, serving, and environment/version issues. None of the proposed soft pairs look like the same underlying bug, so all should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "Issue 43873 is the best broad representative: it is open, heavily discussed, has multiple inbound references, and describes a user-facing quantization/offloading failure that is more central than the other, mostly narrow or CI-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "None of the issues is representative enough to serve as a global cluster canonical because the set is heterogeneous.", + "best_issue_reason": "Issue 43873 is the strongest single issue to anchor triage because it has the most active discussion and a broad symptom statement, but the cluster overall is not a duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Different failures: dtype change during evaluate vs resume-from-checkpoint key mismatch." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Unrelated symptoms: position-id warning vs CLIP class structure change." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Qwen2 tensor-parallel OOM and a PEFT version pin are unrelated." + }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Different bugs: Beit semantic-segmentation legacy-field loading vs BigBirdTokenizer special-token registration." + "reason": "Legacy-field model loading bug vs tokenizer special-token registration bug." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure is not the same as a released PEFT version mismatch." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Both are CI/device issues, but the models and failure mechanisms differ: CPU offload mismatch vs stale device override." + "reason": "CPU offload device mismatch and a GraniteSpeech CI stale test are different problems." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "Different symptoms and code paths: SigLIP2 inference discrepancy vs unexpected position-id keys across models." + "reason": "SigLIP2 bad outputs and unexpected position-id keys do not share the same code path." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Quantization/offloading behavior vs broken save/load after sinq quantization are related only at a broad level." + "reason": "Both involve memory/quantization themes, but one is offloading behavior and the other is broken save/load for sinq quantization." }, { "left": "issue:45588", "right": "issue:46050", "accept": false, - "reason": "Flash-attention import crash on None aux state is unrelated to quantized model serialization failure." + "reason": "Flash-attention s_aux=None crash is unrelated to sinq persistence breakage." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Different serve regressions: multimodal continuous batching crash vs KV-cache continuation tensor indexing." + "reason": "Serve multimodal .to() crash and KV-cache continuation indexing are different serving bugs." }, { "left": "issue:43065", "right": "issue:46369", "accept": false, - "reason": "Same model family, but one is a dummy Conv2d placeholder and the other is a memory overflow during litetext use." + "reason": "Dummy Conv2d in Sam3PixelDecoder and Sam3 litetext memory overflow are distinct issues." }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Unrelated: Flash Attention 2 test failures vs MLflow reporting not updating after a failed job." + "reason": "Flash Attention 2 test failures and MLflow failure-reporting are unrelated CI/product bugs." }, { "left": "issue:42490", "right": "issue:43696", "accept": false, - "reason": "Import warning for trailing-slash paths is not the same as a CUDA out-of-memory error." - }, - { - "left": "issue:43844", - "right": "issue:44964", - "accept": false, - "reason": "Different problems: DeepSpeed/ZeRO-3 gradient growth vs meta-device model loading breakage." + "reason": "Trailing-slash import warning is not the same as a CUDA OOM on GPT-oss-20b." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4 serve crash due to missing _tokenizer is unrelated to rope-parameter TypeError handling." + "reason": "Gemma4Processor missing _tokenizer and rope-parameter JSON list handling are separate errors." }, { - "left": "issue:29159", - "right": "issue:32090", + "left": "issue:43844", + "right": "issue:44964", "accept": false, - "reason": "Tokenizer slow/fast inconsistency is unrelated to the Trainer NoneType broadcast error." + "reason": "ZeRO-3 gradient growth and Phi-4 multimodal model loading failure do not indicate the same defect." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Added-token spacing bug and num_input_tokens_seen not updating are separate tokenizer/state issues." + "reason": "Tokenizer added-space behavior is unrelated to TrainerState token-count tracking." }, { - "left": "issue:43582", - "right": "issue:44977", + "left": "issue:29159", + "right": "issue:32090", "accept": false, - "reason": "Apple Silicon warmup TypeError and Qwen3.5 flash-attention generation bug are different failures." + "reason": "Slow/fast tokenizer inconsistency and _gpu_broadcast_one NoneType are unrelated subsystems." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness with head-dim mismatch is unrelated to the AuxRequest PyTorch version check." + "reason": "MPS attention correctness and AuxRequest PyTorch version checking are different code paths." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Compilation error for non-template nodes and flash_attn import KeyError are different code paths." + "reason": "Template-node compilation error is unrelated to a flash_attn import KeyError on Python 3.13." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon warmup TypeError and Qwen3.5 flash-attention generation bug are not the same bug." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp export issue is unrelated to SigLIP2 zero-probability outputs." + "reason": "TorchScript GeneratorExp support and SigLIP2 zero probabilities are wholly unrelated." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Qwen3OmniMoe video-input validation and Gemma4 text-only mm_token_type_ids defaulting are different bugs." + "reason": "Video-input processing failure and Gemma4 mm_token_type_ids defaults are different multimodal issues." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both touch chat templates, but one is assistant-mask generation for multimodal inputs and the other is tool-call content handling." + "reason": "All-zero assistant masks for multimodal inputs is not the same as tool-call messages with no content crashing." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Gemma4 text-only mm_token_type_ids defaulting and device_map auto device mismatch are distinct issues." + "reason": "Missing mm_token_type_ids defaults and device-map forward mismatch are separate Gemma4 problems." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "Different model-specific regressions: bf16 dtype mismatches vs Wav2Vec2 save_pretrained/tokenization failure." - }, - { - "left": "issue:43493", - "right": "issue:43701", - "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy is unrelated to resume_from_checkpoint key mismatch." + "reason": "Bfloat16 dtype mismatch in inference and Wav2Vec2 save/tokenization failure are unrelated." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "flash_attention_utils import error and meta-device loading breakage are separate loading/import bugs." - }, - { - "left": "issue:44186", - "right": "issue:46192", - "accept": false, - "reason": "LayoutLMv2Tokenizer crash on NER/batched padding is not the same as return_tensors being ignored with text_kwargs." - }, - { - "left": "issue:43232", - "right": "issue:44849", - "accept": false, - "reason": "Generation kwargs after sync_gpus and Qwen3.5 output_hidden_states behavior are different code paths." - }, - { - "left": "issue:45823", - "right": "issue:46192", - "accept": false, - "reason": "Device-mapping mismatch in Gemma4 forward is unrelated to return_tensors/text_kwargs handling." + "reason": "Flash-support import-name error and meta-device model loading breakage are different loading/import issues." } ] }, @@ -7811,7 +7847,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:28218", "issue:35141", @@ -7821,7 +7857,7 @@ "issue:41720", "issue:42490", "issue:42915", - "issue:43278", + "issue:43232", "issue:43295", "issue:43334", "issue:43388", @@ -7832,19 +7868,19 @@ "issue:43550", "issue:43572", "issue:43575", - "issue:43653", + "issue:43701", "issue:43824", "issue:43844", - "issue:43856", "issue:43873", "issue:43950", "issue:43957", "issue:43975", - "issue:44112", + "issue:44186", "issue:44466", "issue:44554", "issue:44610", "issue:44792", + "issue:44849", "issue:44898", "issue:44987", "issue:45127", @@ -7855,6 +7891,7 @@ "issue:45412", "issue:45520", "issue:45563", + "issue:45823", "issue:46178", "issue:46192", "issue:46209", @@ -7862,245 +7899,255 @@ "issue:46365" ], "soft_pairs": [ + "issue:43493|issue:43701", + "issue:44186|issue:46192", + "issue:43232|issue:44849", + "issue:45823|issue:46192", "issue:45381|issue:46209", "issue:43295|issue:43450", "issue:43295|issue:44610", "issue:44554|issue:46365", "issue:39692|issue:46192", - "issue:43295|issue:43873", "issue:43572|issue:44987", + "issue:43295|issue:43873", "issue:44554|issue:45356", "issue:43572|issue:43957", "issue:45520|issue:46178", "issue:44466|issue:45127", - "issue:43950|issue:46285", - "issue:43550|issue:43844", "issue:37428|issue:43824", + "issue:43950|issue:46285", "issue:44898|issue:45412", + "issue:43550|issue:43844", "issue:45200|issue:45325", - "issue:43450|issue:43575", "issue:43334|issue:43525", - "issue:44792|issue:45563", + "issue:43450|issue:43575", + "issue:42915|issue:43824", "issue:42490|issue:43493", "issue:41720|issue:46209", - "issue:42915|issue:43824", + "issue:43824|issue:43975", + "issue:44792|issue:45563", "issue:43388|issue:43526", "issue:28218|issue:35141", - "issue:43824|issue:43975", - "issue:43278|issue:43856", - "issue:38175|issue:43493", - "issue:43526|issue:46192", - "issue:43653|issue:44112" + "issue:38175|issue:43493" ], - "prepared_review_unit_hash": "828e981991eadcb2cd70ab16c93e7cacc52099c9dd2c3b3d01232ecf54ecc01b", + "prepared_review_unit_hash": "f5836354460b9cfae4fe980ef14b5825c9327ffc3bd082fa93d065cd2ccea539", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 36186, - "estimated_input_tokens": 9047, - "estimated_eval_tokens": 18350 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37015, + "estimated_input_tokens": 9254, + "estimated_eval_tokens": 18764 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: it mixes unrelated tokenizer, processor, model-loading, quantization, and backend/runtime bugs. None of the soft pairs look like the same concrete underlying issue, so all soft edges are rejected. The most representative global issue is the broad `from_pretrained()` regression affecting non-persistent buffers.", - "confidence": 0.86, - "canonical_issue_reason": "Issue 43950 is the most broadly applicable and concrete regression in the set: it describes a core model-loading correctness bug that can affect many models, not just one model family or an isolated test case.", + "summary": "Mostly heterogeneous issue cluster. The only plausible duplicates are two small subclusters: SigLIP2 behavior/parity reports, and missing `pad_token_id` on new config classes. The rest are clearly different bugs, import errors, or unrelated model/processor problems.", + "confidence": 0.74, + "canonical_issue_reason": "`issue:43493` is the broadest, most root-cause-oriented report in the SigLIP2 pair; it can plausibly subsume the more specific zero-probability symptom in `issue:38175`.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43950 is the best global representative because it targets a central loading path, has a clear failure mode, and is more generally impactful than the model-specific or symptom-level reports around it.", + "best_issue_reason": "`issue:43493` is the strongest representative of the only broad behavioral duplicate pair, with the clearest description of the underlying implementation discrepancy.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "Different problems: SigLIP2 parity/behavior vs checkpoint resume key mismatch." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "Both involve tokenization/input handling, but the bugs are different and not the same code-path issue." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Unrelated generation-state update bug vs Qwen3.5 hidden-state output bug." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Different subsystems: device-mapped Gemma4 forward mismatch vs `text_kwargs` ignoring `return_tensors`." + }, { "left": "issue:45381", "right": "issue:46209", "accept": false, - "reason": "Qwen2.5-VL video position-id bug and FP8 inference failure on B200 are different code paths and failure modes." + "reason": "Qwen2.5-VL video position-id bug is unrelated to FP8 inference failure on B200." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor/tokenizer API regression is unrelated to the batched video-shape bug." + "reason": "Both are processor/multimodal issues, but they affect different code paths and symptoms." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Different processor regressions: tokenizer/image-handling API breakage vs OmDet-Turbo resize mismatch." + "reason": "Different processor regressions: `processor.tokenizer`/image passing vs OmDet-Turbo image-size mismatch." }, { "left": "issue:44554", "right": "issue:46365", "accept": false, - "reason": "MPS attention correctness bug and Trainer import failure caused by datasets are unrelated." + "reason": "MPS attention correctness bug is unrelated to the Trainer import issue." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "Broken SigLIP2 docs/quantization example is not the same as the `return_tensors` precedence bug." + "reason": "SigLIP2 doc/example errors are unrelated to silent `return_tensors` handling." }, { - "left": "issue:43295", - "right": "issue:43873", + "left": "issue:43572", + "right": "issue:44987", "accept": false, - "reason": "Custom model processor regression and quantization offloading issue are different bugs." + "reason": "Missing `pad_token_idx` in StableLmConfig is not the same bug as the physical-intelligence/fast load failure." }, { - "left": "issue:43572", - "right": "issue:44987", + "left": "issue:43295", + "right": "issue:43873", "accept": false, - "reason": "Missing `pad_token_idx` in StableLmConfig is unrelated to the local-model loading failure." + "reason": "Processor regression vs quantization offloading behavior are different issues." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention bug and Kimi-K2.5 tokenizer codec regression are different problems." + "reason": "Attention correctness on MPS is unrelated to Kimi-K2.5 tokenizer regression." }, { "left": "issue:43572", "right": "issue:43957", "accept": false, - "reason": "Missing config attribute in StableLmConfig is not the same as meta-device loading breakage." + "reason": "Different model-loading regressions with different failure modes." }, { "left": "issue:45520", "right": "issue:46178", "accept": false, - "reason": "Flash-attn import utility error and tensorflow_text backend mapping error are unrelated import failures." + "reason": "Both are import/config errors, but they involve different backend mappings and unrelated symbols." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "Both involve weights, but one is serialization inconsistency and the other is LoRA merge collapse after vocab extension." + "reason": "One is serialization of tied weights; the other is LoRA merge collapse after vocab extension." }, { - "left": "issue:43950", - "right": "issue:46285", + "left": "issue:37428", + "right": "issue:43824", "accept": false, - "reason": "Non-persistent buffer corruption in loading is unrelated to the CLIPTextModel structure inconsistency." + "reason": "Both are import errors, but for different missing names and different modules." }, { - "left": "issue:43550", - "right": "issue:43844", + "left": "issue:43950", + "right": "issue:46285", "accept": false, - "reason": "torch.compile SDPA failure and ZeRO-3 gradient explosion are different runtime/training bugs." + "reason": "Non-persistent buffer corruption on load is not the same as CLIP structure inconsistency." }, { - "left": "issue:37428", - "right": "issue:43824", + "left": "issue:44898", + "right": "issue:45412", "accept": false, - "reason": "Missing flash-attn helper import and missing Qwen2_5_VL export are unrelated import errors." + "reason": "Perceiver resolution failure and RT-DETR memory release are unrelated." }, { - "left": "issue:44898", - "right": "issue:45412", + "left": "issue:43550", + "right": "issue:43844", "accept": false, - "reason": "Perceiver interpolation failure and RT-DETR memory leak are unrelated model-specific bugs." + "reason": "torch.compile/SDPA failure is unrelated to ZeRO-3 gradient inflation." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Gemma 4 text-only token-type defaults and Qwen2.5-VL temporal position scaling are different preprocessing issues." - }, - { - "left": "issue:43450", - "right": "issue:43575", - "accept": false, - "reason": "Batched video processor shape bug and tensor-parallel OOM are not the same issue." + "reason": "Different multimodal bugs: Gemma4 token-type defaults vs Qwen2.5-VL rope/position-id scaling." }, { "left": "issue:43334", "right": "issue:43525", + "accept": true, + "reason": "Same underlying bug class: new config objects missing `pad_token_id`, causing AttributeError in the generation/loading path." + }, + { + "left": "issue:43450", + "right": "issue:43575", "accept": false, - "reason": "Both mention missing `pad_token_id`, but they affect different configs/models and do not look like one concrete bug." + "reason": "Batched video processor shape bug is unrelated to tensor-parallel OOM." }, { - "left": "issue:44792", - "right": "issue:45563", + "left": "issue:42915", + "right": "issue:43824", "accept": false, - "reason": "Janus image-generation test failure and stale `num_return_sequences` warning are unrelated." + "reason": "FineGrainedFP8Config failure is unrelated to the Qwen2_5_VL import error." }, { "left": "issue:42490", "right": "issue:43493", "accept": false, - "reason": "Trailing-slash import warning and SigLIP2 HF-vs-JAX discrepancy are different issues." + "reason": "Trailing-slash local path warning is unrelated to SigLIP2 implementation parity." }, { "left": "issue:41720", "right": "issue:46209", "accept": false, - "reason": "Auto device-mapping CUDA assert and FP8 inference failure on B200 are different hardware/runtime problems." + "reason": "Qwen3 auto device-mapping assert on A800 is unrelated to B200 FP8 inference failure." }, { - "left": "issue:42915", - "right": "issue:43824", + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "ImportError for a missing class and tokenizer detokenization regression are different bugs." + }, + { + "left": "issue:44792", + "right": "issue:45563", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure and missing Qwen2_5_VL import are unrelated." + "reason": "Failed janus image-generation test and stale warning logic are not the same issue." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "gather_for_metrics tuple-label truncation and BeitImageProcessorFast label reduction are different label-handling bugs." + "reason": "Both mention labels, but one is metric gathering and the other is image processor label reduction." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer extra-space behavior and embedding reinitialization on resize are unrelated tokenization vs initialization bugs." - }, - { - "left": "issue:43824", - "right": "issue:43975", - "accept": false, - "reason": "ImportError for a model class and DeepSeek detokenization regression are not the same concrete issue." - }, - { - "left": "issue:43278", - "right": "issue:43856", - "accept": false, - "reason": "Evaluation dtype drift and Qwen3 MoE memory usage are unrelated training/eval bugs." + "reason": "Tokenizer spacing bug and embedding reinitialization after resize are separate issues." }, { "left": "issue:38175", "right": "issue:43493", - "accept": false, - "reason": "SigLIP2 zero-probability behavior and HF-vs-JAX implementation discrepancy are related only by model family, not the same bug." - }, - { - "left": "issue:43526", - "right": "issue:46192", - "accept": false, - "reason": "Beit label reduction and `return_tensors` being ignored are unrelated preprocessing bugs." - }, - { - "left": "issue:43653", - "right": "issue:44112", - "accept": false, - "reason": "BigBirdTokenizer special-token registration and GraniteSpeech stale device override test are unrelated CI/runtime issues." + "accept": true, + "reason": "Same SigLIP2 code-path problem: 38175 is a concrete symptom of the broader HF-vs-JAX implementation discrepancy reported in 43493." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The accepted duplicate pairs are plausible and mergeable into single fixes: the SigLIP2 discrepancy pair (43493/38175) and the missing pad_token_id config pair (43334/43525). The rejected pairs are generally well-justified by distinct titles/symptoms/subsystems, and I don\u2019t see evidence of overclaiming." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -8115,27 +8162,28 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:29942", "issue:32090", "issue:37428", "issue:38175", "issue:43262", + "issue:43278", "issue:43335", "issue:43425", "issue:43450", - "issue:43504", + "issue:43526", "issue:43550", "issue:43606", "issue:43611", "issue:43653", "issue:43819", "issue:43827", + "issue:43856", "issue:43901", "issue:44112", "issue:44206", - "issue:44351", "issue:44442", "issue:44448", "issue:44493", @@ -8152,260 +8200,245 @@ "issue:45200", "issue:45245", "issue:45290", - "issue:45381", "issue:45412", "issue:45520", "issue:45584", "issue:45588", "issue:45593", "issue:45759", - "issue:45823", "issue:46170", + "issue:46192", "issue:46234", "issue:46285", "issue:46326", "issue:46402" ], "soft_pairs": [ + "issue:43278|issue:43856", + "issue:43526|issue:46192", + "issue:43653|issue:44112", "issue:45290|issue:46326", "issue:44442|issue:45198", - "issue:43262|issue:43450", "issue:38175|issue:45412", + "issue:43262|issue:43450", "issue:45520|issue:46170", "issue:45759|issue:46285", "issue:44514|issue:44521", "issue:44610|issue:44857", "issue:44561|issue:46234", - "issue:43819|issue:44112", "issue:44206|issue:45584", + "issue:43819|issue:44112", "issue:43827|issue:43901", "issue:43819|issue:45083", "issue:44857|issue:45593", - "issue:43335|issue:43550", "issue:43611|issue:46402", + "issue:43335|issue:43550", + "issue:37428|issue:43425", "issue:43606|issue:45083", "issue:29942|issue:32090", - "issue:37428|issue:43425", "issue:44448|issue:44743", "issue:45200|issue:45588", "issue:45127|issue:45245", "issue:43335|issue:43653", - "issue:44855|issue:45083", "issue:43819|issue:44493", - "issue:43550|issue:44112", - "issue:44206|issue:45823", - "issue:37428|issue:44351", - "issue:45290|issue:45381", - "issue:43335|issue:43504" + "issue:44855|issue:45083" ], - "prepared_review_unit_hash": "85241ec4ba346e11a0c5860f589e94946e6e83731a45282f34ddc9e6c9ec6176", + "prepared_review_unit_hash": "de7c78fe8cdce43b4d840a134d0ceee598269e5cffc20a89b98c089056027867", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36918, - "estimated_input_tokens": 9230, - "estimated_eval_tokens": 18716 + "soft_pair_count": 28, + "serialized_chars": 36338, + "estimated_input_tokens": 9085, + "estimated_eval_tokens": 18426 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a grab-bag of unrelated single-issue reports. None of the soft pairs appear to be the same underlying bug/change, so I would not merge any of them as duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "issue:44521 is the strongest triage representative in this set: it is open, highly discussed, and has the most external references, but the rest of the items do not form a true duplicate cluster around it.", + "summary": "This cluster is a loose similarity bucket of mostly unrelated Transformers issues; I don\u2019t see any soft pair that looks like the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the items span distinct bugs across models, processors, tokenizers, training, CI, and import/runtime compatibility.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the best issue to keep as the representative user-facing bug because it has the most activity and the broadest downstream impact among the listed issues.", + "best_issue_reason": "No single issue is a good global representative for this cluster because the topics are too heterogeneous.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Different problems: dtype drift between train/eval vs MoE training memory usage." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "Different processor bugs: label reduction vs `return_tensors` being ignored when `text_kwargs` is passed." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Unrelated: BigBirdTokenizer special-token registration vs a GraniteSpeech CI device-override test." + }, { "left": "issue:45290", "right": "issue:46326", "accept": false, - "reason": "Both mention apply_chat_template/chat handling, but they describe different failure modes and code paths: tool-call/content-less assistant messages vs continue_final_message on mistral_common tokenizers." + "reason": "Both involve chat templating, but the concrete bugs differ: tool-call/empty-content crash vs `continue_final_message` on mistral_common tokenizers." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Both are tokenizer/speech-model loading problems, but they affect different models and different operations: AutoTokenizer loading vs save_pretrained/tokenization behavior." + "reason": "Different tokenizer failures for different models and code paths (`AutoTokenizer` load vs save/tokenization behavior)." }, { - "left": "issue:43262", - "right": "issue:43450", + "left": "issue:38175", + "right": "issue:45412", "accept": false, - "reason": "These are different modalities and bugs: audio chat-template sampling-rate defaulting vs batched video processor output shape." + "reason": "Completely different areas: SIGLIP output probabilities vs RT-DETR memory release/GC." }, { - "left": "issue:38175", - "right": "issue:45412", + "left": "issue:43262", + "right": "issue:43450", "accept": false, - "reason": "Completely different subsystems and symptoms: SigLIP zero probabilities vs RT-DETR memory not being released." + "reason": "Different modality processors and bugs: audio chat-template sample-rate default vs batched video shape handling." }, { "left": "issue:45520", "right": "issue:46170", "accept": false, - "reason": "Both are import/runtime compatibility problems, but one is a missing flash_attn key on Python 3.13 and the other is Windows ROCm torch.distributed dependency handling." + "reason": "Different environment/import failures: flash_attn lookup on Python 3.13 vs Windows ROCm/torch.distributed dependency issues." }, { "left": "issue:45759", "right": "issue:46285", "accept": false, - "reason": "Both involve multimodal/model structure changes, but the concrete bugs are different: unwrapping text_config in AutoModelForCausalLM.from_config vs inconsistent CLIP text model structure." + "reason": "Both are model-structure regressions, but for different model families and different missing/unwrapped components." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Same general apply_chat_template area, but one crashes on batched padding=False input while the other returns all-zero assistant masks for multimodal inputs." + "reason": "Same general API, but different bugs: batched padding=False crash vs all-zero assistant masks for multimodal inputs." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Different model families and failures: OmDet-Turbo image size mismatch vs LwDetrImageLoss AMP/CUDA crash." + "reason": "Different detection issues: wrong processor image size vs AMP/CUDA crash in a loss function." }, { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "Both are missing utility symbols, but they are different APIs and breakages: removed is_torch_fx_available vs missing is_falcon_mamba_ssm_available." + "reason": "Both mention missing availability helpers, but they are different symbols causing different breakages." }, { - "left": "issue:43819", - "right": "issue:44112", + "left": "issue:44206", + "right": "issue:45584", "accept": false, - "reason": "Unrelated bugs: DAC.from_latents mismatch with forward pass vs a stale GraniteSpeech CI device override test." + "reason": "Different speech bugs: unsupported `center` arg in a feature extractor vs empty-transcription generation failure." }, { - "left": "issue:44206", - "right": "issue:45584", + "left": "issue:43819", + "right": "issue:44112", "accept": false, - "reason": "Different audio bugs: unsupported center arg in LasrFeatureExtractor vs Whisper generation failing after align_special_tokens." + "reason": "Different code paths: DAC latent reconstruction mismatch vs GraniteSpeech stale CI test." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs-related pipeline issues, but they concern different documentation and different API changes." + "reason": "Both are docs-related, but they refer to different pipeline APIs and different documentation mismatches." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "Different code paths and domains: DAC latent reconstruction vs qwen3_omni_moe feature-length helper behavior." + "reason": "Different subsystems and symptoms: DAC latent behavior vs qwen3_omni_moe feature-length helper semantics." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "Both touch detection/training losses, but one is an AMP CUDA crash and the other is auxiliary losses being skipped when denoising is off." + "reason": "Different detection-training bugs: AMP crash vs missing auxiliary losses when denoising is off." + }, + { + "left": "issue:43611", + "right": "issue:46402", + "accept": false, + "reason": "Different loading regressions affecting different models and versions." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "Unrelated model/config bug vs compile-time SDPA failure; no shared underlying issue." + "reason": "Unrelated model/config bug vs torch.compile/SDPA failure." }, { - "left": "issue:43611", - "right": "issue:46402", + "left": "issue:37428", + "right": "issue:43425", "accept": false, - "reason": "Different model-loading regressions: base_model_prefix handling vs Sam3 text encoder weights not loading." + "reason": "Different compatibility issues: missing flash-attention helper import vs general Torch 2.10 incompatibility." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "No meaningful overlap beyond being model bugs; CPU offload device mismatch is unrelated to qwen3_omni_moe helper behavior." + "reason": "Different problems in different areas: CPU-offload device mismatch vs feature-extraction length helper behavior." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are separate failures with different causes." - }, - { - "left": "issue:37428", - "right": "issue:43425", - "accept": false, - "reason": "A missing flash attention symbol import is not the same as a broad Torch 2.10 incompatibility report." + "reason": "Unrelated: Flash Attention 2 test failures vs Trainer `_gpu_broadcast_one` NoneType error." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Different models and bugs: Pegasus output differences across versions vs recurrent state reset in modular_qwen3_5 cache handling." + "reason": "Different model behaviors: Pegasus output regression vs Qwen recurrent-state reset with cache." }, { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Different multimodal bugs: Gemma 4 token type defaults vs flash_attention integration crashing on s_aux=None." + "reason": "Different multimodal issues: missing default `mm_token_type_ids` vs flash-attention `s_aux=None` crash." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA merge collapse with tied embeddings is unrelated to the category-count runtime error." + "reason": "Unrelated: LoRA merge/tied embeddings collapse vs category-count limit error." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "SwitchTransformers sparse-layer configuration and BigBirdTokenizer special-token registration are unrelated." - }, - { - "left": "issue:44855", - "right": "issue:45083", - "accept": false, - "reason": "Python 3.13 JIT parsing/import failure is unrelated to qwen3_omni_moe length-helper behavior." + "reason": "Different model/config bug vs tokenizer special-token registration issue." }, { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "Different symptoms and model families: DAC latent mismatch vs unexpected position-id keys across many models." - }, - { - "left": "issue:43550", - "right": "issue:44112", - "accept": false, - "reason": "torch.compile+SDPA failure in Bamba-9B-v2 is unrelated to a GraniteSpeech CI device override test." - }, - { - "left": "issue:44206", - "right": "issue:45823", - "accept": false, - "reason": "These are different failures: unsupported feature-extractor arg vs Gemma4 PLE device mismatch during forward." - }, - { - "left": "issue:37428", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but they involve different missing symbols and different modules; not the same underlying bug." + "reason": "Different failures: DAC latent mismatch vs unexpected key in position-id state dicts." }, { - "left": "issue:45290", - "right": "issue:45381", - "accept": false, - "reason": "Both are chat/multimodal-related, but one is a chat-template crash with tool calls and the other is wrong vision_position_ids for Qwen2.5-VL video input." - }, - { - "left": "issue:43335", - "right": "issue:43504", + "left": "issue:44855", + "right": "issue:45083", "accept": false, - "reason": "Completely different issues: SwitchTransformers sparse-layer creation vs BEiT pretrained loading with a legacy field." + "reason": "Different issues entirely: Python 3.13 import/torch.jit parsing vs qwen3_omni_moe helper behavior." } ] }, @@ -8427,317 +8460,310 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:28218", "issue:29127", "issue:29942", "issue:32090", - "issue:33357", "issue:34689", "issue:35141", + "issue:37428", "issue:38175", "issue:39692", "issue:40444", "issue:42490", "issue:43278", "issue:43335", + "issue:43504", "issue:43526", "issue:43550", - "issue:43575", "issue:43606", "issue:43701", "issue:43720", "issue:43746", "issue:43825", "issue:43854", - "issue:43873", - "issue:43927", "issue:44112", + "issue:44206", + "issue:44351", "issue:44360", "issue:44448", "issue:44493", "issue:44561", "issue:44779", - "issue:44855", "issue:44857", "issue:44977", "issue:45072", "issue:45083", "issue:45229", + "issue:45290", "issue:45310", + "issue:45381", "issue:45459", "issue:45478", "issue:45588", "issue:45593", - "issue:45698", + "issue:45823", "issue:45859", "issue:45920", "issue:46192", "issue:46209", - "issue:46234", - "issue:46399" + "issue:46234" ], "soft_pairs": [ + "issue:37428|issue:44351", + "issue:43550|issue:44112", + "issue:44206|issue:45823", + "issue:45290|issue:45381", + "issue:43335|issue:43504", "issue:29127|issue:32090", - "issue:45588|issue:45859", "issue:43278|issue:43825", + "issue:45588|issue:45859", "issue:39692|issue:43526", "issue:45310|issue:45478", "issue:45459|issue:46234", - "issue:43335|issue:45083", "issue:39692|issue:44493", + "issue:43335|issue:45083", "issue:43720|issue:44112", "issue:44448|issue:44779", "issue:43335|issue:44112", "issue:43335|issue:45593", + "issue:43606|issue:44857", "issue:42490|issue:43854", "issue:34689|issue:35141", - "issue:43606|issue:44857", "issue:44360|issue:44561", - "issue:45920|issue:46192", "issue:38175|issue:44493", "issue:44857|issue:45072", - "issue:44977|issue:45229", + "issue:45920|issue:46192", "issue:44448|issue:44977", + "issue:44977|issue:45229", "issue:28218|issue:32090", "issue:40444|issue:43526", "issue:43746|issue:46209", "issue:43701|issue:45083", - "issue:32090|issue:33357", - "issue:29942|issue:35141", - "issue:44855|issue:46399", - "issue:43550|issue:45083", - "issue:42490|issue:43575", - "issue:43927|issue:45698", - "issue:43873|issue:45229" + "issue:29942|issue:35141" ], - "prepared_review_unit_hash": "ebc79c910bedc643978da39b338650ea32703edf5c76e0a8cb421d4642c18912", + "prepared_review_unit_hash": "be3f77fb01506dda01a53602ed0d11d8ebd4a78a581a33508732b40439543869", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37340, - "estimated_input_tokens": 9335, - "estimated_eval_tokens": 18926 + "soft_pair_count": 31, + "serialized_chars": 37049, + "estimated_input_tokens": 9263, + "estimated_eval_tokens": 18782 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "One soft edge is a plausible duplicate: the two Qwen3.5 MoE from_pretrained error reports. The remaining pairs do not look like the same underlying bug from the visible metadata and should stay separate.", - "confidence": 0.91, - "canonical_issue_reason": "issue:45310 is the better canonical issue because it is the earlier report and has at least one inbound reference; issue:45478 has the same title, so the visible metadata supports duplicate status but not a claim that one title is more descriptive.", + "summary": "The cluster is mostly heterogeneous, with one likely duplicate pair centered on the Qwen3.5 MoE `from_pretrained` regression (45310/45478). The rest of the items appear to be unrelated bugs, feature regressions, or documentation/test issues.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 45310 is the best canonical issue for the Qwen3.5 MoE `from_pretrained` bug because it is an open, concrete report and 45478 appears to describe the same underlying failure with essentially the same scope.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 is the best issue anchor for this cluster because it is the earliest of the matching reports and already has some linkage, making it the most stable representative of the duplicate pair.", + "best_issue_reason": "45310 is the best representative issue: it is open, specific, and the earlier of the two highly similar Qwen3.5 MoE `from_pretrained` reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:29127", - "right": "issue:32090", + "left": "issue:37428", + "right": "issue:44351", "accept": false, - "reason": "Different subsystems and symptoms: LayoutLMv3 error-message clarity vs Trainer GPU broadcast TypeError." + "reason": "Different missing-import regressions in different code paths (`_flash_supports_window_size` vs `HybridCache`)." }, { - "left": "issue:45588", - "right": "issue:45859", + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba SDPA/torch.compile failure and GraniteSpeech CI flake are unrelated bugs." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Unsupported `center` arg in LasrFeatureExtractor is unrelated to Gemma4 device-map forward mismatch." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash and Qwen2.5-VL video position-id bug are different failures." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug and BEiT legacy-field load failure are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:32090", "accept": false, - "reason": "Different code paths: flash_attention sinkless-model crash vs Qwen3.5 rotary embedding CPU-offload incompatibility." + "reason": "LayoutLMv3 error-message clarity and Trainer `_gpu_broadcast_one` NoneType failure are unrelated." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Different bugs: training/eval dtype drift vs pipeline translation-task warning text." + "reason": "BF16/FP32 dtype drift and pipeline translation-warning behavior are different issues." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Flash-attention `s_aux=None` crash and Qwen3_5Moe CPU-offload incompatibility are distinct code paths." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "Documentation example failures for SigLIP2 are unrelated to BeitImageProcessorFast label reduction." + "reason": "SigLIP2 docs/example issues do not match the BEiTImageProcessorFast `reduce_labels` bug." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Same title and highly similar topic: both are Qwen3.5 MoE from_pretrained error reports, so they look like the same underlying bug." + "reason": "Same Qwen3.5 MoE `from_pretrained` regression with the same version scope and no meaningful distinguishing behavior." }, { "left": "issue:45459", "right": "issue:46234", "accept": false, - "reason": "Tokenizer/protobuf error masking is unrelated to a missing import_utils capability flag." + "reason": "Tokenizer protobuf-error masking and missing `is_falcon_mamba_ssm_available` export are unrelated API issues." }, { - "left": "issue:43335", - "right": "issue:45083", + "left": "issue:39692", + "right": "issue:44493", "accept": false, - "reason": "Different model/config bugs: SwitchTransformers sparse-layer creation vs qwen3_omni_moe helper behavior." + "reason": "SigLIP2 documentation problems and unexpected position-id keys are different bugs." }, { - "left": "issue:39692", - "right": "issue:44493", + "left": "issue:43335", + "right": "issue:45083", "accept": false, - "reason": "Documentation example errors are not the same as the unexpected position-id key behavior described here." + "reason": "SwitchTransformers config bug and qwen3_omni_moe helper behavior are unrelated." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "Both are CI-related, but they hit different models and failure modes." + "reason": "Packed-weights loading bug and GraniteSpeech CI device-override flake are unrelated." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Both are v4/v5 regressions, but one is Pegasus output drift and the other is DeepSeek tokenizer behavior." + "reason": "Pegasus output regression and DeepSeek tokenizer regression are distinct model-specific problems." }, { "left": "issue:43335", "right": "issue:44112", "accept": false, - "reason": "Sparse-layer config bug vs GraniteSpeech stale device override test are unrelated." + "reason": "SwitchTransformers config bug and GraniteSpeech CI failure are unrelated." }, { "left": "issue:43335", "right": "issue:45593", "accept": false, - "reason": "Different behavior: SwitchTransformers sparse-layer setup vs D-FINE auxiliary-loss omission." + "reason": "SwitchTransformers sparse-layer bug and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch and LwDetr AMP crash are separate failures." }, { "left": "issue:42490", "right": "issue:43854", "accept": false, - "reason": "Local-path import warning is unrelated to GLM-4.7 model-loading failures." + "reason": "Local-path import warning and GLM-4.7 test loading issue are unrelated." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Both concern loading/model setup, but they describe different bugs." - }, - { - "left": "issue:43606", - "right": "issue:44857", - "accept": false, - "reason": "CPU offload device mismatch is unrelated to an AMP/CUDA crash in a different model." + "reason": "Llama 3.2 Vision loading regression and embedding reinitialization after resize are different bugs." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer activation behavior is unrelated to the v5 removal of is_torch_fx_available." - }, - { - "left": "issue:45920", - "right": "issue:46192", - "accept": false, - "reason": "Tokenizer ID regression across models is different from return_tensors being ignored when text_kwargs is passed." + "reason": "DSA indexer/ReLU discussion and `is_torch_fx_available` removal breakage are unrelated." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Zero probabilities in SigLIP2 are unrelated to the unexpected position-id key warning." + "reason": "SigLIP2 zero-probability behavior and unexpected position-id keys are distinct bugs." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Different numeric-stability/device-dtype problems in different models; not the same bug." + "reason": "LwDetr AMP loss crash and dtype mismatches in SwitchTransformers/TimmWrapper inference are different issues." }, { - "left": "issue:44977", - "right": "issue:45229", + "left": "issue:45920", + "right": "issue:46192", "accept": false, - "reason": "Flash-attention generation issue is unrelated to Gemma4 multi-GPU CUDA OOM." + "reason": "AutoTokenizer wrong-token-ID regression and `return_tensors` being ignored with `text_kwargs` are not the same bug." }, { "left": "issue:44448", "right": "issue:44977", "accept": false, - "reason": "Both affect generation, but they are different model-specific regressions with different causes." + "reason": "Pegasus output regression and Qwen3.5 flash-attention generation issue are unrelated." }, { - "left": "issue:28218", - "right": "issue:32090", + "left": "issue:44977", + "right": "issue:45229", "accept": false, - "reason": "Tokenizer added-space bug is unrelated to Trainer broadcast TypeError." + "reason": "Flash-attention generation bug and Gemma4 multi-GPU OOM are different problems." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer added-space behavior and Trainer NoneType broadcast error are unrelated." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "Multimodal IterableDataset finetuning failure is unrelated to BeitImageProcessorFast label reduction." + "reason": "Qwen2.5-VL multi-image finetuning failure and BEiT reduce_labels bug are unrelated." }, { "left": "issue:43746", "right": "issue:46209", "accept": false, - "reason": "Local checkpoint loading with PEFT adapters is unrelated to FP8 inference on B200." + "reason": "Local checkpoint PEFT loading and FP8 B200 inference failure are distinct." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "resume_from_checkpoint key mismatch is unrelated to qwen3_omni_moe helper behavior." - }, - { - "left": "issue:32090", - "right": "issue:33357", - "accept": false, - "reason": "Trainer NoneType broadcast error is unrelated to the MacOS bus error loading CLIP." + "reason": "Resume-from-checkpoint key mismatch and qwen3_omni_moe helper behavior are unrelated." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures are unrelated to embedding reinitialization after resize." - }, - { - "left": "issue:44855", - "right": "issue:46399", - "accept": false, - "reason": "Both involve import-time/decorator parsing, but they are different errors in different code paths." - }, - { - "left": "issue:43550", - "right": "issue:45083", - "accept": false, - "reason": "torch.compile + SDPA failure in Bamba is unrelated to qwen3_omni_moe helper behavior." - }, - { - "left": "issue:42490", - "right": "issue:43575", - "accept": false, - "reason": "Local-path slash warning is unrelated to Qwen2-57B-A14B-Instruct tensor-parallel OOM." - }, - { - "left": "issue:43927", - "right": "issue:45698", - "accept": false, - "reason": "Both are save/load regressions, but they describe different underlying defects." - }, - { - "left": "issue:43873", - "right": "issue:45229", - "accept": false, - "reason": "Quantization/offloading behavior is unrelated to Gemma4 multi-GPU OOM." + "reason": "Flash Attention 2 CI failures and embedding reinitialization after resize are different bugs." } ] }, "evaluator_result": { "accept": false, - "feedback": "Reject: the duplicate-pair call on issue:45310 / issue:45478 is plausible, but the canonical_issue_reason overstates the evidence. In the packet both issues have the same title, so 45478 does not have a shorter title, and nothing shown supports that 45310 is more descriptive. Keep the verdict conservative and grounded only in the visible metadata." + "feedback": "The cluster-level conclusion is too strong for the packet. The evidence supports that 45310 and 45478 are highly similar and likely about the same Qwen3.5 MoE `from_pretrained` bug, but the packet does not explicitly confirm them as duplicates. Phrases like \"only confirmed duplicate set,\" \"exact regression,\" and \"later duplicate\" overstate the evidence. The soft-edge rejections are generally conservative, but the duplicate verdict should be downgraded to likely/same-topic rather than confirmed." }, "evaluator_used": true, "retried": true, @@ -8756,29 +8782,29 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:10000", "issue:15354", "issue:25251", - "issue:29127", + "issue:27957", "issue:32090", "issue:33290", "issue:33357", - "issue:34567", + "issue:34689", "issue:37428", - "issue:39692", "issue:41720", "issue:42490", "issue:43278", "issue:43329", "issue:43381", "issue:43412", - "issue:43611", + "issue:43550", + "issue:43575", "issue:43653", "issue:43873", + "issue:43927", "issue:43976", - "issue:44186", "issue:44206", "issue:44448", "issue:44466", @@ -8787,36 +8813,43 @@ "issue:44568", "issue:44743", "issue:44779", + "issue:44855", "issue:44857", - "issue:44871", "issue:44977", "issue:44987", + "issue:45083", "issue:45198", "issue:45229", "issue:45245", "issue:45372", "issue:45406", "issue:45440", - "issue:45459", "issue:45491", "issue:45563", "issue:45588", - "issue:45712", + "issue:45698", "issue:45803", "issue:45823", - "issue:46121" + "issue:46399" ], "soft_pairs": [ + "issue:32090|issue:33357", + "issue:44855|issue:46399", + "issue:43550|issue:45083", + "issue:43927|issue:45698", + "issue:42490|issue:43575", "issue:10000|issue:15354", "issue:44561|issue:44779", + "issue:37428|issue:44987", + "issue:43873|issue:45229", "issue:43278|issue:43329", "issue:41720|issue:45229", - "issue:37428|issue:44987", - "issue:44977|issue:45563", "issue:44857|issue:45588", + "issue:27957|issue:34689", + "issue:44977|issue:45563", "issue:44568|issue:45491", - "issue:43278|issue:43381", "issue:45372|issue:45588", + "issue:43278|issue:43381", "issue:43412|issue:45588", "issue:45588|issue:45823", "issue:45372|issue:45406", @@ -8826,218 +8859,205 @@ "issue:25251|issue:32090", "issue:45245|issue:45563", "issue:43653|issue:44743", - "issue:32090|issue:33290", "issue:44206|issue:44479", - "issue:43611|issue:46121", - "issue:29127|issue:34567", - "issue:42490|issue:43873", - "issue:33357|issue:39692", - "issue:44871|issue:44977", - "issue:44186|issue:45198", - "issue:45459|issue:45712" + "issue:32090|issue:33290" ], - "prepared_review_unit_hash": "426e6f8bc369488b2761af88a6f1e3ed9fdf553ab696fd5752234619c780936d", + "prepared_review_unit_hash": "08388b72805ea8acd4665019f04a9816810c2d5ec72aca918c34de2c015d2ff8", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35950, - "estimated_input_tokens": 8988, - "estimated_eval_tokens": 18232 + "serialized_chars": 35784, + "estimated_input_tokens": 8946, + "estimated_eval_tokens": 18148 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the paired items mostly share broad wording or subsystem names, but they describe different bugs, models, or code paths. I found no true duplicate pair to merge, so no single canonical issue stands out.", - "confidence": 0.91, + "summary": "The set is highly heterogeneous: the soft pairs share superficial vocabulary or broad subsystem overlap, but none look like the same underlying bug/change. No PRs are present, and I would reject every soft edge as a duplicate candidate.", + "confidence": 0.94, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45588 is the strongest standalone issue: it has a concrete crash, a specific file/function (`integrations/flash_attention.py`), and a clear repro path. It is actionable, but it still does not unify the other items.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Both are runtime failures, but one is a Trainer broadcast NoneType error and the other is a MacOS bus error on a community CLIP model; different code paths and symptoms." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Both mention decorator/script parsing, but one is a Python 3.13 torch.jit import error and the other is a Qwen3.5 registration bug; not the same failure." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Different models and bugs: torch.compile/SDPA failure in Bamba vs a feature-length helper bug in qwen3_omni_moe." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load, but one is custom token IDs being lost and the other is the wrong custom module being loaded; distinct serialization bugs." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "A trailing-slash import warning is unrelated to a tensor-parallel OOM when loading Qwen2-57B-A14B-Instruct." + }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "Unrelated model/export failures: German DistilBertModel issue vs torch.jit.script GeneratorExp export problem on ViT." + "reason": "Completely different models and failures: DistilBertModel issue vs torch.jit.script GeneratorExp/export problem for ViT." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Different regressions: removal of is_torch_fx_available affects trust_remote_code loading, while the other is a DeepSeek tokenizer behavior change." + "reason": "One is a removed API breaking trust_remote_code models; the other is tokenizer output corruption in Deepseek. Related release era, not the same bug." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "An import error from flash-attention utils and a model-loading failure for a specific repo are not the same code-path problem." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization/offloading behavior and multi-GPU OOM are different resource-management issues with different likely fixes." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Separate code paths: dtype mismatch in eval vs a multimodal token-counting bug in the video branch." + "reason": "BF16-to-FP32 dtype drift in eval is unrelated to undefined variables in a video token counting helper." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Both are model/runtime issues, but one is a Qwen3 auto device-map assert and the other is Gemma4 multi-GPU OOM; not the same bug." + "reason": "A cudaErrorAssert during auto device mapping is not the same as a multi-GPU OOM while loading Gemma4." }, { - "left": "issue:37428", - "right": "issue:44987", + "left": "issue:44857", + "right": "issue:45588", "accept": false, - "reason": "Both mention loading/import failures, but one is a missing flash-attention utility import and the other is a model-loading failure for a specific repo." + "reason": "AMP/CUDA loss crash in LwDetrImageLoss is unrelated to an AttributeError in flash_attention handling of `s_aux=None`." }, { - "left": "issue:44977", - "right": "issue:45563", + "left": "issue:27957", + "right": "issue:34689", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug vs a stale warning in paged generate(); different functionality and symptoms." + "reason": "Both concern model loading/attention, but XLMRoberta Flash Attention 2 and Llama 3.2 90B Vision load break are distinct issues." }, { - "left": "issue:44857", - "right": "issue:45588", + "left": "issue:44977", + "right": "issue:45563", "accept": false, - "reason": "CUDA AMP loss crash vs flash_attention.py AttributeError on s_aux=None; different components and failures." + "reason": "Flash-attention generation regression and a stale warning for num_return_sequences do not share the same bug surface." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token handling vs Gemma3 NaN embeddings from sliding-window attention; not the same issue." + "reason": "Tokenizer special-token insertion bug vs Gemma3 NaN embeddings on GPU; different layers of the stack." }, { - "left": "issue:43278", - "right": "issue:43381", + "left": "issue:45372", + "right": "issue:45588", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction is unrelated to BF16->FP32 dtype drift during evaluation." + "reason": "Gemma4 processor loading fails because a dependency import is missing, while the other crashes inside flash_attention on `s_aux=None`; different root causes." }, { - "left": "issue:45372", - "right": "issue:45588", + "left": "issue:43278", + "right": "issue:43381", "accept": false, - "reason": "Gemma4 processor import dependency issue vs flash-attention sink-less model crash; different call sites." + "reason": "Eval-mode gradient checkpointing restriction is unrelated to embedding dtype changes." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "Executorch export failure with training guard/erfinv is unrelated to flash_attention.py crashing on s_aux=None." + "reason": "Executorch export issues for gemma3n are a different path from a runtime flash_attention AttributeError." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "Both involve Gemma-related inference paths, but one is a flash-attention AttributeError and the other is a device-map mismatch in PLE." + "reason": "A flash_attention crash and a Gemma4 PLE device mismatch are separate bugs with different fixes." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Both concern Gemma4 processor loading, but one is an import error from mistral_common and the other is a missing _tokenizer attribute in serve." + "reason": "Missing `ReasoningEffort` import during processor loading is unrelated to a missing `_tokenizer` attribute in serve." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Generation output mismatch and lm_head.weight serialization inconsistency are different bugs with different code paths." + "reason": "Both are v5 regressions, but one is output divergence and the other is serialization of tied weights; not the same concrete defect." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility breakage vs DeepSeekV3MoE divergence from remote implementation; unrelated." + "reason": "Python version compatibility and DeepseekV3MoE behavioral divergence are unrelated." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure vs a bare-except bug in Fuyu tensor conversion; different models and fixes." + "reason": "A save_pretrained/tokenization failure is not the same as a bare-except bug swallowing exceptions." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline top_k shape change is unrelated to Trainer broadcast failing on NoneType." + "reason": "top_k output nesting in pipeline is unrelated to a Trainer GPU broadcast NoneType error." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category cardinality limit error vs paged-generate warning; no shared underlying defect." + "reason": "Category-count limits and a stale warning in paged generate() are different issues." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBirdTokenizer special-token registration bug vs Qwen3.5 recurrent-state reset in cache handling; different subsystems." - }, - { - "left": "issue:32090", - "right": "issue:33290", - "accept": false, - "reason": "Trainer broadcast NoneType error and DeepSpeed Adafactor OOM are both training-related but not the same issue." + "reason": "Tokenizer special-token registration and recurrent-state reset in a model file are unrelated." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Image feature extractor center-arg regression and video-input regression affect different preprocessing/code paths." - }, - { - "left": "issue:43611", - "right": "issue:46121", - "accept": false, - "reason": "base_model_prefix loading breakage is unrelated to rope-parameter JSON list handling." - }, - { - "left": "issue:29127", - "right": "issue:34567", - "accept": false, - "reason": "LayoutLMv3 error messaging and TrainerState token counting are distinct issues." - }, - { - "left": "issue:42490", - "right": "issue:43873", - "accept": false, - "reason": "Trailing-slash import warning is unrelated to quantization offloading not working." - }, - { - "left": "issue:33357", - "right": "issue:39692", - "accept": false, - "reason": "MacOS bus error on CLIP vs documentation example errors for SigLIP2; different failures and remediation." - }, - { - "left": "issue:44871", - "right": "issue:44977", - "accept": false, - "reason": "Gemma eos_token_id config mismatch is unrelated to Qwen3.5 flash-attention generation failure." - }, - { - "left": "issue:44186", - "right": "issue:45198", - "accept": false, - "reason": "Different tokenizers/models and different failure modes: LayoutLMv2 batching crash vs Wav2Vec2 save/tokenization failure." - }, - { - "left": "issue:45459", - "right": "issue:45712", - "accept": false, - "reason": "Tokenizer error masking by missing protobuf is unrelated to leftover dummy classes leaking into dir(transformers)." + "reason": "Unsupported feature arg in an extractor and a video-input regression in Qwen3.5 are not the same bug." }, { - "left": "issue:43278", - "right": "issue:43381", + "left": "issue:32090", + "right": "issue:33290", "accept": false, - "reason": "Duplicate of earlier verdict: different evaluation behaviors, not the same bug." + "reason": "Unsupported NoneType passed to `_gpu_broadcast_one` is unrelated to an OOM caused by Adafactor in DeepSpeed." } ] }, @@ -9059,14 +9079,16 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:27957", "issue:29127", "issue:29942", "issue:30064", "issue:33357", + "issue:34567", "issue:35141", - "issue:36010", + "issue:39692", "issue:42490", "issue:42915", "issue:42994", @@ -9077,26 +9099,24 @@ "issue:43421", "issue:43526", "issue:43575", + "issue:43611", "issue:43653", "issue:43701", "issue:43756", - "issue:43761", "issue:43824", - "issue:43931", + "issue:43873", "issue:43994", "issue:44186", "issue:44265", "issue:44279", "issue:44361", - "issue:44387", "issue:44442", "issue:44462", + "issue:44464", "issue:44625", - "issue:44857", + "issue:44871", "issue:44977", - "issue:45072", "issue:45198", - "issue:45245", "issue:45372", "issue:45412", "issue:45440", @@ -9104,273 +9124,273 @@ "issue:45491", "issue:45584", "issue:45593", + "issue:45712", "issue:45834", - "issue:45859", "issue:46121", "issue:46369", "issue:46399" ], "soft_pairs": [ "issue:43065|issue:45412", + "issue:44871|issue:44977", + "issue:44186|issue:45198", + "issue:43611|issue:46121", + "issue:42490|issue:43873", + "issue:33357|issue:39692", "issue:44361|issue:45198", + "issue:29127|issue:34567", "issue:43065|issue:43994", + "issue:45459|issue:45712", "issue:42915|issue:42994", "issue:43066|issue:44462", - "issue:43278|issue:43575", "issue:29942|issue:30064", "issue:45372|issue:45459", "issue:45412|issue:46369", + "issue:43278|issue:43575", "issue:43994|issue:45412", "issue:44265|issue:45198", "issue:43381|issue:43701", "issue:43756|issue:45440", - "issue:43421|issue:45491", "issue:43653|issue:44442", + "issue:43421|issue:45491", + "issue:27957|issue:29942", "issue:29127|issue:33357", "issue:43824|issue:44279", - "issue:45440|issue:45593", "issue:43994|issue:46369", + "issue:45440|issue:45593", "issue:43526|issue:45834", - "issue:29942|issue:36010", + "issue:44464|issue:45584", "issue:46121|issue:46399", - "issue:29127|issue:35141", "issue:43526|issue:44186", - "issue:44625|issue:45584", - "issue:43761|issue:45859", - "issue:43575|issue:43701", - "issue:42915|issue:43931", - "issue:44442|issue:45072", - "issue:42490|issue:43381", - "issue:43575|issue:44387", - "issue:44977|issue:45245", - "issue:43526|issue:44857" + "issue:29127|issue:35141", + "issue:44625|issue:45584" ], - "prepared_review_unit_hash": "2a908b2e33d8723c0ff27bfc655068f5fe110432e5e477bc5a5011521d5462f9", + "prepared_review_unit_hash": "18c624fbb8ddcf6058e491a869c7ec5be175777bcfb4d96fbb1ca41c4b90089f", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37450, - "estimated_input_tokens": 9363, - "estimated_eval_tokens": 18982 + "serialized_chars": 37492, + "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 highly heterogeneous: the apparent matches are mostly same-subsystem or same-error-class similarities, not true duplicates. I would not merge any of the soft-edge pairs. If one issue must serve as the anchor, the cleanest and most representative core-path bug is the AutoTokenizer tokenizer.json loading regression.", - "confidence": 0.79, - "canonical_issue_reason": "issue:44462 is the clearest, most actionable core-path regression in the set and has the cleanest repro/impact story, making it the best anchor for this otherwise mixed cluster.", + "summary": "The cluster is very heterogeneous; most soft pairs are only keyword-similar (often sharing a subsystem like tokenizers, quantization, or attention) but do not describe the same underlying bug. No PRs are present.", + "confidence": 0.69, + "canonical_issue_reason": "issue:44462 is the cleanest, most self-contained regression in the set (AutoTokenizer ignoring tokenizer.json), while most other issues are unrelated, model-specific bugs.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44462 is the strongest single issue candidate overall: well-scoped, user-facing, and centered on tokenizer loading behavior that is easy to triage and fix.", + "best_issue_reason": "issue:44462 has the clearest repro and narrow scope, making it the best representative anchor for this otherwise mixed cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Different bugs: a dummy Conv2d placeholder in Sam3PixelDecoder vs an RT-DETR memory leak on deletion." + "reason": "Different problems: a dummy Conv2d placeholder in Sam3PixelDecoder vs RT-DETR memory not being released." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Same broad area only; eos_token_id mismatch in Gemma-3 is unrelated to Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizers and failure modes: LayoutLMv2 NER/padding crash vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Unrelated code paths: base_model_prefix loading regression vs rope validation type error from JSON-loaded lists." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Path-with-slash import warning and quantization offloading behavior are different bugs." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error with a community CLIP model is not the same as SigLIP2 documentation example errors." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "Different tokenizer/model failures: MLukeTokenizer AttributeError vs Wav2Vec2 save_pretrained/tokenization failure." + "reason": "Different tokenizers and symptoms; MLuke AttributeError is not the same as Wav2Vec2 save/tokenization breakage." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and TrainerState token accounting are unrelated issues." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Same broad vision area, but not the same bug or code path; dummy module placeholder vs incorrect inference outputs." + "reason": "Different model bugs: Sam3 pixel decoder placeholder vs SigLIP2 producing nonsensical outputs." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer exception masking and leftover dummy classes are separate maintenance bugs." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Different quantization-related issues: FineGrainedFP8Config load failure vs quantized model saving failure." + "reason": "Both mention quantization, but one is Qwen3Moe FP8 load failure and the other is quantized model saving; not the same code path." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Both concern tokenizer loading, but one is a decoder-type regression in v5 while the other is repository tokenizer.json resolution; distinct failures." - }, - { - "left": "issue:43278", - "right": "issue:43575", - "accept": false, - "reason": "Unrelated problems: embedding dtype drift during eval vs OOM when loading a large model with tensor parallelism." + "reason": "Tokenizer decoder-type migration bug is unrelated to AutoTokenizer ignoring tokenizer.json from the repo." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Different areas entirely: Flash Attention 2 test failures vs void segmentation map processing." + "reason": "Flash Attention 2 test failures and void segmentation-map processing are different subsystems." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Different tokenizer/processor failures: missing ReasoningEffort import vs protobuf-related exception masking." + "reason": "Gemma 4 processor import failure is unrelated to protobuf error masking in tokenizer decoding." }, { "left": "issue:45412", "right": "issue:46369", "accept": false, - "reason": "Different memory bugs in different models/code paths: RT-DETR GC leak vs sam3-litetext memory overflow." + "reason": "Both are memory-related, but one is a leak on deletion and the other is an OOM during sam3-litetext usage." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "BF16-to-FP32 embedding dtype change in eval is unrelated to tensor-parallel OOM on Qwen2-57B." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "Different symptoms and root causes: bad CLIP/Siglip outputs vs model deletion memory retention." + "reason": "SigLIP2 incorrect outputs and RT-DETR memory release are unrelated bugs." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "Different code paths: torch.export failure around torch_compilable_check vs Wav2Vec2 save/tokenization failure." + "reason": "torch.export failure with torch_compilable_check is not the same as Wav2Vec2 save_pretrained/tokenization breakage." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Different training-flow bugs: gradient checkpointing in eval mode vs resume_from_checkpoint key mismatch." + "reason": "Gradient checkpointing in eval and resume_from_checkpoint key mismatch are different issues." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Different model-specific regressions: Smollm3 RoPE layer handling vs DeepseekV3MoE divergence from remote implementation." + "reason": "Different model-implementation divergences: Smollm3 RoPE-layer dropping vs DeepSeekV3MoE remote/native mismatch." }, { - "left": "issue:43421", - "right": "issue:45491", + "left": "issue:43653", + "right": "issue:44442", "accept": false, - "reason": "Tokenizer post-processor update behavior vs Gemma3 NaN embeddings; unrelated failure modes." + "reason": "BigBird tokenizer special-token registration bug is not the same as AutoTokenizer failing to load FastSpeech2ConformerTokenizer." }, { - "left": "issue:43653", - "right": "issue:44442", + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Updating tokenizers post-processors at runtime is unrelated to Gemma3 NaN embeddings with sliding-window attention." + }, + { + "left": "issue:27957", + "right": "issue:29942", "accept": false, - "reason": "Both are tokenizer-related, but BigBird special-token registration and FastSpeech2ConformerTokenizer loading are distinct bugs." + "reason": "XLMRoberta Flash Attention 2 support and failing FA2 tests overlap only by topic, not by concrete bug." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "Completely different issues: LayoutLMv3 error-message clarity vs CLIP bus error on macOS." + "reason": "LayoutLMv3 error-message issue and CLIP bus error on MacOS are unrelated." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "One is a missing Qwen2.5-VL import, the other is a vague dependency issue; not the same underlying bug." + "reason": "Qwen2_5_VL import error is a specific symbol/version issue, while the other is a broad dependency problem." }, { - "left": "issue:45440", - "right": "issue:45593", + "left": "issue:43994", + "right": "issue:46369", "accept": false, - "reason": "Different model families and code paths: DeepseekV3MoE divergence vs D-FINE auxiliary-loss behavior." + "reason": "SigLIP2 wrong outputs and SAM3 memory overflow do not share the same failing code path." }, { - "left": "issue:43994", - "right": "issue:46369", + "left": "issue:45440", + "right": "issue:45593", "accept": false, - "reason": "Different models and symptoms: Siglip2 nonsensical outputs vs SAM3 memory overflow." + "reason": "DeepSeekV3MoE divergence and D-FINE auxiliary-loss behavior are different model bugs." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "Different image-processing bugs: BeitImageProcessorFast label reduction vs Kosmos2.5 OCR long-input index error." + "reason": "BeitImageProcessorFast label reduction and Kosmos2.5 long-OCR index error are unrelated image-processing failures." }, { - "left": "issue:29942", - "right": "issue:36010", + "left": "issue:44464", + "right": "issue:45584", "accept": false, - "reason": "Unrelated failures: Flash Attention 2 tests vs GenerationMixin import error." + "reason": "Compiled-forward chunked generation inconsistency is unrelated to Whisper empty-transcription failure after align_special_tokens." }, { "left": "issue:46121", "right": "issue:46399", "accept": false, - "reason": "Related to RoPE, but different concrete failures: JSON-loaded list handling vs missing decorator for apply_rotary_pos_emb registration." - }, - { - "left": "issue:29127", - "right": "issue:35141", - "accept": false, - "reason": "Different bug classes: clearer error reporting vs embeddings reinitialized during post_init after resizing." + "reason": "Rope-validation list typing and apply_rotary_pos_emb registration errors are separate implementation defects." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "Different components and symptoms: Beit image label reduction vs LayoutLMv2Tokenizer crashes on NER/batched padding." - }, - { - "left": "issue:44625", - "right": "issue:45584", - "accept": false, - "reason": "Different models and behaviors: Qwen3.5 num_labels propagation vs Whisper generation on empty transcription after token alignment." - }, - { - "left": "issue:43761", - "right": "issue:45859", - "accept": false, - "reason": "Different regressions: CLIPVisionModel hidden_states bug vs Qwen3_5Moe rotary embedding CPU-offload incompatibility." - }, - { - "left": "issue:43575", - "right": "issue:43701", - "accept": false, - "reason": "Different issues: tensor-parallel OOM on load vs checkpoint key mismatch during resume." - }, - { - "left": "issue:42915", - "right": "issue:43931", - "accept": false, - "reason": "Different Qwen model failures: FineGrainedFP8Config load issue vs weight-shape mismatch on Qwen3-VL." - }, - { - "left": "issue:44442", - "right": "issue:45072", - "accept": false, - "reason": "Tokenizer loading failure vs dtype mismatch in bfloat16 inference; no shared root cause." - }, - { - "left": "issue:42490", - "right": "issue:43381", - "accept": false, - "reason": "Unrelated: local-path trailing-slash import warning vs gradient checkpointing eval-mode limitation." + "reason": "ImageProcessor label handling and tokenizer crashes on NER inputs are different bug classes." }, { - "left": "issue:43575", - "right": "issue:44387", - "accept": false, - "reason": "Both mention memory pressure, but one is load-time OOM for a large Qwen model and the other is reserved-memory growth under int4 quantization; not the same bug." - }, - { - "left": "issue:44977", - "right": "issue:45245", + "left": "issue:29127", + "right": "issue:35141", "accept": false, - "reason": "Completely unrelated: Qwen3.5 flash-attention generation bug vs category-count limit RuntimeError." + "reason": "LayoutLMv3 error messaging and post-init embedding reinitialization are unrelated." }, { - "left": "issue:43526", - "right": "issue:44857", + "left": "issue:44625", + "right": "issue:45584", "accept": false, - "reason": "Different bugs in different subsystems: Beit image label handling vs LwDetrImageLoss AMP/CUDA crash." + "reason": "Qwen3.5 num_labels propagation and Whisper empty transcription are unrelated." } ] }, @@ -9392,290 +9412,297 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:27957", + "issue:29127", "issue:33290", "issue:34567", "issue:34689", "issue:35141", - "issue:36010", - "issue:36331", "issue:42175", "issue:42490", - "issue:43232", + "issue:42915", "issue:43262", "issue:43316", "issue:43334", "issue:43381", - "issue:43388", "issue:43452", - "issue:43493", "issue:43525", + "issue:43526", "issue:43531", + "issue:43575", "issue:43611", "issue:43646", + "issue:43701", "issue:43720", "issue:43723", "issue:43761", - "issue:43827", "issue:43873", "issue:43931", "issue:44062", "issue:44077", "issue:44079", - "issue:44188", "issue:44206", + "issue:44387", "issue:44442", "issue:44464", - "issue:44479", "issue:44743", "issue:44811", + "issue:44857", "issue:44898", - "issue:44987", + "issue:44977", "issue:45003", - "issue:45216", + "issue:45072", + "issue:45245", "issue:45440", "issue:45479", "issue:45491", "issue:45588", - "issue:45803", + "issue:45859", "issue:45907", - "issue:46178", "issue:46234" ], "soft_pairs": [ + "issue:43761|issue:45859", + "issue:44442|issue:45072", + "issue:43575|issue:43701", + "issue:42915|issue:43931", + "issue:42490|issue:43381", + "issue:43575|issue:44387", + "issue:43526|issue:44857", + "issue:44977|issue:45245", "issue:43334|issue:43531", - "issue:44464|issue:45491", "issue:43262|issue:44206", - "issue:44743|issue:44811", + "issue:44464|issue:45491", + "issue:27957|issue:29127", "issue:43452|issue:45588", + "issue:44743|issue:44811", + "issue:27957|issue:33290", "issue:43646|issue:44077", + "issue:44062|issue:45907", "issue:33290|issue:34689", - "issue:33290|issue:35141", "issue:33290|issue:34567", - "issue:44062|issue:45907", + "issue:33290|issue:35141", "issue:43381|issue:43873", "issue:44442|issue:44898", "issue:45003|issue:45440", "issue:43723|issue:45479", "issue:43761|issue:44079", "issue:42490|issue:44743", + "issue:43316|issue:43525", "issue:42175|issue:46234", "issue:34689|issue:43611", - "issue:43316|issue:43525", - "issue:43232|issue:43388", - "issue:43720|issue:44442", - "issue:43761|issue:44188", - "issue:34689|issue:36010", - "issue:43931|issue:45216", - "issue:44077|issue:44479", - "issue:33290|issue:36331", - "issue:43493|issue:43827", - "issue:44442|issue:45803", - "issue:44987|issue:46178" + "issue:43720|issue:44442" ], - "prepared_review_unit_hash": "535fecc6d3deafd8f19c991b18c6d512a39c2a255dcccf778f7ddefaa16b6e9d", + "prepared_review_unit_hash": "31f961d01637e5125352a4c445b9e3bf2cd415ea1069820fbfdcbf8f310526c6", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36685, - "estimated_input_tokens": 9172, - "estimated_eval_tokens": 18600 + "soft_pair_count": 30, + "serialized_chars": 36630, + "estimated_input_tokens": 9158, + "estimated_eval_tokens": 18572 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft pairs point to different bug reports or regressions in nearby areas, but not the same underlying issue. I would reject all soft edges and not force a canonical duplicate.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "The cluster is heterogeneous: most candidates only share broad themes like loading, memory, or config regressions, but not the same concrete bug. I would keep them separate and use the most central memory/offload regression as the representative issue.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43873 is the most discussion-heavy and broadly phrased memory/offloading regression in the set, so it is the best representative if one issue must stand in for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "43873 is the strongest global issue candidate because it has the most activity and sits closest to the recurring quantization/offloading/memory theme; however, the rest of the set is too diverse for true deduping.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43334", - "right": "issue:43531", + "left": "issue:43761", + "right": "issue:45859", "accept": false, - "reason": "Qwen3-VL pad_token_id loading failure vs Qwen3-MoE sliding_window behavior; different model families and bugs." + "reason": "Different bugs: CLIP hidden-states regression vs Qwen3 MoE CPU-offload incompatibility." }, { - "left": "issue:44464", - "right": "issue:45491", + "left": "issue:44442", + "right": "issue:45072", "accept": false, - "reason": "Chunked generation with compiled forward is unrelated to Gemma3 NaN embeddings from sliding-window/all-padding batches." + "reason": "Tokenizer loading failure vs bfloat16 dtype mismatch in inference; unrelated code paths." }, { - "left": "issue:43262", - "right": "issue:44206", + "left": "issue:43575", + "right": "issue:43701", "accept": false, - "reason": "Audio chat-template sampling-rate default mismatch vs LasrFeatureExtractor unsupported center arg; different code paths." + "reason": "OOM while loading a large model with TP vs resume_from_checkpoint key mismatch." }, { - "left": "issue:44743", - "right": "issue:44811", + "left": "issue:42915", + "right": "issue:43931", "accept": false, - "reason": "Qwen3_5 recurrent state reset bug vs Whisper batch_decode skip_special_tokens handling; not the same issue." + "reason": "FineGrainedFP8Config failure vs Qwen3-VL weight-shape mismatch; different loading problems." }, { - "left": "issue:43452", - "right": "issue:45588", + "left": "issue:42490", + "right": "issue:43381", "accept": false, - "reason": "gguf_file tokenizer/model loading break vs flash_attention s_aux=None crash; unrelated failures." + "reason": "Local-path import warning vs gradient-checkpointing-in-eval restriction; unrelated." }, { - "left": "issue:43646", - "right": "issue:44077", + "left": "issue:43575", + "right": "issue:44387", "accept": false, - "reason": "Both touch initialization, but one is custom model init regression and the other is patchtsmixer post_init policy." + "reason": "Both involve memory, but one is TP load OOM and the other is int4 reserved-memory regression." }, { - "left": "issue:33290", - "right": "issue:34689", + "left": "issue:43526", + "right": "issue:44857", "accept": false, - "reason": "Adafactor OOM in deepspeed is unrelated to TrainerState num_input_tokens_seen not updating." + "reason": "ImageProcessor label reduction bug vs AMP/CUDA crash in loss computation." }, { - "left": "issue:33290", - "right": "issue:35141", + "left": "issue:44977", + "right": "issue:45245", "accept": false, - "reason": "OOM in deepspeed optimizer vs token embedding/output reinit bug; different subsystems." + "reason": "Flash-attention generation bug vs category-count limit error; different failures." }, { - "left": "issue:33290", - "right": "issue:34567", + "left": "issue:43334", + "right": "issue:43531", "accept": false, - "reason": "Optimizer memory issue vs TrainerState token counter update bug; no shared underlying defect." + "reason": "Missing pad_token_id on Qwen3-VL config vs sliding_window behavior in Qwen3-MoE." }, { - "left": "issue:44062", - "right": "issue:45907", + "left": "issue:43262", + "right": "issue:44206", "accept": false, - "reason": "AddedToken special-arg conflict is unrelated to list_repo_files forgetting revision." + "reason": "Chat-template sample-rate default vs unsupported center arg in feature extraction." }, { - "left": "issue:43381", - "right": "issue:43873", + "left": "issue:44464", + "right": "issue:45491", "accept": false, - "reason": "Gradient checkpointing in eval mode vs quantization offloading behavior; different runtime problems." + "reason": "Compiled-forward chunked generation inconsistency vs NaNs from all-padding sliding-window attention." }, { - "left": "issue:44442", - "right": "issue:44898", + "left": "issue:27957", + "right": "issue:29127", "accept": false, - "reason": "Tokenizer load failure vs Perceiver non-default-resolution interpolation failure; not the same path." + "reason": "Flash Attention 2 support request vs a LayoutLMv3 error-message clarity issue." }, { - "left": "issue:45003", - "right": "issue:45440", + "left": "issue:43452", + "right": "issue:45588", "accept": false, - "reason": "Unsafe sys.modules access is unrelated to DeepseekV3MoE implementation divergence." + "reason": "Tokenizer/model loading via gguf_file vs flash_attention sinkless AttributeError; different subsystems." }, { - "left": "issue:43723", - "right": "issue:45479", + "left": "issue:44743", + "right": "issue:44811", "accept": false, - "reason": "AutoTokenizer loading regression vs zero-loss sequence-classification bug; different failure modes." + "reason": "Recurrent-state reset bug vs Whisper batch_decode skip_special_tokens handling." }, { - "left": "issue:43761", - "right": "issue:44079", + "left": "issue:27957", + "right": "issue:33290", "accept": false, - "reason": "CLIPVision hidden_states regression vs ModelOutput None-key assignment logic; unrelated components." + "reason": "Flash-attention model support vs deepspeed Adafactor OOM; not the same bug." }, { - "left": "issue:42490", - "right": "issue:44743", + "left": "issue:43646", + "right": "issue:44077", "accept": false, - "reason": "Trailing-slash import warning is unrelated to Qwen3_5 recurrent-state reset." + "reason": "General custom-init regression vs patchtsmixer-specific post_init policy change." }, { - "left": "issue:42175", - "right": "issue:46234", + "left": "issue:44062", + "right": "issue:45907", "accept": false, - "reason": "Missing tensorflow in a torch install extra vs missing import_utils helper symbol; different packaging/API issues." + "reason": "AddedToken special-arg duplication vs list_repo_files revision forwarding." }, { - "left": "issue:34689", - "right": "issue:43611", + "left": "issue:33290", + "right": "issue:34689", "accept": false, - "reason": "Both are loading regressions, but one is Llama 3.2 Vision breakage and the other is base_model_prefix handling in v5." + "reason": "Adafactor/deepspeed OOM vs Llama 3.2 Vision loading breakage; different root causes." }, { - "left": "issue:43316", - "right": "issue:43525", + "left": "issue:33290", + "right": "issue:34567", "accept": false, - "reason": "Gemma3TextConfig API mismatch vs Llama4Config missing pad_token_id; similar shape, but not the same bug." + "reason": "Memory/OOM issue vs TrainerState token-count update bug." }, { - "left": "issue:43232", - "right": "issue:43388", + "left": "issue:33290", + "right": "issue:35141", "accept": false, - "reason": "Generation kwargs after sync_gpus vs gather_for_metrics dropping tuple labels; different trainer/runtime code." + "reason": "OOM in deepspeed vs embedding reinitialization after resize_token_embeddings." }, { - "left": "issue:43720", - "right": "issue:44442", + "left": "issue:43381", + "right": "issue:43873", "accept": false, - "reason": "BitNet accelerate loading/unpacking bug vs FastSpeech2ConformerTokenizer loading failure; unrelated." + "reason": "Eval-mode checkpointing restriction vs quantization/offloading memory regression; related area but different bug." }, { - "left": "issue:43761", - "right": "issue:44188", + "left": "issue:44442", + "right": "issue:44898", "accept": false, - "reason": "CLIPVision hidden_states regression vs attention-kernel divergence under torch.compile; different models and causes." + "reason": "Tokenizer load failure vs Perceiver image-classification interpolation failure." }, { - "left": "issue:34689", - "right": "issue:36010", + "left": "issue:45003", + "right": "issue:45440", "accept": false, - "reason": "Model loading breakage vs GenerationMixin import error; separate regressions." + "reason": "Unsafe sys.modules access vs DeepseekV3 implementation divergence." }, { - "left": "issue:43931", - "right": "issue:45216", + "left": "issue:43723", + "right": "issue:45479", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch vs Qwen3.5 save_pretrained checkpoint corruption; both Qwen-related but different defects." + "reason": "Tokenizer loading in v5 vs degenerate zero-loss classification behavior." }, { - "left": "issue:44077", - "right": "issue:44479", + "left": "issue:43761", + "right": "issue:44079", "accept": false, - "reason": "patchtsmixer post_init restriction is unrelated to the v5.3.0 video-input regression." + "reason": "CLIP hidden_states regression vs ModelOutput key assignment bug." }, { - "left": "issue:33290", - "right": "issue:36331", + "left": "issue:42490", + "right": "issue:44743", "accept": false, - "reason": "Deepspeed OOM in optimizer use vs CustomTrainer.compute_loss signature mismatch; distinct training issues." + "reason": "Trailing-slash import warning vs recurrent-state reset in cached generation." }, { - "left": "issue:43493", - "right": "issue:43827", + "left": "issue:43316", + "right": "issue:43525", "accept": false, - "reason": "SigLIP2 implementation discrepancy vs docs still referencing pipeline(); one is model parity, the other documentation." + "reason": "Gemma3TextConfig API discrepancy vs Llama4Config missing pad_token_id; similar family, different attribute." }, { - "left": "issue:44442", - "right": "issue:45803", + "left": "issue:42175", + "right": "issue:46234", "accept": false, - "reason": "Tokenizer loading failure vs bare except swallowing errors in Fuyu tensor conversion; not duplicates." + "reason": "Package extra/backend dependency issue vs missing import_utils symbol." }, { - "left": "issue:44987", - "right": "issue:46178", + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Model-loading regression for a specific model vs base_model_prefix-related v5 loading breakage." + }, + { + "left": "issue:43720", + "right": "issue:44442", "accept": false, - "reason": "Model-loading regression on transformers>=5.1.0 vs backend mapping error for tensorflow_text; different root causes." + "reason": "BitNet packed-weight unpacking bug vs tokenizer load failure; unrelated." } ] }, @@ -9697,7 +9724,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:16998", "issue:22355", @@ -9706,268 +9733,260 @@ "issue:42617", "issue:42907", "issue:43065", + "issue:43232", "issue:43295", "issue:43316", - "issue:43404", - "issue:43441", - "issue:43540", + "issue:43388", + "issue:43493", "issue:43550", "issue:43575", "issue:43606", "issue:43643", "issue:43650", "issue:43701", - "issue:43716", - "issue:43742", "issue:43749", + "issue:43761", "issue:43825", + "issue:43827", "issue:43873", "issue:43931", "issue:44077", - "issue:44162", + "issue:44188", "issue:44190", + "issue:44442", + "issue:44464", "issue:44479", "issue:44496", - "issue:44514", - "issue:44534", - "issue:44610", "issue:44617", - "issue:44792", "issue:44855", "issue:44871", - "issue:44877", "issue:44913", + "issue:44987", "issue:45137", + "issue:45216", "issue:45290", - "issue:45584", + "issue:45632", "issue:45698", "issue:45759", "issue:45803", - "issue:46170", + "issue:46178", "issue:46202", "issue:46365", - "issue:46369" + "issue:46421", + "issue:46424" ], "soft_pairs": [ - "issue:43873|issue:44871", + "issue:43232|issue:43388", + "issue:43931|issue:45216", + "issue:43761|issue:44188", + "issue:44077|issue:44479", + "issue:44464|issue:46421", + "issue:44987|issue:46178", + "issue:43493|issue:43827", + "issue:44442|issue:45803", "issue:42907|issue:45698", "issue:44479|issue:45290", + "issue:43873|issue:44871", "issue:43749|issue:43931", "issue:43643|issue:44913", "issue:43316|issue:45759", - "issue:44855|issue:45137", "issue:44190|issue:44496", + "issue:44855|issue:45137", "issue:43650|issue:46202", + "issue:45632|issue:46424", "issue:45803|issue:46365", "issue:25251|issue:28282", "issue:43295|issue:43825", "issue:43550|issue:43701", - "issue:42617|issue:43575", "issue:43606|issue:43701", - "issue:43575|issue:43742", - "issue:16998|issue:22355", + "issue:42617|issue:43575", "issue:43065|issue:44617", - "issue:43540|issue:44514", - "issue:43316|issue:43716", - "issue:44871|issue:45584", - "issue:44610|issue:46369", - "issue:43295|issue:44871", - "issue:43316|issue:44877", - "issue:44162|issue:44792", - "issue:44077|issue:44534", - "issue:43404|issue:43441", - "issue:45803|issue:46170" + "issue:16998|issue:22355" ], - "prepared_review_unit_hash": "0a958a507953b70ab02f4aee38f70bebb59805862d1773e588121427d7d37018", + "prepared_review_unit_hash": "2db86bf115ed068c7ff84163fa651cf060f9139a4ce10bdb9e4db8c91398913b", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 35742, - "estimated_input_tokens": 8936, - "estimated_eval_tokens": 18128 + "soft_pair_count": 26, + "serialized_chars": 35402, + "estimated_input_tokens": 8851, + "estimated_eval_tokens": 17958 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: the soft pairs mostly connect unrelated transformer bugs that share broad themes like loading, memory, or multimodal regressions but not the same concrete code path. No soft edge should be merged.", - "confidence": 0.98, - "canonical_issue_reason": "issue:43873 is the broadest and most actively discussed open report in this set, with multiple inbound references; it is the best single representative even though the cluster itself is not a true duplicate group.", + "summary": "This cluster is a loose basket of unrelated Transformers issues; none of the soft pairs look like true duplicates. The closest matches are still different bug classes or different code paths.", + "confidence": 0.95, + "canonical_issue_reason": "If one issue must serve as the cluster anchor, issue 43873 is the broadest and most central by activity/inbound refs, but it is not a duplicate of the others.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the strongest global issue candidate because it has the highest apparent triage value (open, discussed, referenced) and sits on a broadly relevant offloading/quantization topic.", + "best_issue_reason": "Issue 43873 is the best global issue anchor because it is active, broadly scoped, and has the strongest discussion/reference signal in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43873", - "right": "issue:44871", - "accept": false, - "reason": "Offloading/quantization behavior vs Gemma eos_token_id mismatch are different bugs and different code paths." - }, - { - "left": "issue:42907", - "right": "issue:45698", + "left": "issue:43232", + "right": "issue:43388", "accept": false, - "reason": "Saving dequantized models is unrelated to loading the wrong custom module after save_pretrained." + "reason": "Generation kwarg syncing after sync_gpus vs metric gathering truncating tuple labels; different subsystems and failure modes." }, { - "left": "issue:44479", - "right": "issue:45290", + "left": "issue:43931", + "right": "issue:45216", "accept": false, - "reason": "Video-input regression across VL models is not the same as apply_chat_template crashing on tool-call assistant messages." + "reason": "Model weight-shape mismatch on load vs save_pretrained regression for Qwen3.5 checkpoints; related family, different bug." }, { - "left": "issue:43749", - "right": "issue:43931", + "left": "issue:43761", + "right": "issue:44188", "accept": false, - "reason": "FSDP CPU RAM efficient loading and a Qwen3-VL weight-shape mismatch are distinct loading failures." + "reason": "CLIPVision hidden_states regression vs attention-kernel divergence under torch.compile; unrelated code paths." }, { - "left": "issue:43643", - "right": "issue:44913", + "left": "issue:44077", + "right": "issue:44479", "accept": false, - "reason": "Missing fields from trust_remote_code AutoConfig loading is unrelated to GPTNeoXConfig rotary_pct not persisting on reload." + "reason": "Optional post_init policy for patchtsmixer vs video-input regression in Qwen VL models; no shared concrete defect." }, { - "left": "issue:43316", - "right": "issue:45759", + "left": "issue:44464", + "right": "issue:46421", "accept": false, - "reason": "Gemma3TextConfig API discrepancy is not the same as AutoModelForCausalLM failing to unwrap text_config for Qwen multimodal configs." + "reason": "Both mention chunked/compiled behavior, but one is wrong outputs and the other is spurious recompilation from lazy cache init." }, { - "left": "issue:44855", - "right": "issue:45137", + "left": "issue:44987", + "right": "issue:46178", "accept": false, - "reason": "Python 3.13 IndentationError in DebertaV2Model parsing is unrelated to DeepSpeed ZeRO3 deque underflow." + "reason": "Loading a specific model under transformers>=5.1 vs backend mapping missing tensorflow_text; unrelated failures." }, { - "left": "issue:44190", - "right": "issue:44496", + "left": "issue:43493", + "right": "issue:43827", "accept": false, - "reason": "Local dataset loading in a training script and an unrecognized model/config.json error are unrelated." + "reason": "SigLIP2 implementation mismatch vs docs still referencing removed pipeline APIs; one is model parity, the other documentation." }, { - "left": "issue:43650", - "right": "issue:46202", + "left": "issue:44442", + "right": "issue:45803", "accept": false, - "reason": "An empty-data issue does not match Gemma pipeline roundtrip behavior on the underscore token." + "reason": "Tokenizer load failure for FastSpeech2ConformerTokenizer vs bare except swallowing errors in Fuyu batch conversion; unrelated bugs." }, { - "left": "issue:45803", - "right": "issue:46365", + "left": "issue:42907", + "right": "issue:45698", "accept": false, - "reason": "Bare except in FuyuBatchFeature.convert_to_tensors() is unrelated to Trainer import failure caused by datasets installation." + "reason": "Save/dequantization problem for specific models vs wrong custom module loaded after save/load; both custom-model related but not the same bug." }, { - "left": "issue:25251", - "right": "issue:28282", + "left": "issue:44479", + "right": "issue:45290", "accept": false, - "reason": "Pipeline top_k nesting behavior is unrelated to AutoModel/PyTorch missing-library import errors." + "reason": "Video-input regression for several Qwen VL models vs chat-template crash on tool-call assistant messages; different APIs and code paths." }, { - "left": "issue:43295", - "right": "issue:43825", + "left": "issue:43873", + "right": "issue:44871", "accept": false, - "reason": "Processor.tokenizer regression and an incorrect translation-task error message are different API regressions." + "reason": "Quantization/offloading behavior vs Gemma eos_token_id config inconsistency; no duplicate evidence." }, { - "left": "issue:43550", - "right": "issue:43701", + "left": "issue:43749", + "right": "issue:43931", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to resume_from_checkpoint key mismatch." + "reason": "Broken FSDP CPU RAM-efficient loading vs a model weight-shape mismatch at load time; different root causes." }, { - "left": "issue:42617", - "right": "issue:43575", + "left": "issue:43643", + "right": "issue:44913", "accept": false, - "reason": "3d_parallel.py runtime issues are not the same as Qwen2-57B-A14B-Instruct loading OOM under tensor parallelism." + "reason": "Missing fields from trust_remote_code AutoConfig result vs rotary_pct not persisting on reload; both config-related but distinct." }, { - "left": "issue:43606", - "right": "issue:43701", + "left": "issue:43316", + "right": "issue:45759", "accept": false, - "reason": "CPU offload device mismatch and resume_from_checkpoint key mismatch are separate bugs." + "reason": "Gemma3TextConfig API mismatch vs AutoModelForCausalLM.from_config not unwrapping text_config for Qwen multimodal configs; different model families and fixes." }, { - "left": "issue:43575", - "right": "issue:43742", + "left": "issue:44190", + "right": "issue:44496", "accept": false, - "reason": "TP OOM when loading Qwen2-57B-A14B-Instruct is unrelated to a key error loading MobileLLM-125M." + "reason": "Local dataset loading failure in an example script vs unrecognized model_type in config.json; unrelated." }, { - "left": "issue:16998", - "right": "issue:22355", + "left": "issue:44855", + "right": "issue:45137", "accept": false, - "reason": "Model max length behavior for DeBERTa-V3 has nothing to do with a missing transformers.onnx module." + "reason": "Python 3.13 import parsing error from torch.jit.script comment placement vs DeepSpeed ZeRO3 deque underflow; unrelated." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:43650", + "right": "issue:46202", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same as CUDA OOM in Sam3Video." + "reason": "Placeholder issue vs Gemma pipeline roundtrip failure on a specific token; clearly unrelated." }, { - "left": "issue:43540", - "right": "issue:44514", + "left": "issue:45632", + "right": "issue:46424", "accept": false, - "reason": "Qwen3OmniMoe video processing ValueError and Qwen2_5_VL batched chat-template crash are different processor bugs." + "reason": "trust_remote_code cache-path collision for local models vs cache max_cache_length only honored for quantized cache; different cache layers." }, { - "left": "issue:43316", - "right": "issue:43716", + "left": "issue:45803", + "right": "issue:46365", "accept": false, - "reason": "Gemma3TextConfig API discrepancy is unrelated to Mistral-3 image-preprocessor dtype mismatch." + "reason": "Bare except in Fuyu tensor conversion vs importing Trainer broken until datasets is uninstalled; unrelated." }, { - "left": "issue:44871", - "right": "issue:45584", + "left": "issue:25251", + "right": "issue:28282", "accept": false, - "reason": "Gemma eos_token_id inconsistency is unrelated to Whisper empty-transcription generation after align_special_tokens." + "reason": "Pipeline top_k nesting bug vs AutoModel requiring PyTorch not installed; no overlap." }, { - "left": "issue:44610", - "right": "issue:46369", + "left": "issue:43295", + "right": "issue:43825", "accept": false, - "reason": "OmDet-Turbo input-size mismatch and Sam3-litetext memory overflow are not the same underlying defect." + "reason": "Custom code broke around processor.tokenizer/image passing vs v5 pipeline error-message wording; different regressions." }, { - "left": "issue:43295", - "right": "issue:44871", + "left": "issue:43550", + "right": "issue:43701", "accept": false, - "reason": "processor.tokenizer regression and Gemma eos_token_id mismatch do not share the same code path." + "reason": "torch.compile + SDPA failure in Bamba vs resume_from_checkpoint key mismatch; unrelated." }, { - "left": "issue:44162", - "right": "issue:44792", + "left": "issue:43606", + "right": "issue:43701", "accept": false, - "reason": "ESM2 breakage and janus test failure are unrelated model-specific issues." + "reason": "CPU offload device mismatch in bark-small vs checkpoint resume key mismatch; unrelated." }, { - "left": "issue:44077", - "right": "issue:44534", + "left": "issue:42617", + "right": "issue:43575", "accept": false, - "reason": "PatchTSMixer post_init policy and non-persistent buffer junk are different v5 concerns." + "reason": "3d_parallel.py execution failure vs tensor-parallel OOM when loading a large Qwen2 model; not the same bug." }, { - "left": "issue:43404", - "right": "issue:43441", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "lm_head weight tying in Mistral3 is not the same as FlashAttention failure in Ministral-3." + "reason": "Dummy Conv2d in Sam3PixelDecoder vs Sam3Video CUDA OOM; same product area, but different concrete issues." }, { - "left": "issue:45803", - "right": "issue:46170", + "left": "issue:16998", + "right": "issue:22355", "accept": false, - "reason": "Fuyu tensor conversion exception handling is unrelated to Windows ROCm import/dependency issues." + "reason": "Model_max_length question for DeBERTa-V3 vs missing transformers.onnx import; unrelated." } ] }, @@ -9989,296 +10008,278 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:27957", "issue:32090", - "issue:35141", "issue:41950", - "issue:42490", + "issue:43295", + "issue:43316", "issue:43388", + "issue:43404", + "issue:43441", "issue:43452", "issue:43475", "issue:43504", - "issue:43525", - "issue:43531", + "issue:43540", "issue:43572", + "issue:43575", "issue:43646", "issue:43673", "issue:43688", + "issue:43716", + "issue:43742", "issue:43761", - "issue:43827", - "issue:43931", - "issue:43994", - "issue:44060", - "issue:44190", - "issue:44242", - "issue:44464", + "issue:44077", + "issue:44162", "issue:44479", "issue:44496", + "issue:44514", "issue:44534", + "issue:44610", "issue:44617", "issue:44734", "issue:44743", "issue:44792", - "issue:44855", "issue:44871", + "issue:44877", "issue:44913", - "issue:44964", - "issue:44977", "issue:45127", "issue:45200", - "issue:45381", "issue:45406", "issue:45542", - "issue:45563", "issue:45584", - "issue:45588", - "issue:45600", "issue:45698", + "issue:45803", "issue:46050", + "issue:46170", "issue:46178", "issue:46234", - "issue:46369" + "issue:46369", + "issue:46421" ], "soft_pairs": [ + "issue:43575|issue:43742", + "issue:43540|issue:44514", + "issue:27957|issue:32090", + "issue:43316|issue:43716", + "issue:44610|issue:46369", + "issue:43316|issue:44877", + "issue:43295|issue:44871", + "issue:44871|issue:45584", + "issue:44162|issue:44792", + "issue:43404|issue:43441", + "issue:44077|issue:44534", + "issue:45803|issue:46170", "issue:44871|issue:46050", + "issue:41950|issue:43475", "issue:43761|issue:44479", - "issue:43646|issue:44913", "issue:43452|issue:44496", - "issue:41950|issue:43475", + "issue:43646|issue:44913", "issue:43388|issue:43688", "issue:45127|issue:45200", "issue:43673|issue:44743", - "issue:32090|issue:35141", "issue:43572|issue:46178", - "issue:43504|issue:45698", "issue:44617|issue:45584", + "issue:43504|issue:45698", + "issue:43673|issue:46421", "issue:44734|issue:45406", - "issue:45542|issue:46234", - "issue:44792|issue:44977", - "issue:43525|issue:43531", - "issue:45563|issue:45584", - "issue:43761|issue:44242", - "issue:45381|issue:46369", - "issue:43931|issue:44060", - "issue:44792|issue:45600", - "issue:44534|issue:44913", - "issue:44190|issue:44964", - "issue:35141|issue:42490", - "issue:44479|issue:44734", - "issue:44464|issue:45563", - "issue:44855|issue:45588", - "issue:43827|issue:43994" + "issue:45542|issue:46234" ], - "prepared_review_unit_hash": "91d28e8597fadad82008caf7a3be97054c3c712111e556ca8506eea7cea84ee7", + "prepared_review_unit_hash": "a485d0044123188b4c3718aefe92f05643deeb5f249902b847776323186df492", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36380, - "estimated_input_tokens": 9095, - "estimated_eval_tokens": 18446 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35010, + "estimated_input_tokens": 8753, + "estimated_eval_tokens": 17762 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of superficially similar but distinct regressions; I don\u2019t see any true duplicate PRs, and none of the soft pairs look mergeable as the same bug/change. The broadest issue is the generic v5 initialization regression, but it is only an umbrella candidate, not a confirmed duplicate match.", - "confidence": 0.43, - "canonical_issue_reason": "issue:43646 is the broadest, most umbrella-like report in the set (\u201cTransformers 5.0.0 breaks custom model initialization\u201d), so it is the best single representative even though the cluster is not a true duplicate set.", + "summary": "These issues are largely unrelated across different models and subsystems. The only plausible duplicate pair is the chunked-prefill cache regression in the generation path; the rest differ in concrete code path, model, or failure mode.", + "confidence": 0.83, + "canonical_issue_reason": "issue:46421 is the best canonical representative for the chunked-prefill cache regression because it names the underlying lazy-initialization problem directly; issue:43673 reads more like the narrower symptom report (missing cache during chunked prefill).", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 best captures the common v5 init/load regression theme across several nearby reports; it is the most generally reusable reference point among these issues.", + "best_issue_reason": "issue:46421 is the strongest issue candidate overall among the soft-edge set: it is specific, actionable, and captures the shared generation/cache code path more clearly than the paired report.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44871", - "right": "issue:46050", + "left": "issue:43575", + "right": "issue:43742", "accept": false, - "reason": "One is an eos_token_id config mismatch, the other is a quantized-model save/load failure; same area, different bug." + "reason": "Both are loading failures, but they involve different models and different root causes (TP OOM vs config key error)." }, { - "left": "issue:43761", - "right": "issue:44479", + "left": "issue:43540", + "right": "issue:44514", "accept": false, - "reason": "Different models and failure modes: CLIPVision hidden-states handling vs Qwen2.5-VL video regression." + "reason": "Both involve Qwen multimodal/video processing, but one is a video-input processing bug and the other is a batched apply_chat_template/padding bug." }, { - "left": "issue:43646", - "right": "issue:44913", + "left": "issue:27957", + "right": "issue:32090", "accept": false, - "reason": "Both involve config/reload behavior, but one is a broad v5 init breakage and the other is a GPTNeoX rotary_pct persistence bug." + "reason": "Unrelated subsystems: FlashAttention support for XLM-Roberta vs Trainer GPU broadcast TypeError." }, { - "left": "issue:43452", - "right": "issue:44496", + "left": "issue:43316", + "right": "issue:43716", "accept": false, - "reason": "Different loading problems: gguf_file path handling vs unrecognized custom model/config.json model_type." + "reason": "Different models and different issues: Gemma3TextConfig API mismatch vs Mistral-3 image preprocessor dtype mismatch." }, { - "left": "issue:41950", - "right": "issue:43475", + "left": "issue:44610", + "right": "issue:46369", "accept": false, - "reason": "Video-classification processor lookup and SAM3 video attribute access are unrelated bugs." + "reason": "Both are SAM/vision-related, but one is a processor size mismatch and the other is a memory overflow; not the same bug." }, { - "left": "issue:43388", - "right": "issue:43688", + "left": "issue:43316", + "right": "issue:44877", "accept": false, - "reason": "Metrics batch truncation and auxiliary-loss normalization are different training-path issues." + "reason": "Both are config-related, but they concern different models and different failures (API discrepancy vs strict config loading)." }, { - "left": "issue:45127", - "right": "issue:45200", + "left": "issue:43295", + "right": "issue:44871", "accept": false, - "reason": "LoRA merge/vocab extension collapse is unrelated to Gemma 4 mm_token_type_ids defaults." + "reason": "A regression in processor.tokenizer/image handling is unrelated to the Gemma-3 eos_token_id config inconsistency." }, { - "left": "issue:43673", - "right": "issue:44743", + "left": "issue:44871", + "right": "issue:45584", "accept": false, - "reason": "Both touch cache/state handling, but they are separate code paths and model behaviors." + "reason": "Different models and symptoms: Gemma-3 EOS config mismatch vs Whisper empty transcription failure." }, { - "left": "issue:32090", - "right": "issue:35141", + "left": "issue:44162", + "right": "issue:44792", "accept": false, - "reason": "Trainer GPU broadcast NoneType error and token-embedding reinit on post_init are distinct." + "reason": "ESM2 workflow breakage is unrelated to the janus image-generation test failure." }, { - "left": "issue:43572", - "right": "issue:46178", + "left": "issue:43404", + "right": "issue:43441", "accept": false, - "reason": "StableLm pad_token_idx regression and tensorflow_text backend registration are unrelated." + "reason": "Same general Mistral family, but the concrete bugs differ: weight tying vs FlashAttention compatibility." }, { - "left": "issue:43504", - "right": "issue:45698", - "accept": false, - "reason": "Legacy preset loading for BEiT segmentation is not the same as loading the wrong custom module after save_pretrained." - }, - { - "left": "issue:44617", - "right": "issue:45584", - "accept": false, - "reason": "CUDA OOM in SAM3Video is not the same as Whisper empty-transcription failure after align_special_tokens." - }, - { - "left": "issue:44734", - "right": "issue:45406", - "accept": false, - "reason": "KV-cache continuation tensor indexing and missing _tokenizer on Gemma4Processor are separate serve crashes." - }, - { - "left": "issue:45542", - "right": "issue:46234", + "left": "issue:44077", + "right": "issue:44534", "accept": false, - "reason": "TensorBoard/TensorFlow backend confusion is unrelated to the missing is_falcon_mamba_ssm_available helper." + "reason": "Optional post_init on patchtsmixer is unrelated to non-persistent buffer initialization in v5." }, { - "left": "issue:44792", - "right": "issue:44977", + "left": "issue:45803", + "right": "issue:46170", "accept": false, - "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation bug are different model-path issues." + "reason": "Different bugs and platforms: bare except in Fuyu tensor conversion vs Windows/ROCm torch.distributed import failure." }, { - "left": "issue:43525", - "right": "issue:43531", + "left": "issue:44871", + "right": "issue:46050", "accept": false, - "reason": "Missing pad_token_id on Llama4Config and Qwen3-MoE sliding_window behavior are unrelated." + "reason": "Gemma-3 EOS token inconsistency is unrelated to quantized model save/load breakage after sinq." }, { - "left": "issue:45563", - "right": "issue:45584", + "left": "issue:41950", + "right": "issue:43475", "accept": false, - "reason": "A stale warning for num_return_sequences is not the same as Whisper failing on empty transcription." + "reason": "Video-classification processor lookup and SAM3 video missing attribute are different code paths and failures." }, { "left": "issue:43761", - "right": "issue:44242", + "right": "issue:44479", "accept": false, - "reason": "CLIPVision hidden_states regression and MoE load-balancing loss omission are different bugs." + "reason": "CLIPVision hidden_states regression is unrelated to Qwen VL video-input regression." }, { - "left": "issue:45381", - "right": "issue:46369", + "left": "issue:43452", + "right": "issue:44496", "accept": false, - "reason": "Qwen2.5-VL video position ids and SAM3-LiteText memory overflow are unrelated." + "reason": "Both are model-loading issues, but one is gguf/from_pretrained handling and the other is a missing model_type in config.json." }, { - "left": "issue:43931", - "right": "issue:44060", + "left": "issue:43646", + "right": "issue:44913", "accept": false, - "reason": "Weight-shape mismatch on model load and incorrect tied-weights warning are different Qwen3-VL/Next issues." + "reason": "Transformers 5 custom model initialization breakage is broader and different from GPTNeoXConfig rotary_pct not persisting on reload." }, { - "left": "issue:44792", - "right": "issue:45600", + "left": "issue:43388", + "right": "issue:43688", "accept": false, - "reason": "Janus generate-image test failure and removed Sam3LiteText auto-mapping references are not the same bug." + "reason": "Different training/eval bugs: gather_for_metrics tuple truncation vs auxiliary-loss normalization." }, { - "left": "issue:44534", - "right": "issue:44913", + "left": "issue:45127", + "right": "issue:45200", "accept": false, - "reason": "Non-persistent buffer junk in v5 is a different regression from GPTNeoX rotary_pct not persisting on reload." + "reason": "LoRA merge collapse with tied embeddings is unrelated to Gemma 4 mm_token_type_ids defaults." }, { - "left": "issue:44190", - "right": "issue:44964", + "left": "issue:43673", + "right": "issue:44743", "accept": false, - "reason": "Local dataset loading and Phi-4 multimodal model loading are unrelated failures." + "reason": "Both mention cache/recurrent state, but one is a GenerationMixin chunked-prefill cache initialization issue and the other is a model-specific recurrent-state reset bug." }, { - "left": "issue:35141", - "right": "issue:42490", + "left": "issue:43572", + "right": "issue:46178", "accept": false, - "reason": "Embedding reinit on post_init is unrelated to the local-path trailing-slash import warning." + "reason": "StableLm pad_token_idx regression is unrelated to the tensorflow_text backend registration error." }, { - "left": "issue:44479", - "right": "issue:44734", + "left": "issue:44617", + "right": "issue:45584", "accept": false, - "reason": "Qwen video-input regression and serve KV-cache indexing crash are different code paths." + "reason": "Sam3Video OOM and Whisper empty transcription failure are unrelated symptoms in different models." }, { - "left": "issue:44464", - "right": "issue:45563", + "left": "issue:43504", + "right": "issue:45698", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is not the same as a stale generate warning." + "reason": "Legacy-field preset loading and wrong custom-module selection after save_pretrained are both load-time issues, but they are different bugs." }, { - "left": "issue:44855", - "right": "issue:45588", - "accept": false, - "reason": "Python 3.13 torch.jit parsing failure and flash-attention s_aux=None crash are unrelated." + "left": "issue:43673", + "right": "issue:46421", + "accept": true, + "reason": "Both describe the same chunked-prefill cache initialization regression; one reports a missing cache, the other the lazy-init/recompilation consequence, and one fix could plausibly address both." }, { - "left": "issue:43827", - "right": "issue:43994", + "left": "issue:44734", + "right": "issue:45406", "accept": false, - "reason": "Docs still referencing pipeline() and SigLIP2 producing nonsense via pipeline/AutoModel are not the same change." + "reason": "KV-cache continuation tensor indexing and Gemma4Processor missing _tokenizer are unrelated serving bugs." }, { - "left": "issue:45588", - "right": "issue:46369", + "left": "issue:45542", + "right": "issue:46234", "accept": false, - "reason": "Flash-attention s_aux=None attribute crash and SAM3-LiteText memory overflow are different problems." + "reason": "TensorBoard/TensorFlow backend detection and missing is_falcon_mamba_ssm_available are unrelated import-utils issues." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft-edge pairs. The only accepted duplicate pair, issue:43673 and issue:46421, is well supported by the shared chunked-prefill cache regression theme; the rest of the false verdicts are reasonably differentiated by model, subsystem, or failure mode. " + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -10293,276 +10294,283 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ - "issue:36246", - "issue:41628", + "issue:25251", + "issue:27957", + "issue:35141", "issue:41720", - "issue:42175", + "issue:42490", "issue:43065", - "issue:43116", - "issue:43295", - "issue:43335", - "issue:43404", + "issue:43525", + "issue:43531", "issue:43540", "issue:43550", "issue:43644", - "issue:43650", + "issue:43761", + "issue:43827", + "issue:43931", "issue:43976", + "issue:43994", + "issue:44060", "issue:44077", - "issue:44112", - "issue:44186", "issue:44190", - "issue:44220", - "issue:44279", + "issue:44242", "issue:44462", + "issue:44464", + "issue:44479", "issue:44521", + "issue:44534", "issue:44610", "issue:44617", "issue:44734", - "issue:44737", + "issue:44792", "issue:44855", "issue:44877", "issue:44913", "issue:44933", - "issue:44938", + "issue:44964", + "issue:44977", "issue:45127", - "issue:45216", "issue:45245", - "issue:45276", "issue:45357", "issue:45362", "issue:45381", - "issue:45542", + "issue:45563", "issue:45584", + "issue:45588", "issue:45600", "issue:45632", - "issue:45803", "issue:45907", - "issue:46082", - "issue:46121", "issue:46234", "issue:46369" ], "soft_pairs": [ - "issue:41720|issue:45245", + "issue:43525|issue:43531", + "issue:45381|issue:46369", + "issue:43761|issue:44242", + "issue:44792|issue:44977", + "issue:45563|issue:45584", + "issue:43931|issue:44060", + "issue:44792|issue:45600", + "issue:44190|issue:44964", + "issue:44534|issue:44913", + "issue:44479|issue:44734", + "issue:35141|issue:42490", + "issue:44855|issue:45588", "issue:44617|issue:46369", + "issue:43827|issue:43994", + "issue:41720|issue:45245", "issue:43065|issue:44610", "issue:44521|issue:45381", + "issue:44464|issue:45563", "issue:44933|issue:45362", "issue:43644|issue:44913", "issue:43976|issue:44855", "issue:43550|issue:44190", "issue:44077|issue:44877", - "issue:45127|issue:45357", "issue:43540|issue:44734", - "issue:45632|issue:46234", + "issue:45127|issue:45357", "issue:44462|issue:45907", - "issue:44112|issue:45600", - "issue:44737|issue:45803", - "issue:43335|issue:44190", - "issue:43116|issue:44190", - "issue:45276|issue:45584", - "issue:43295|issue:45276", - "issue:44938|issue:45542", - "issue:46082|issue:46121", - "issue:45245|issue:45584", - "issue:36246|issue:41628", - "issue:43404|issue:45216", - "issue:42175|issue:44279", - "issue:43650|issue:44220", - "issue:44186|issue:44734" + "issue:25251|issue:27957", + "issue:45632|issue:46234" ], - "prepared_review_unit_hash": "a93c11792068d665be18742d8d2ce31d5cab55dc4854b9246125aa92061f1cc6", + "prepared_review_unit_hash": "756b80033388320a3318424fd8e44e8c06af536ad67239cd022469c56843969b", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35782, - "estimated_input_tokens": 8946, - "estimated_eval_tokens": 18148 + "soft_pair_count": 28, + "serialized_chars": 35901, + "estimated_input_tokens": 8976, + "estimated_eval_tokens": 18208 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most items are unrelated issues spanning different models, configs, import/export regressions, and CI/install problems. None of the soft pairs look like the same underlying bug or a mergeable duplicate PR case.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue is a safe duplicate target here; the issues are largely unrelated, so choosing one would collapse distinct bugs into one bucket.", + "summary": "This set is a mixed bag of unrelated Transformers bugs and regressions. None of the soft-similarity pairs look like the same underlying issue/change, so no duplicates should be merged here. The most canonical standalone issue is the multimodal chat-template assistant-mask regression, which has clear repros and inbound references.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44521 is the strongest standalone bug report: it is current, well-scoped, has multiple inbound references, and describes a concrete multimodal regression with clear user impact.", "canonical_pr_reason": null, - "best_issue_reason": "No issue stands out as a global canonical representative because the cluster does not form one coherent bug family.", + "best_issue_reason": "issue:44521 is the best representative issue because it has the clearest bug description, the highest apparent external interest, and is actionable without needing broader context.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41720", - "right": "issue:45245", + "left": "issue:43525", + "right": "issue:43531", "accept": false, - "reason": "Different failures: Qwen3 auto device mapping cudaErrorAssert vs category-cardinality runtime error." + "reason": "Both involve Qwen3-family models, but one is a missing `pad_token_id` attribute and the other is a `sliding_window` behavior problem; different failure modes and code paths." }, { - "left": "issue:44617", + "left": "issue:45381", "right": "issue:46369", "accept": false, - "reason": "Both are memory-related, but they affect different Sam3 variants and likely different code paths." + "reason": "Qwen2.5-VL video `vision_position_ids` corruption and Sam3-litetext memory overflow are unrelated multimodal issues with different symptoms and fixes." }, { - "left": "issue:43065", - "right": "issue:44610", + "left": "issue:43761", + "right": "issue:44242", "accept": false, - "reason": "Sam3PixelDecoder dummy conv vs OmDet-Turbo processor size mismatch are unrelated bugs." + "reason": "CLIPVision hidden-states regression and MoE load-balancing loss omission are distinct model behaviors, not the same bug." }, { - "left": "issue:44521", - "right": "issue:45381", + "left": "issue:44792", + "right": "issue:44977", "accept": false, - "reason": "Both are multimodal Qwen issues, but assistant mask generation and video position IDs are different defects." + "reason": "A janus image-generation test failure and Qwen3.5 flash-attention generation failure are different model-specific issues." }, { - "left": "issue:44933", - "right": "issue:45362", + "left": "issue:45563", + "right": "issue:45584", "accept": false, - "reason": "Nonexistent image_utils import is unrelated to Qwen3.5 chat crash behavior." + "reason": "A stale `num_return_sequences` warning in paged generate is unrelated to Whisper empty-transcription failure after token alignment." }, { - "left": "issue:43644", - "right": "issue:44913", + "left": "issue:43931", + "right": "issue:44060", "accept": false, - "reason": "Different config/state issues: junk buffers on load vs rotary_pct not persisting on reload." + "reason": "Shape mismatch when loading Qwen3-VL weights and an incorrect tied-weights warning in Qwen3-Next are separate loading/config bugs." }, { - "left": "issue:43976", - "right": "issue:44855", + "left": "issue:44792", + "right": "issue:45600", "accept": false, - "reason": "Python version compatibility problem vs DebertaV2 JIT parsing/import syntax issue." + "reason": "A janus test failure and broken `auto_mappings.py` references to removed configs are not the same underlying problem." }, { - "left": "issue:43550", - "right": "issue:44190", + "left": "issue:44190", + "right": "issue:44964", "accept": false, - "reason": "Bamba torch.compile/SDPA failure is unrelated to local dataset loading in an example script." + "reason": "Local dataset loading in an example script and model loading for Phi-4 multimodal instruct are unrelated." }, { - "left": "issue:44077", - "right": "issue:44877", + "left": "issue:44534", + "right": "issue:44913", "accept": false, - "reason": "Both involve config handling, but patchtsmixer post_init and granite_speech strict loading are different bugs." + "reason": "Non-persistent buffer junk initialization and GPTNeoX `rotary_pct` reset on reload affect different subsystems." }, { - "left": "issue:45127", - "right": "issue:45357", + "left": "issue:44479", + "right": "issue:44734", "accept": false, - "reason": "Both are regressions, but LoRA/vocab collapse and wrong visual encoder keys are distinct save/weight issues." + "reason": "Both are generation-related, but one is a video-input regression in Qwen multimodal models and the other is a KV-cache continuation indexing crash in `transformers serve`." }, { - "left": "issue:43540", - "right": "issue:44734", + "left": "issue:35141", + "right": "issue:42490", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError and serve KV-cache indexing are different code paths." + "reason": "Output embedding reinitialization after resize and import warnings from trailing-slash local paths are unrelated bugs." }, { - "left": "issue:45632", - "right": "issue:46234", + "left": "issue:44855", + "right": "issue:45588", "accept": false, - "reason": "Cache-path collision for trust_remote_code is unrelated to a missing import-utils helper." + "reason": "A Python 3.13 `torch.jit.script` parsing issue and a flash-attention `AttributeError` on `s_aux=None` are different failure modes." }, { - "left": "issue:44462", - "right": "issue:45907", + "left": "issue:44617", + "right": "issue:46369", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json and list_repo_files revision forwarding are unrelated repo/file handling bugs." + "reason": "Sam3Video OOM and Sam3-litetext memory overflow are both memory issues, but they concern different models and likely different causes." }, { - "left": "issue:44112", - "right": "issue:45600", + "left": "issue:43827", + "right": "issue:43994", "accept": false, - "reason": "Both broke CI, but stale GraniteSpeech test and removed Sam3LiteText config references are different failures." + "reason": "Docs still referencing removed `pipeline()` APIs is not the same as SigLIP2 producing nonsensical outputs through AutoModel/pipeline." }, { - "left": "issue:44737", - "right": "issue:45803", + "left": "issue:41720", + "right": "issue:45245", "accept": false, - "reason": "XLNet device placement bug and bare-except error swallowing are unrelated." + "reason": "Qwen3 auto device mapping CUDA assert and the 2^24 category limit error are unrelated problems." }, { - "left": "issue:43335", - "right": "issue:44190", + "left": "issue:43065", + "right": "issue:44610", "accept": false, - "reason": "SwitchTransformers config bug and image-classification local dataset loading are unrelated." + "reason": "A dummy `nn.Conv2d` in Sam3PixelDecoder and a processor/model input-size mismatch are different issues, even within Sam3." }, { - "left": "issue:43116", - "right": "issue:44190", + "left": "issue:44521", + "right": "issue:45381", "accept": false, - "reason": "Example multi-label classification output bug is unrelated to local dataset loading." + "reason": "Both are multimodal regressions, but assistant-mask generation and video `vision_position_ids` are different pipeline stages and fixes." }, { - "left": "issue:45276", - "right": "issue:45584", + "left": "issue:44464", + "right": "issue:45563", "accept": false, - "reason": "Gemma4 embedding resize issue and Whisper empty-transcription generation bug are unrelated." + "reason": "Compiled-forward chunked generation inconsistency and a stale warning about `num_return_sequences` are not the same bug." }, { - "left": "issue:43295", - "right": "issue:45276", + "left": "issue:44933", + "right": "issue:45362", "accept": false, - "reason": "Processor/tokenizer regression and Gemma4 embedding resize problem do not share the same underlying fix." + "reason": "A nonexistent import from `image_utils` and a Qwen3.5 chat crash are unrelated." }, { - "left": "issue:44938", - "right": "issue:45542", + "left": "issue:43644", + "right": "issue:44913", "accept": false, - "reason": "Python 3.14 import failure and missing TensorFlow backend dependency are unrelated." + "reason": "Global non-persistent buffer initialization junk and config reload losing `rotary_pct` are different classes of regression." }, { - "left": "issue:46082", - "right": "issue:46121", + "left": "issue:43976", + "right": "issue:44855", "accept": false, - "reason": "LlamaConfig head_dim validation and rope-params JSON list handling are separate config bugs." + "reason": "Python-version compatibility for Transformers 5.1.0 and a Python 3.13 decorator/comment parsing failure are not the same issue." }, { - "left": "issue:45245", - "right": "issue:45584", + "left": "issue:43550", + "right": "issue:44190", "accept": false, - "reason": "Category-count overflow and Whisper empty transcription are unrelated runtime failures." + "reason": "Bamba torch.compile/SDPA failure and local dataset loading in an example script are unrelated." }, { - "left": "issue:36246", - "right": "issue:41628", + "left": "issue:44077", + "right": "issue:44877", "accept": false, - "reason": "Both are import errors, but they concern different symbols and different public APIs." + "reason": "Optional `post_init` on patchtsmixer and strict config rejection for granite_speech are distinct config/model-loading concerns." }, { - "left": "issue:43404", - "right": "issue:45216", + "left": "issue:43540", + "right": "issue:44734", "accept": false, - "reason": "Mistral3 weight tying and Qwen3.5 save_pretrained checkpoint keys are different model serialization bugs." + "reason": "Qwen3OmniMoe video-input validation and serve KV-cache indexing are different code paths." }, { - "left": "issue:42175", - "right": "issue:44279", + "left": "issue:45127", + "right": "issue:45357", "accept": false, - "reason": "Package extra omission for TensorFlow and a generic dependency issue are not the same bug." + "reason": "LoRA merge collapse with extended vocab and incorrect visual encoder keys on save are separate serialization/training issues." }, { - "left": "issue:43650", - "right": "issue:44220", + "left": "issue:44462", + "right": "issue:45907", "accept": false, - "reason": "'ADD THE DATA' lacks evidence of the same audio feature-extraction issue." + "reason": "AutoTokenizer ignoring repository `tokenizer.json` and `list_repo_files` not forwarding `revision` are unrelated Hub behaviors." }, { - "left": "issue:44186", - "right": "issue:44734", + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "Pipeline `top_k` nesting behavior and XLMRoberta Flash Attention 2 support are not the same bug or change." + }, + { + "left": "issue:45632", + "right": "issue:46234", "accept": false, - "reason": "LayoutLMv2 tokenizer/padding crash and serve KV-cache indexing are unrelated." + "reason": "A local-model cache path collision and a missing `is_falcon_mamba_ssm_available` import utility are unrelated." } ] }, @@ -10584,61 +10592,77 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:27957", + "issue:35141", "issue:36246", "issue:38617", + "issue:41628", "issue:41950", - "issue:42907", + "issue:42175", + "issue:43116", "issue:43295", - "issue:43299", - "issue:43381", - "issue:43441", + "issue:43335", + "issue:43404", "issue:43479", - "issue:43526", - "issue:43531", "issue:43643", + "issue:43650", "issue:43673", - "issue:43824", "issue:43854", - "issue:43901", "issue:43976", "issue:44016", - "issue:44062", "issue:44112", + "issue:44186", + "issue:44190", "issue:44220", "issue:44246", "issue:44279", - "issue:44336", "issue:44496", - "issue:44561", "issue:44683", "issue:44734", - "issue:44933", + "issue:44737", + "issue:44938", "issue:44964", "issue:44987", "issue:45003", - "issue:45042", "issue:45216", - "issue:45335", - "issue:45372", + "issue:45245", + "issue:45276", "issue:45375", "issue:45446", - "issue:45588", - "issue:45593", + "issue:45542", + "issue:45584", "issue:45600", "issue:45632", - "issue:45759", + "issue:45803", "issue:46082", + "issue:46121", "issue:46123", - "issue:46234", "issue:46369", - "issue:46399" + "issue:46421", + "issue:46424" ], "soft_pairs": [ + "issue:44112|issue:45600", + "issue:44737|issue:45803", + "issue:43335|issue:44190", + "issue:43116|issue:44190", + "issue:45276|issue:45584", + "issue:43295|issue:45276", + "issue:44938|issue:45542", + "issue:46082|issue:46121", + "issue:45632|issue:46421", + "issue:45245|issue:45584", + "issue:36246|issue:41628", + "issue:42175|issue:44279", + "issue:43404|issue:45216", + "issue:43650|issue:44220", "issue:44246|issue:45003", - "issue:41950|issue:44016", "issue:36246|issue:38617", + "issue:41950|issue:44016", + "issue:44186|issue:44734", + "issue:43673|issue:46424", "issue:43479|issue:45375", "issue:44683|issue:45446", "issue:43643|issue:44496", @@ -10647,234 +10671,211 @@ "issue:44016|issue:44220", "issue:44112|issue:44964", "issue:46123|issue:46369", - "issue:44561|issue:45632", - "issue:43976|issue:44987", - "issue:44734|issue:45632", - "issue:43531|issue:45600", - "issue:42907|issue:43441", - "issue:43526|issue:43901", - "issue:46082|issue:46399", - "issue:43295|issue:45335", - "issue:44496|issue:45042", - "issue:43441|issue:45588", - "issue:45372|issue:46234", - "issue:44016|issue:44062", - "issue:43299|issue:45216", - "issue:43673|issue:44336", - "issue:43901|issue:44496", - "issue:43381|issue:45593", - "issue:43824|issue:44933", - "issue:43976|issue:44279", - "issue:44496|issue:45759" + "issue:27957|issue:35141", + "issue:43976|issue:44987" ], - "prepared_review_unit_hash": "08261460e5cff007d9aafb2d7ae349735f16b82b2022b7f0b75e9464ac9d4a2c", + "prepared_review_unit_hash": "bf4877cbe6ec624cd2224d7240f26e0682e6b132101ff1823774bf3d1a0edf95", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36840, - "estimated_input_tokens": 9210, - "estimated_eval_tokens": 18676 + "soft_pair_count": 29, + "serialized_chars": 36513, + "estimated_input_tokens": 9129, + "estimated_eval_tokens": 18514 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes unrelated import errors, config/model-loading regressions, pipeline bugs, and docs/test issues. I would not merge any of the soft-edge pairs as duplicates.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue: the items span multiple unrelated defects rather than one underlying bug.", + "summary": "This cluster is heterogeneous: it mixes unrelated issues across model configs, import/export errors, cache behavior, pipelines, notebooks, and CI/test failures. None of the soft pairs look like the same underlying bug or change, so there is no meaningful canonical duplicate here.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue fits this cluster; the items are not duplicates and span unrelated subsystems.", "canonical_pr_reason": null, - "best_issue_reason": "If forced to pick a representative, issue 44964 is the most concrete and broadly discussed user-facing regression, but it still does not explain the rest of the cluster.", + "best_issue_reason": "No single best issue is suitable because the cluster does not represent one bug family or one fixable change.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44246", - "right": "issue:45003", + "left": "issue:44112", + "right": "issue:45600", "accept": false, - "reason": "Different problems: slow import timing vs unsafe sys.modules access; not the same concrete bug." + "reason": "Both mention CI-like failures, but one is a GraniteSpeech test issue and the other is a broken auto-mapping reference for Sam3LiteText; different bugs." }, { - "left": "issue:41950", - "right": "issue:44016", + "left": "issue:44737", + "right": "issue:45803", "accept": false, - "reason": "A pipeline bug and a notebook syntax error are unrelated." + "reason": "XLNet CPU device placement and a bare-except bug in Fuyu tensor conversion are unrelated code paths." }, { - "left": "issue:36246", - "right": "issue:38617", + "left": "issue:43335", + "right": "issue:44190", "accept": false, - "reason": "Both are import errors, but for different symbols in different modules with different causes." + "reason": "SwitchTransformers config construction and an image-classification script failing on local datasets are different problems." }, { - "left": "issue:43479", - "right": "issue:45375", + "left": "issue:43116", + "right": "issue:44190", "accept": false, - "reason": "Different model configs and different missing/defaulted fields; not one underlying config bug." + "reason": "Both are example/script bugs, but one is multi-label classification output handling and the other is local dataset loading in a different script." }, { - "left": "issue:44683", - "right": "issue:45446", + "left": "issue:45276", + "right": "issue:45584", "accept": false, - "reason": "Both touch flex_attention, but one is a runtime failure on torch 2.9+ and the other is a version-check/import gating bug." + "reason": "Gemma4 embedding resizing and Whisper empty transcription after token alignment are unrelated model behaviors." }, { - "left": "issue:43643", - "right": "issue:44496", + "left": "issue:43295", + "right": "issue:45276", "accept": false, - "reason": "Both are config-loading related, but one is about trust_remote_code dropping fields and the other is about an unrecognized model lacking model_type." + "reason": "A processor/tokenizer regression and Gemma4 embedding resizing are different failures." }, { - "left": "issue:44964", - "right": "issue:45600", + "left": "issue:44938", + "right": "issue:45542", "accept": false, - "reason": "Both are model-loading issues, but for different models and different failure modes." + "reason": "Python 3.14 import failure and missing TensorFlow backend under torch install are distinct dependency issues." }, { - "left": "issue:43854", - "right": "issue:44964", + "left": "issue:46082", + "right": "issue:46121", "accept": false, - "reason": "Different multimodal models and different import/load failures; not the same bug." + "reason": "Llama head_dim validation and rope-parameter list handling are unrelated config-validation bugs." }, { - "left": "issue:44016", - "right": "issue:44220", + "left": "issue:45632", + "right": "issue:46421", "accept": false, - "reason": "A notebook syntax error is unrelated to a feature implementation issue." + "reason": "Both involve cache, but one is a trust_remote_code path collision and the other is chunked-prefill recompilation from lazy cache init." }, { - "left": "issue:44112", - "right": "issue:44964", + "left": "issue:45245", + "right": "issue:45584", "accept": false, - "reason": "CI test flakiness in GraniteSpeech is unrelated to Phi-4 multimodal loading." + "reason": "Category cardinality overflow and Whisper empty transcription are unrelated." }, { - "left": "issue:46123", - "right": "issue:46369", + "left": "issue:36246", + "right": "issue:41628", "accept": false, - "reason": "MaskGenerationPipeline batching logic and sam3-litetext memory overflow are unrelated." + "reason": "Both are import errors, but they concern different missing exports/classes and do not look like one shared bug." }, { - "left": "issue:44561", - "right": "issue:45632", + "left": "issue:42175", + "right": "issue:44279", "accept": false, - "reason": "Both involve trust_remote_code, but one is a missing helper API and the other is a cache-path collision." + "reason": "TensorFlow backend packaging under pip extras and a generic dependency issue are not the same concrete problem." }, { - "left": "issue:43976", - "right": "issue:44987", + "left": "issue:43404", + "right": "issue:45216", "accept": false, - "reason": "A Python-version compatibility issue is not the same as a model-specific load failure." + "reason": "Mistral3 lm_head tying and Qwen3.5 save_pretrained checkpoint correctness are different model-specific bugs." }, { - "left": "issue:44734", - "right": "issue:45632", + "left": "issue:43650", + "right": "issue:44220", "accept": false, - "reason": "Serve/KV-cache indexing crash and local-model cache-path collision are different code paths." + "reason": "A vague data request and an fbank feature extraction issue are unrelated." }, { - "left": "issue:43531", - "right": "issue:45600", + "left": "issue:44246", + "right": "issue:45003", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and Sam3LiteText auto_mapping references are unrelated." + "reason": "Slow import timing and unsafe sys.modules access are different import-related concerns, not the same bug." }, { - "left": "issue:42907", - "right": "issue:43441", + "left": "issue:36246", + "right": "issue:38617", "accept": false, - "reason": "Different Ministral-family failures with different triggers: save_pretrained/dequantization vs FlashAttention loading." + "reason": "Two separate missing-import symbols from different modules; similar symptom, but not one underlying defect." }, { - "left": "issue:43526", - "right": "issue:43901", + "left": "issue:41950", + "right": "issue:44016", "accept": false, - "reason": "An image-processor label reduction bug and outdated docs for return_all_scores are not the same issue." + "reason": "A video-classification pipeline bug and a notebook syntax error are unrelated." }, { - "left": "issue:46082", - "right": "issue:46399", + "left": "issue:44186", + "right": "issue:44734", "accept": false, - "reason": "LlamaConfig head_dim validation and Qwen3.5 rotary-kernel registration failure are unrelated." + "reason": "Tokenizer crashes on NER padding/truncation and serve KV-cache continuation indexing are different code paths." }, { - "left": "issue:43295", - "right": "issue:45335", + "left": "issue:43673", + "right": "issue:46424", "accept": false, - "reason": "One is a processor/tokenizer regression; the other is an embedding resize bug." + "reason": "Both are cache-related, but one is missing cache in chunked_prefill and the other is max_cache_length handling for quantized cache only." }, { - "left": "issue:44496", - "right": "issue:45042", + "left": "issue:43479", + "right": "issue:45375", "accept": false, - "reason": "Missing model_type in config loading is unrelated to PIL image processors wrongly requiring torchvision." + "reason": "Different multimodal config bugs: default subconfig initialization vs a missing field dropped by strict validation." }, { - "left": "issue:43441", - "right": "issue:45588", + "left": "issue:44683", + "right": "issue:45446", "accept": false, - "reason": "Different flash-attention-related failures in different code paths and models." + "reason": "Both touch flex_attention and torch version checks, but one is a compiled flex_attention failure on torch>=2.9 and the other is an AuxRequest import guard; not clearly the same fix." }, { - "left": "issue:45372", - "right": "issue:46234", + "left": "issue:43643", + "right": "issue:44496", "accept": false, - "reason": "Different missing-import symbols in different modules; not a duplicate bug." + "reason": "trust_remote_code missing fields and unrecognized model_type are separate config-loading issues." }, { - "left": "issue:44016", - "right": "issue:44062", + "left": "issue:44964", + "right": "issue:45600", "accept": false, - "reason": "Notebook syntax error versus tokenizer AddedToken argument conflict; unrelated." + "reason": "Both are model-loading problems, but one is Phi-4-multimodal and the other is Sam3LiteText auto-mapping breakage." }, { - "left": "issue:43299", - "right": "issue:45216", + "left": "issue:43854", + "right": "issue:44964", "accept": false, - "reason": "Both are Qwen regressions, but one is loading/import behavior and the other is save_pretrained checkpoint correctness." + "reason": "Different model-loading failures for different models and different failure modes." }, { - "left": "issue:43673", - "right": "issue:44336", + "left": "issue:44016", + "right": "issue:44220", "accept": false, - "reason": "Generation cache behavior and ANSI output formatting are unrelated." + "reason": "Notebook syntax and audio feature extraction are unrelated." }, { - "left": "issue:43901", - "right": "issue:44496", + "left": "issue:44112", + "right": "issue:44964", "accept": false, - "reason": "Documentation mismatch and model-type loading failure are different issues." + "reason": "CI test flakiness in GraniteSpeech and Phi-4-multimodal loading are unrelated." }, { - "left": "issue:43381", - "right": "issue:45593", + "left": "issue:46123", + "right": "issue:46369", "accept": false, - "reason": "Gradient checkpointing eval-mode behavior and D-FINE auxiliary-loss handling are unrelated." + "reason": "Partial-batch pipeline output dropping and memory overflow in sam3-litetext are different symptoms and code paths." }, { - "left": "issue:43824", - "right": "issue:44933", + "left": "issue:27957", + "right": "issue:35141", "accept": false, - "reason": "Both are import-related, but they target different symbols and different modules." + "reason": "Flash Attention support for XLMRoberta and embedding reinitialization after resize are unrelated bugs." }, { "left": "issue:43976", - "right": "issue:44279", - "accept": false, - "reason": "Python version support and a generic dependency issue are not the same concrete bug." - }, - { - "left": "issue:44496", - "right": "issue:45759", + "right": "issue:44987", "accept": false, - "reason": "Unrecognized-model config loading and composite Qwen multimodal unwrapping are different failures." + "reason": "Python-version compatibility and loading a specific model on transformers>=5.1.0 are different issues." } ] }, @@ -10896,283 +10897,288 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:28282", + "issue:35141", "issue:38175", - "issue:39692", - "issue:41950", + "issue:42907", "issue:43295", "issue:43299", + "issue:43381", "issue:43441", - "issue:43450", - "issue:43479", - "issue:43525", + "issue:43526", + "issue:43531", "issue:43577", "issue:43650", - "issue:43854", + "issue:43673", + "issue:43824", "issue:43901", "issue:43976", "issue:44016", + "issue:44062", "issue:44075", "issue:44242", "issue:44279", - "issue:44291", - "issue:44297", - "issue:44351", + "issue:44336", "issue:44368", - "issue:44462", - "issue:44466", + "issue:44496", "issue:44521", "issue:44617", - "issue:44683", - "issue:44779", - "issue:45020", - "issue:45081", + "issue:44792", + "issue:44933", + "issue:45042", "issue:45103", "issue:45200", "issue:45216", - "issue:45276", - "issue:45278", "issue:45325", "issue:45335", "issue:45357", "issue:45372", "issue:45405", - "issue:45542", "issue:45584", - "issue:45812", - "issue:45823", - "issue:45834", - "issue:45907", - "issue:46132", - "issue:46209" + "issue:45588", + "issue:45593", + "issue:45600", + "issue:45759", + "issue:46082", + "issue:46209", + "issue:46234", + "issue:46399", + "issue:46424" ], "soft_pairs": [ - "issue:44075|issue:44368", - "issue:45335|issue:45584", - "issue:43650|issue:46209", - "issue:45103|issue:45372", - "issue:43577|issue:45216", - "issue:44242|issue:45200", - "issue:38175|issue:44617", - "issue:44521|issue:45325", - "issue:43299|issue:45357", - "issue:44279|issue:45405", - "issue:41950|issue:45020", - "issue:44351|issue:45278", - "issue:39692|issue:43901", - "issue:44779|issue:45812", - "issue:43295|issue:44779", - "issue:43479|issue:43525", - "issue:41950|issue:45823", - "issue:44683|issue:45542", - "issue:43976|issue:45405", - "issue:43299|issue:43854", - "issue:44462|issue:45081", - "issue:44297|issue:44462", - "issue:43450|issue:44279", - "issue:43650|issue:46132", - "issue:44466|issue:45276", - "issue:45081|issue:45907", - "issue:44016|issue:44291", - "issue:43441|issue:45834" + "issue:42907|issue:43441", + "issue:43531|issue:45600", + "issue:43526|issue:43901", + "issue:44792|issue:46424", + "issue:43295|issue:45335", + "issue:28282|issue:35141", + "issue:44496|issue:45042", + "issue:46082|issue:46399", + "issue:43441|issue:45588", + "issue:45372|issue:46234", + "issue:44016|issue:44062", + "issue:43299|issue:45216", + "issue:43901|issue:44496", + "issue:43824|issue:44933", + "issue:43673|issue:44336", + "issue:43381|issue:45593", + "issue:43976|issue:44279", + "issue:44075|issue:44368", + "issue:44496|issue:45759", + "issue:45335|issue:45584", + "issue:43650|issue:46209", + "issue:43577|issue:45216", + "issue:45103|issue:45372", + "issue:44242|issue:45200", + "issue:38175|issue:44617", + "issue:44521|issue:45325", + "issue:43299|issue:45357", + "issue:44279|issue:45405" ], - "prepared_review_unit_hash": "aba66aae3d1128930173d1426da858d149aaf54f09f72441b57cb06959b6e136", + "prepared_review_unit_hash": "1aa4d24ba2e5151a2a8559d0b05d52f1cad3b886e5d273a02cb6bf16ad43174b", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36238, - "estimated_input_tokens": 9060, - "estimated_eval_tokens": 18376 + "serialized_chars": 35662, + "estimated_input_tokens": 8916, + "estimated_eval_tokens": 18088 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All listed soft pairs are similar at the subsystem level but do not describe the same underlying bug/change, so they should be kept separate. The strongest standalone issue here is the open multimodal assistant_masks regression.", - "confidence": 0.84, - "canonical_issue_reason": "issue:44521 is the clearest high-signal standalone bug: it is open, reproducible, and has concrete multimodal behavior with multiple inbound references.", + "summary": "This cluster is mostly heterogeneous and does not form one coherent duplicate set. The only strong duplicate-looking pair is the Qwen3.5 `save_pretrained` checkpoint serialization regression; the rest are distinct model/import/config/documentation issues across unrelated code paths.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45357 is the clearest statement of the underlying bug: Qwen3.5 `save_pretrained` is writing incorrect visual-encoder keys, which matches the broader checkpoint corruption report in 45216.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the best overall issue candidate because it is specific, actionable, currently open, and appears to affect an active code path rather than a narrow doc/example path.", + "best_issue_reason": "Issue 45357 is the best representative of the only real duplicate pair because it pinpoints the concrete faulty save path and the incorrect serialized keys.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44075", - "right": "issue:44368", + "left": "issue:42907", + "right": "issue:43441", "accept": false, - "reason": "Both mention behavior around optimization/config warnings, but one is SGD args being ignored and the other is a tied-word-embeddings warning in Qwen3.5 LoRA; different bugs and code paths." + "reason": "Both involve Ministral/flash-attention, but one is about saving a dequantized model and the other is a runtime attention failure; different bugs." }, { - "left": "issue:45335", - "right": "issue:45584", + "left": "issue:43531", + "right": "issue:45600", "accept": false, - "reason": "One is about resize_token_embeddings not affecting t5gemma decoder embeddings; the other is Whisper generation failing on empty transcription after align_special_tokens. Different models and failures." + "reason": "Qwen3-MoE sliding-window behavior and a removed Sam3LiteText config reference are unrelated code paths." }, { - "left": "issue:43650", - "right": "issue:46209", + "left": "issue:43526", + "right": "issue:43901", "accept": false, - "reason": "'ADD THE DATA' is not a technical bug report, while the other is a specific FP8 inference failure on B200. No shared underlying issue." + "reason": "BeitImageProcessorFast label reduction and TextClassificationPipeline docs for `return_all_scores` are unrelated." }, { - "left": "issue:45103", - "right": "issue:45372", + "left": "issue:44792", + "right": "issue:46424", "accept": false, - "reason": "Auto-docstring crashes with future annotations in one case; the other is Gemma 4 processor loading broken by a mistral_common import error. Different root causes." + "reason": "A janus image-generation test failure is not the same bug as cache-length handling for quantized cache." }, { - "left": "issue:43577", - "right": "issue:45216", + "left": "issue:43295", + "right": "issue:45335", "accept": false, - "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained serialization are unrelated failures, even though both touch model state handling." + "reason": "Custom processor/tokenizer handling regression and t5gemma decoder embedding resize are different failures." }, { - "left": "issue:44242", - "right": "issue:45200", + "left": "issue:28282", + "right": "issue:35141", "accept": false, - "reason": "Load-balancing loss gating with router logits is unrelated to Gemma 4 mm_token_type_ids defaults for text-only fine-tuning." + "reason": "PyTorch import missing and embedding reinitialization on resize are completely different issues." }, { - "left": "issue:38175", - "right": "issue:44617", + "left": "issue:44496", + "right": "issue:45042", "accept": false, - "reason": "SigLIP2 zero probabilities and Sam3Video CUDA OOM are completely different symptoms and code paths." + "reason": "Unrecognized model config and incorrect torchvision dependency checks are unrelated." }, { - "left": "issue:44521", - "right": "issue:45325", + "left": "issue:46082", + "right": "issue:46399", "accept": false, - "reason": "assistant_masks being all zero in multimodal chat templates is not the same as Qwen2.5-VL rope index scaling for still-image temporal positions." + "reason": "LlamaConfig head_dim validation and Qwen3.5 rotary-pos-emb registration are different model/config bugs." }, { - "left": "issue:43299", - "right": "issue:45357", + "left": "issue:43441", + "right": "issue:45588", "accept": false, - "reason": "Qwen3VL model loading regression and Qwen3.5 save_pretrained visual-key serialization are related only by model family, not by the same bug." + "reason": "Both mention flash attention, but one is a Ministral model crash and the other is an `s_aux=None` sink-less-model edge case." }, { - "left": "issue:44279", - "right": "issue:45405", + "left": "issue:45372", + "right": "issue:46234", "accept": false, - "reason": "A generic dependency issue is not the same as a specific MIN_PEFT_VERSION bump to an unreleased PyPI version." + "reason": "Missing `ReasoningEffort` import and missing `is_falcon_mamba_ssm_available` helper are unrelated import errors." }, { - "left": "issue:41950", - "right": "issue:45020", + "left": "issue:44016", + "right": "issue:44062", "accept": false, - "reason": "video-classification pipeline using image processors and remote_code breakage are distinct failures with different triggers." + "reason": "Notebook syntax error and `AddedToken` keyword collision are unrelated." }, { - "left": "issue:44351", - "right": "issue:45278", + "left": "issue:43299", + "right": "issue:45216", "accept": false, - "reason": "Both are import-related, but one is a missing HybridCache symbol and the other is broad import errors after upgrading; too different to merge." + "reason": "Qwen3.5/Qwen3.6 model-loading failure and incorrect `save_pretrained` checkpoint output are different stages and bugs." }, { - "left": "issue:39692", - "right": "issue:43901", + "left": "issue:43901", + "right": "issue:44496", "accept": false, - "reason": "SigLIP2 docs example errors and TextClassificationPipeline docs mentioning return_all_scores are different documentation issues and different behaviors." + "reason": "Pipeline docs mismatch and model-type detection failure are not the same issue." }, { - "left": "issue:44779", - "right": "issue:45812", + "left": "issue:43824", + "right": "issue:44933", "accept": false, - "reason": "Deepseek tokenizer regressions and Granite tokenizer regressions are separate model-specific tokenizer bugs." + "reason": "Missing Qwen2.5-VL import and a nonexistent image_utils import are different import-path problems." }, { - "left": "issue:43295", - "right": "issue:44779", + "left": "issue:43673", + "right": "issue:44336", "accept": false, - "reason": "Processor/tokenizer regression with custom model code is unrelated to Deepseek tokenizer producing incorrect results." + "reason": "Generation cache missing in chunked prefill and ANSI code emission in loading_report are unrelated." }, { - "left": "issue:43479", - "right": "issue:43525", + "left": "issue:43381", + "right": "issue:45593", "accept": false, - "reason": "One bug is default vision/audio config initialization in Phi4MultimodalConfig; the other is Llama4Config missing pad_token_id. Different attributes and fixes." + "reason": "Gradient checkpointing eval-mode restriction and D-FINE auxiliary-loss handling are unrelated training bugs." }, { - "left": "issue:41950", - "right": "issue:45823", + "left": "issue:43976", + "right": "issue:44279", "accept": false, - "reason": "A video-classification processor lookup bug is not the same as a Gemma4 device_map auto placement mismatch." + "reason": "Python version compatibility breakage and a generic dependency issue are not the same concrete bug." }, { - "left": "issue:44683", - "right": "issue:45542", + "left": "issue:44075", + "right": "issue:44368", "accept": false, - "reason": "Compiled flex_attention failures on torch >= 2.9 are unrelated to the TensorBoard-without-TensorFlow backend error." + "reason": "SGD optimizer args being ignored and a tie_word_embeddings warning are unrelated." }, { - "left": "issue:43976", - "right": "issue:45405", + "left": "issue:44496", + "right": "issue:45759", "accept": false, - "reason": "Python version compatibility breakage and an unreleased PEFT minimum version bump are different dependency problems." + "reason": "Unrecognized model config and `from_config` not unwrapping `text_config` are different auto-model loading problems." }, { - "left": "issue:43299", - "right": "issue:43854", + "left": "issue:45335", + "right": "issue:45584", "accept": false, - "reason": "Qwen3VL loading failures and GLM-4.7-Flash unit test loading issues concern different models and likely different loaders." + "reason": "t5gemma embedding resize and Whisper empty-transcription generation failure are unrelated." }, { - "left": "issue:44462", - "right": "issue:45081", + "left": "issue:43650", + "right": "issue:46209", "accept": false, - "reason": "AutoTokenizer ignoring repository tokenizer.json and Mistral regex patch crashing on backend_tokenizer access are separate tokenizer-loading bugs." + "reason": "A vague data request and an FP8 inference-on-B200 bug are unrelated." }, { - "left": "issue:44297", - "right": "issue:44462", + "left": "issue:43577", + "right": "issue:45216", "accept": false, - "reason": "A tokenizer_class mismatch on save_pretrained is a serialization bug, while ignoring tokenizer.json is a loading/resolution bug; not the same underlying issue." + "reason": "Blip2 dtype propagation and Qwen3.5 checkpoint saving are different model behaviors." }, { - "left": "issue:43450", - "right": "issue:44279", + "left": "issue:45103", + "right": "issue:45372", "accept": false, - "reason": "Incorrect batched video processor shapes and a generic dependency issue are unrelated." + "reason": "Auto-docstring annotation handling and Gemma 4 processor import failure are unrelated." }, { - "left": "issue:43650", - "right": "issue:46132", + "left": "issue:44242", + "right": "issue:45200", "accept": false, - "reason": "Placeholder/non-bug report versus an AttentionInterface.register behavior change; no shared code defect." + "reason": "MoE load-balancing loss logic and Gemma 4 multimodal token ids are unrelated." }, { - "left": "issue:44466", - "right": "issue:45276", + "left": "issue:38175", + "right": "issue:44617", "accept": false, - "reason": "lm_head.weight serialization/tied-weights behavior and resize_token_embeddings not updating Gemma4 embeddings are different model weight-handling bugs." + "reason": "SigLIP2 zero probabilities and Sam3Video OOM are unrelated model/runtime issues." }, { - "left": "issue:45081", - "right": "issue:45907", + "left": "issue:44521", + "right": "issue:45325", "accept": false, - "reason": "Mistral tokenizer regex crash and list_repo_files missing revision forwarding are unrelated API bugs." + "reason": "All-zero assistant masks and Qwen2.5-VL rope-index scaling are different multimodal bugs." }, { - "left": "issue:44016", - "right": "issue:44291", + "left": "issue:43299", + "right": "issue:45357", "accept": false, - "reason": "A notebook syntax error is not the same as init_empty_weights receiving an unexpected _is_hf_initialized argument." + "reason": "Qwen3.5 loading failure is not the same as incorrect `save_pretrained` serialization output." }, { - "left": "issue:43441", - "right": "issue:45834", + "left": "issue:44279", + "right": "issue:45405", "accept": false, - "reason": "Ministral FlashAttention failure and Kosmos2.5 long-OCR index error are different model-specific issues." + "reason": "Generic dependency issue and a bumped-but-unreleased PEFT minimum version are different packaging problems." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Both report the same Qwen3.5 `save_pretrained` regression producing incorrect checkpoint contents, specifically around serialized keys." } ] }, @@ -11194,283 +11200,276 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:33290", - "issue:41628", - "issue:42175", + "issue:39692", + "issue:41950", "issue:42994", - "issue:43262", + "issue:43295", + "issue:43299", "issue:43317", "issue:43441", + "issue:43450", + "issue:43479", "issue:43525", - "issue:43643", + "issue:43650", "issue:43653", - "issue:43673", - "issue:43723", - "issue:43761", - "issue:43784", - "issue:43824", - "issue:43881", + "issue:43854", "issue:43901", "issue:43976", - "issue:43994", "issue:44016", - "issue:44162", - "issue:44188", - "issue:44246", "issue:44261", "issue:44263", + "issue:44279", + "issue:44291", "issue:44297", - "issue:44336", + "issue:44351", "issue:44361", - "issue:44373", - "issue:44521", - "issue:44589", + "issue:44462", + "issue:44466", "issue:44623", "issue:44671", "issue:44683", "issue:44749", "issue:44779", - "issue:45103", - "issue:45362", + "issue:45020", + "issue:45081", + "issue:45276", + "issue:45278", "issue:45397", + "issue:45405", "issue:45542", - "issue:45561", - "issue:45563", "issue:45593", "issue:45600", "issue:45656", "issue:45712", + "issue:45812", "issue:45823", - "issue:46399" + "issue:45834", + "issue:45907", + "issue:46132" ], "soft_pairs": [ + "issue:41950|issue:45020", + "issue:44351|issue:45278", + "issue:44779|issue:45812", + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:41950|issue:45823", + "issue:43479|issue:43525", + "issue:44683|issue:45542", + "issue:43976|issue:45405", + "issue:43299|issue:43854", + "issue:44462|issue:45081", + "issue:44297|issue:44462", + "issue:43450|issue:44279", + "issue:43650|issue:46132", + "issue:45081|issue:45907", + "issue:44466|issue:45276", + "issue:44016|issue:44291", + "issue:43441|issue:45834", "issue:42994|issue:43317", "issue:43653|issue:45593", - "issue:45600|issue:45712", "issue:44623|issue:45397", "issue:44623|issue:44749", + "issue:45600|issue:45712", "issue:33290|issue:45656", - "issue:44361|issue:45593", "issue:44261|issue:44263", + "issue:44361|issue:45593", "issue:44671|issue:44779", - "issue:44016|issue:45823", - "issue:43441|issue:44683", - "issue:43723|issue:44162", - "issue:42175|issue:43976", - "issue:43901|issue:43994", - "issue:43262|issue:44521", - "issue:43673|issue:43761", - "issue:43824|issue:43976", - "issue:44297|issue:45397", - "issue:43901|issue:44246", - "issue:43525|issue:43881", - "issue:43525|issue:43643", - "issue:41628|issue:43976", - "issue:43784|issue:45542", - "issue:44589|issue:45362", - "issue:44188|issue:44336", - "issue:45103|issue:46399", - "issue:45561|issue:45563", - "issue:44336|issue:44373" + "issue:44016|issue:45823" ], - "prepared_review_unit_hash": "22cab030fc016f50a5980262141bff7a53321f5d017877d05ebfdd801692d1a5", + "prepared_review_unit_hash": "649fca8708ad4cb4775fb907b85ebf5a932b02321ce0f42fedaea65f59834b95", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35904, - "estimated_input_tokens": 8976, - "estimated_eval_tokens": 18208 + "serialized_chars": 35383, + "estimated_input_tokens": 8846, + "estimated_eval_tokens": 17948 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: the soft pairs mostly share only broad subsystem keywords, not the same underlying bug. I would reject every soft edge and avoid selecting a single canonical issue from the set.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "This cluster is mostly a loose set of Transformers v5 regressions and model-specific bugs, not a set of true duplicates. I would use the broad remote_code regression report as the canonical issue, but no soft pair looks like the same concrete bug/change.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45020 is the broadest and most umbrella-like report here, covering recent-version breakage for model loading/custom code and fitting the general regression theme of the cluster.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:45020 is the best single issue to anchor the cluster because it is the most general, well-scoped regression report and can subsume nearby v5 breakage discussions without overfitting to one model.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42994", - "right": "issue:43317", - "accept": false, - "reason": "Both concern quantization/loading, but one is about saving quantized models and the other about loading a dequantized model with device_map offload; different failure modes and code paths." - }, - { - "left": "issue:43653", - "right": "issue:45593", + "left": "issue:41950", + "right": "issue:45020", "accept": false, - "reason": "Completely different topics: BigBirdTokenizer special-token registration vs D-FINE auxiliary losses/denoising." + "reason": "Different failures: video-classification pipeline processor lookup vs general remote_code loading regressions." }, { - "left": "issue:45600", - "right": "issue:45712", + "left": "issue:44351", + "right": "issue:45278", "accept": false, - "reason": "Both are CI/cleanup-related, but one is stale auto_mappings references and the other is leftover dummy PT classes; unrelated fixes." + "reason": "Both are import-related, but one is HybridCache import and the other is many import errors after upgrade; not the same concrete bug." }, { - "left": "issue:44623", - "right": "issue:45397", + "left": "issue:44779", + "right": "issue:45812", "accept": false, - "reason": "Processor save_pretrained missing files vs gemma-4 zero3 from_pretrained loading; save-path and load-path bugs are not the same." + "reason": "Both affect tokenization, but they concern different models and different incorrect outputs, so they are not the same underlying tokenizer bug." }, { - "left": "issue:44623", - "right": "issue:44749", + "left": "issue:43295", + "right": "issue:44779", "accept": false, - "reason": "Missing files during processor saving is unrelated to a model inference slowdown report." + "reason": "Processor/tokenizer API regression vs DeepSeek tokenization regression; different code paths and symptoms." }, { - "left": "issue:33290", - "right": "issue:45656", + "left": "issue:39692", + "right": "issue:43901", "accept": false, - "reason": "Both mention deepspeed/optimizer behavior, but OOM with Adafactor and optimizer.step() being called twice are distinct bugs." + "reason": "Both are docs issues, but one is a SigLIP2 example failure and the other is stale pipeline docs wording; unrelated." }, { - "left": "issue:44361", - "right": "issue:45593", + "left": "issue:41950", + "right": "issue:45823", "accept": false, - "reason": "Tokenizer AttributeError on tasks vs D-FINE loss behavior; no shared underlying code path." + "reason": "Video pipeline processor bug and Gemma4 device-mapping mismatch are unrelated." }, { - "left": "issue:44261", - "right": "issue:44263", + "left": "issue:43479", + "right": "issue:43525", "accept": false, - "reason": "Separate model-specific numerical/config issues: missing rms_norm_eps in MLA q_a_layernorm vs torch.split return values in GlmMoeDsaIndexer." + "reason": "Different config classes and different initialization/attribute problems; not the same bug." }, { - "left": "issue:44671", - "right": "issue:44779", + "left": "issue:44683", + "right": "issue:45542", "accept": false, - "reason": "Both are v5 regressions, but they affect different model families and different components (masked LM predictions vs tokenizer output)." + "reason": "Torch flex_attention compile failure vs TensorFlow/tensorboard backend error; no shared code path." }, { - "left": "issue:44016", - "right": "issue:45823", + "left": "issue:43976", + "right": "issue:45405", "accept": false, - "reason": "Notebook syntax error and Gemma4 device mismatch are unrelated." + "reason": "Python version support regression and PEFT version pin issue are unrelated." }, { - "left": "issue:43441", - "right": "issue:44683", + "left": "issue:43299", + "right": "issue:43854", "accept": false, - "reason": "Both involve attention kernels, but one is a model/FlashAttention compatibility issue and the other is compiled flex_attention failing on newer torch; not the same bug." + "reason": "Both are model-loading issues, but they are model-specific regressions with different failure modes." }, { - "left": "issue:43723", - "right": "issue:44162", + "left": "issue:44462", + "right": "issue:45081", "accept": false, - "reason": "Tokenizer loading in v5 vs ESM2 model breakage; only broad release context overlaps." + "reason": "Tokenizer repository loading vs Mistral regex patch crash are different tokenizer-loading bugs." }, { - "left": "issue:42175", - "right": "issue:43976", + "left": "issue:44297", + "right": "issue:44462", "accept": false, - "reason": "Packaging/backend dependency mismatch is unrelated to Python-version support for Transformers 5.1.0." + "reason": "Save-pretrained metadata mismatch is not the same as ignoring tokenizer.json on load." }, { - "left": "issue:43901", - "right": "issue:43994", + "left": "issue:43450", + "right": "issue:44279", "accept": false, - "reason": "Docs drift for return_all_scores is not the same as incorrect SigLIP2 inference results." + "reason": "Video processor shape bug and generic dependency issue do not match." }, { - "left": "issue:43262", - "right": "issue:44521", + "left": "issue:43650", + "right": "issue:46132", "accept": false, - "reason": "Both mention apply_chat_template, but one is about audio sampling-rate defaults and the other about assistant mask generation for multimodal inputs." + "reason": "Unrelated: placeholder data issue vs behavior change in AttentionInterface.register." }, { - "left": "issue:43673", - "right": "issue:43761", + "left": "issue:45081", + "right": "issue:45907", "accept": false, - "reason": "Generation cache/chunked_prefill regression and CLIPVisionModel hidden_states bug are unrelated." + "reason": "Tokenizer regex crash and list_repo_files revision forwarding are unrelated." }, { - "left": "issue:43824", - "right": "issue:43976", + "left": "issue:44466", + "right": "issue:45276", "accept": false, - "reason": "Import failure for a specific Qwen VL class vs general Python compatibility issue; different problem types." + "reason": "Serialization inconsistency for tied weights vs resize_token_embeddings not updating Gemma4 outputs; different code paths." }, { - "left": "issue:44297", - "right": "issue:45397", + "left": "issue:44016", + "right": "issue:44291", "accept": false, - "reason": "Tokenizer save metadata mismatch and Gemma4 zero3 loading are unrelated." + "reason": "Notebook syntax error and init_empty_weights argument regression are unrelated." }, { - "left": "issue:43901", - "right": "issue:44246", + "left": "issue:43441", + "right": "issue:45834", "accept": false, - "reason": "Documentation mismatch vs import performance issue." + "reason": "FlashAttention failure for Ministral-3 vs long-OCR index error in Kosmos2.5; different model-specific bugs." }, { - "left": "issue:43525", - "right": "issue:43881", + "left": "issue:42994", + "right": "issue:43317", "accept": false, - "reason": "Both are model-specific load errors, but for different models and different missing fields/code paths." + "reason": "Both involve quantization/offload, but one is saving and the other is loading with device_map; not the same bug." }, { - "left": "issue:43525", - "right": "issue:43643", + "left": "issue:43653", + "right": "issue:45593", "accept": false, - "reason": "There is some superficial overlap around missing config fields, but one is a concrete Llama4 AttributeError and the other is a generic trust_remote_code missing-fields report; not enough evidence of the same bug." + "reason": "Tokenizer special-token registration bug and D-FINE denoising-loss behavior are unrelated." }, { - "left": "issue:41628", - "right": "issue:43976", + "left": "issue:44623", + "right": "issue:45397", "accept": false, - "reason": "Importing AutoImageProcessor and Python version support are unrelated." + "reason": "Missing processor save files vs Gemma-4 Zero3 loading are different issues." }, { - "left": "issue:43784", - "right": "issue:45542", + "left": "issue:44623", + "right": "issue:44749", "accept": false, - "reason": "One is a missing nn import in sentence-transformers integration; the other is a TensorFlow backend/dependency issue." + "reason": "Processor save-pretrained problem and data filtering slowdown are unrelated." }, { - "left": "issue:44589", - "right": "issue:45362", + "left": "issue:45600", + "right": "issue:45712", "accept": false, - "reason": "Float8 storage type error and Qwen3.5 chat crash are unrelated." + "reason": "Both are CI/repo hygiene issues, but one is stale auto_mappings entries and the other is leftover dummy classes; different fixes." }, { - "left": "issue:44188", - "right": "issue:44336", + "left": "issue:33290", + "right": "issue:45656", "accept": false, - "reason": "Attention-kernel divergence under torch.compile has nothing to do with ANSI escape codes in loading_report." + "reason": "Both mention DeepSpeed/optimizer behavior, but OOM with Adafactor and optimizer.step being called twice are different concrete bugs." }, { - "left": "issue:45103", - "right": "issue:46399", + "left": "issue:44261", + "right": "issue:44263", "accept": false, - "reason": "auto_docstring annotation handling and Qwen3.5 apply_rotary_pos_emb registration are unrelated." + "reason": "Different model internals: q_a_layernorm epsilon config vs torch.split return values in a separate indexer." }, { - "left": "issue:45561", - "right": "issue:45563", + "left": "issue:44361", + "right": "issue:45593", "accept": false, - "reason": "pytest-xdist race in testing utils vs a stale generation warning; different areas and symptoms." + "reason": "Tokenizer AttributeError on tasks and D-FINE loss behavior are unrelated." }, { - "left": "issue:44336", - "right": "issue:44373", + "left": "issue:44671", + "right": "issue:44779", "accept": false, - "reason": "ANSI-code leakage in loading_report and a wrong docstring are unrelated." + "reason": "Different model families and different failure modes: CamemBERT MLM predictions vs DeepSeek tokenization." } ] }, @@ -11492,297 +11491,282 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ - "issue:36246", + "issue:41628", "issue:41950", + "issue:42175", "issue:42491", - "issue:42757", "issue:42886", - "issue:43257", "issue:43262", "issue:43299", - "issue:43352", - "issue:43475", + "issue:43441", + "issue:43525", "issue:43526", "issue:43531", - "issue:43596", + "issue:43643", "issue:43673", + "issue:43723", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:43881", "issue:43901", + "issue:43976", + "issue:43994", "issue:44162", + "issue:44188", "issue:44220", "issue:44242", "issue:44246", "issue:44295", + "issue:44297", "issue:44336", - "issue:44351", - "issue:44485", + "issue:44373", "issue:44496", - "issue:44561", - "issue:44749", - "issue:44779", - "issue:44857", + "issue:44521", + "issue:44589", + "issue:44683", + "issue:44734", "issue:44871", "issue:45003", - "issue:45137", - "issue:45200", - "issue:45250", - "issue:45292", + "issue:45103", "issue:45362", - "issue:45375", "issue:45397", - "issue:45412", + "issue:45542", "issue:45561", - "issue:45584", + "issue:45563", "issue:45593", - "issue:45759", - "issue:45859", - "issue:46123", "issue:46132", - "issue:46291", - "issue:46369", - "issue:46399" + "issue:46399", + "issue:46424" ], "soft_pairs": [ + "issue:43723|issue:44162", + "issue:43441|issue:44683", + "issue:42175|issue:43976", + "issue:43901|issue:43994", + "issue:43262|issue:44521", + "issue:43673|issue:43761", + "issue:44297|issue:45397", + "issue:43824|issue:43976", + "issue:43525|issue:43881", + "issue:43901|issue:44246", + "issue:43525|issue:43643", + "issue:44589|issue:45362", + "issue:41628|issue:43976", + "issue:43784|issue:45542", + "issue:44188|issue:44336", + "issue:45103|issue:46399", + "issue:45561|issue:45563", + "issue:44336|issue:44373", "issue:44246|issue:44496", "issue:44295|issue:45397", "issue:44242|issue:45593", "issue:42491|issue:43299", - "issue:43526|issue:45003", "issue:43531|issue:45362", + "issue:43526|issue:45003", "issue:41950|issue:44220", "issue:42886|issue:46132", "issue:44246|issue:44871", - "issue:44336|issue:44485", - "issue:36246|issue:42757", - "issue:36246|issue:44351", - "issue:45003|issue:45561", - "issue:44779|issue:45292", - "issue:43475|issue:46369", - "issue:44242|issue:44857", - "issue:45859|issue:46399", - "issue:44295|issue:44749", - "issue:46291|issue:46399", - "issue:43352|issue:45250", - "issue:43673|issue:44561", - "issue:36246|issue:43596", - "issue:43901|issue:45003", - "issue:44162|issue:45200", - "issue:44496|issue:45003", - "issue:43257|issue:45137", - "issue:43262|issue:44246", - "issue:45292|issue:45584", - "issue:45412|issue:46123", - "issue:45375|issue:45759" + "issue:44734|issue:46424" ], - "prepared_review_unit_hash": "cf04e9e1f532aabf52c15d982c7cd04a10043f2880a9048e6a5c0b56522710f9", + "prepared_review_unit_hash": "0813e514ba7a1ce29c1caab06c5158473d1c46bbfc6770d91ee3daa7b0827de0", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36644, - "estimated_input_tokens": 9161, - "estimated_eval_tokens": 18578 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35380, + "estimated_input_tokens": 8845, + "estimated_eval_tokens": 17946 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: many unrelated model- or subsystem-specific regressions, with no soft pair looking like a true duplicate. I picked the broadest v5 compatibility regression as the best representative issue.", - "confidence": 0.27, - "canonical_issue_reason": "Issue 44561 is the broadest concrete regression in the set: a v5 break that impacts trust_remote_code models generally, so it best represents the cluster\u2019s main compatibility theme.", + "summary": "A heterogeneous set of Transformers regressions across tokenizer/config loading, multimodal processors, model families, attention/caching, and docs. They share a broad 'v5 broke something' theme, but most pairs are not the same underlying bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43723 is the most general, user-facing loading regression here: it targets a core AutoTokenizer.from_pretrained path and is broad enough to represent nearby import/load failures better than the model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "44561 is the most globally representative issue because it is user-visible, cross-model, and describes a framework-level break rather than a narrow model-specific symptom.", + "best_issue_reason": "issue:43723 is the best overall representative issue because it describes a common entrypoint failure in v5 and is broad, well-scoped, and close to several other loading-related reports in the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44246", - "right": "issue:44496", + "left": "issue:43723", + "right": "issue:44162", "accept": false, - "reason": "One is a performance complaint about importing transformers; the other is a model config/import error for Olmo-Hybrid. Different problems." + "reason": "Both are loading-related, but one is a generic AutoTokenizer regression and the other is an ESM2-specific failure; different concrete bug surfaces." }, { - "left": "issue:44295", - "right": "issue:45397", + "left": "issue:43441", + "right": "issue:44683", "accept": false, - "reason": "Different failure modes and different models: buffer/position_ids access vs Gemma-4 Zero3 from_pretrained." + "reason": "Both involve attention/flash/flex kernels, but one is a Ministral-3 FlashAttention failure and the other is a torch>=2.9 compiled flex_attention regression; not the same bug." }, { - "left": "issue:44242", - "right": "issue:45593", + "left": "issue:42175", + "right": "issue:43976", "accept": false, - "reason": "Both mention missing losses when a setting is disabled, but they affect different models and code paths; not the same bug." + "reason": "Package/backend dependency issue vs Python version compatibility issue; different underlying causes." }, { - "left": "issue:42491", - "right": "issue:43299", + "left": "issue:43901", + "right": "issue:43994", "accept": false, - "reason": "Both involve Qwen3 MoE/v5 loading regressions, but the models and reported breakages differ enough that this is not a clear duplicate." + "reason": "Docs mismatch for return_all_scores vs a SigLIP2 model output problem; unrelated." }, { - "left": "issue:43526", - "right": "issue:45003", + "left": "issue:43262", + "right": "issue:44521", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug vs unsafe sys.modules access in modeling_utils; unrelated." + "reason": "Both mention apply_chat_template, but one is about audio sampling rate defaults and the other about multimodal assistant masks; different code paths." }, { - "left": "issue:43531", - "right": "issue:45362", + "left": "issue:43673", + "right": "issue:43761", "accept": false, - "reason": "One is a sliding_window issue in Qwen3-MoE; the other is a chat crash in Qwen3.5-35B. Same family, different code path." + "reason": "Cache handling in generation vs CLIPVisionModel hidden_states output; unrelated regressions." }, { - "left": "issue:41950", - "right": "issue:44220", + "left": "issue:44297", + "right": "issue:45397", "accept": false, - "reason": "Video-classification pipeline image-processor lookup vs an fbank feature extraction issue; not the same underlying bug." + "reason": "Tokenizer serialization metadata bug vs a gemma-4 from_pretrained/zero3 loading issue; not the same failure." }, { - "left": "issue:42886", - "right": "issue:46132", + "left": "issue:43824", + "right": "issue:43976", "accept": false, - "reason": "Offline cache loading failure and AttentionInterface.register behavior change are unrelated." + "reason": "Specific Qwen2.5-VL import error vs general Python support regression; different problems." }, { - "left": "issue:44246", - "right": "issue:44871", + "left": "issue:43525", + "right": "issue:43881", "accept": false, - "reason": "Import slowdown and Gemma-3 eos_token_id config mismatch are unrelated issues." + "reason": "Llama4Config missing pad_token_id vs glm-4v-9b loading failure; different model-specific issues." }, { - "left": "issue:44336", - "right": "issue:44485", + "left": "issue:43901", + "right": "issue:44246", "accept": false, - "reason": "ANSI-output cleanup bug vs GLM-5 RoPE implementation discussion; no shared code-path evidence." + "reason": "Documentation mismatch vs intermittent import slowdown; unrelated." }, { - "left": "issue:36246", - "right": "issue:42757", + "left": "issue:43525", + "right": "issue:43643", "accept": false, - "reason": "Both are import errors, but for different symbols and packages; they do not look like the same defect." + "reason": "Both involve missing config fields, but they are different models and different manifestations; too broad to treat as the same bug." }, { - "left": "issue:36246", - "right": "issue:44351", + "left": "issue:44589", + "right": "issue:45362", "accept": false, - "reason": "Different missing exports from different modules; shared symptom only." + "reason": "Float8 storage deserialization issue vs Qwen3.5 chat crash; unrelated." }, { - "left": "issue:45003", - "right": "issue:45561", - "accept": false, - "reason": "sys.modules access bug vs xdist race in testing utils; unrelated." - }, - { - "left": "issue:44779", - "right": "issue:45292", + "left": "issue:41628", + "right": "issue:43976", "accept": false, - "reason": "Tokenizer correctness regression vs resize_token_embeddings not updating output embeddings; different subsystems." + "reason": "AutoImageProcessor import failure vs Python version compatibility; no shared concrete bug." }, { - "left": "issue:43475", - "right": "issue:46369", + "left": "issue:43784", + "right": "issue:45542", "accept": false, - "reason": "SAM3 attribute error and SAM3 LiteText memory overflow are different failures." + "reason": "Different backend/import regressions: sentence-transformers import NameError vs TensorFlow backend detection error." }, { - "left": "issue:44242", - "right": "issue:44857", + "left": "issue:44188", + "right": "issue:44336", "accept": false, - "reason": "Moe loss aggregation bug vs RT-DETR AMP/CUDA crash; different models and code paths." + "reason": "Attention kernel divergence on torch.compile vs ANSI codes in loading_report; unrelated." }, { - "left": "issue:45859", + "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "Both are Qwen3.5/kernels-related, but CPU offload incompatibility and missing-decorator registration crash are distinct issues." + "reason": "Auto-docstring annotation handling bug vs Qwen3.5 kernel registration crash; different subsystems." }, { - "left": "issue:44295", - "right": "issue:44749", + "left": "issue:45561", + "right": "issue:45563", "accept": false, - "reason": "Position_ids buffer access bug vs a general transformer upgrade slowdown; not duplicates." + "reason": "Test harness race condition vs stale generation warning; unrelated." }, { - "left": "issue:46291", - "right": "issue:46399", + "left": "issue:44336", + "right": "issue:44373", "accept": false, - "reason": "Kernel version/revision requirement in evaluate vs Qwen3.5 kernel registration crash; different layers and symptoms." + "reason": "ANSI output formatting bug vs docstring text error; not the same bug." }, { - "left": "issue:43352", - "right": "issue:45250", + "left": "issue:44246", + "right": "issue:44496", "accept": false, - "reason": "A specific model lacking Flash Attention 2 support vs a broad Flash Attention 2 topic; not the same concrete bug." + "reason": "Import performance issue vs unrecognized model/config problem; different symptoms and causes." }, { - "left": "issue:43673", - "right": "issue:44561", + "left": "issue:44295", + "right": "issue:45397", "accept": false, - "reason": "Chunked prefill cache regression and trust_remote_code breakage are separate v5 issues." + "reason": "Position_ids buffer access bug vs gemma-4 loading issue; unrelated." }, { - "left": "issue:36246", - "right": "issue:43596", + "left": "issue:44242", + "right": "issue:45593", "accept": false, - "reason": "Qwen2.5 VL import error vs DeepSpeed ZeRO3 deque IndexError; unrelated." + "reason": "Both concern missing losses, but they are different models and different loss terms; too distinct to merge." }, { - "left": "issue:43901", - "right": "issue:45003", + "left": "issue:42491", + "right": "issue:43299", "accept": false, - "reason": "Documentation mismatch vs unsafe sys.modules access; no overlap in code path." + "reason": "Both involve Qwen3 MoE models, but one is LoRA compatibility across branches and the other is a model loading regression; not the same concrete bug." }, { - "left": "issue:44162", - "right": "issue:45200", + "left": "issue:43531", + "right": "issue:45362", "accept": false, - "reason": "ESM2 breakage and Gemma-4 mm_token_type_ids defaulting are unrelated model-specific problems." + "reason": "Qwen3-MoE sliding_window bug vs Qwen3.5 chat crash; related family, but different defects." }, { - "left": "issue:44496", + "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "Model config validation failure vs sys.modules access bug; different root causes." - }, - { - "left": "issue:43257", - "right": "issue:45137", - "accept": false, - "reason": "Both involve DeepSpeed, but one is Qwen3 MoE weight conversion and the other is a zero3 deque index error; not the same bug." + "reason": "BeitImageProcessorFast label reduction bug vs unsafe sys.modules access in modeling_utils; unrelated." }, { - "left": "issue:43262", - "right": "issue:44246", + "left": "issue:41950", + "right": "issue:44220", "accept": false, - "reason": "Audio chat-template sampling-rate default bug vs import slowdown; unrelated." + "reason": "Video-classification processor mismatch vs audio fbank extraction issue; different modalities and code paths." }, { - "left": "issue:45292", - "right": "issue:45584", + "left": "issue:42886", + "right": "issue:46132", "accept": false, - "reason": "Embedding resize behavior and Whisper empty-transcription generation failure are different defects." + "reason": "Offline tokenizer cache loading bug vs AttentionInterface.register behavior change; unrelated." }, { - "left": "issue:45412", - "right": "issue:46123", + "left": "issue:44246", + "right": "issue:44871", "accept": false, - "reason": "RT-DETR memory cleanup and mask-generation batch handling are unrelated." + "reason": "Import slowdown vs Gemma-3 eos_token_id configuration mismatch; different bug classes." }, { - "left": "issue:45375", - "right": "issue:45759", + "left": "issue:44734", + "right": "issue:46424", "accept": false, - "reason": "Both are Qwen 3.5 multimodal config issues, but one is a missing config field and the other is text_config unwrapping; distinct code paths." + "reason": "Serve KV-continuation indexing crash vs cache max length configuration only applying to quantized cache; different cache issues." } ] }, @@ -11804,283 +11788,297 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ - "issue:42886", - "issue:43065", - "issue:43232", - "issue:43408", + "issue:36246", + "issue:42757", + "issue:43257", + "issue:43262", + "issue:43352", "issue:43475", - "issue:43540", - "issue:43653", + "issue:43596", "issue:43673", - "issue:43931", - "issue:44060", - "issue:44077", + "issue:43901", "issue:44162", - "issue:44188", - "issue:44297", + "issue:44242", + "issue:44246", + "issue:44295", "issue:44336", - "issue:44361", - "issue:44368", + "issue:44351", "issue:44466", "issue:44479", - "issue:44483", + "issue:44485", + "issue:44496", "issue:44561", - "issue:44623", - "issue:44734", "issue:44749", + "issue:44779", + "issue:44857", "issue:44877", - "issue:44977", "issue:44987", "issue:44995", + "issue:45003", "issue:45137", "issue:45200", - "issue:45229", + "issue:45250", "issue:45292", "issue:45335", "issue:45375", "issue:45397", + "issue:45412", "issue:45459", - "issue:45563", + "issue:45561", "issue:45584", - "issue:45588", - "issue:45600", "issue:45656", "issue:45701", - "issue:45987", - "issue:46082", + "issue:45759", + "issue:45859", "issue:46123", "issue:46170", - "issue:46202", - "issue:46399" + "issue:46291", + "issue:46369", + "issue:46399", + "issue:46424" ], "soft_pairs": [ - "issue:45292|issue:45701", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:44336|issue:44485", + "issue:45003|issue:45561", + "issue:44336|issue:46424", + "issue:44779|issue:45292", + "issue:43475|issue:46369", + "issue:44242|issue:44857", + "issue:44295|issue:44749", + "issue:45859|issue:46399", + "issue:46291|issue:46399", + "issue:36246|issue:43596", + "issue:44162|issue:45200", + "issue:43352|issue:45250", + "issue:43673|issue:44561", + "issue:43901|issue:45003", + "issue:44496|issue:45003", + "issue:43257|issue:45137", + "issue:43262|issue:44246", + "issue:45292|issue:45584", + "issue:45412|issue:46123", + "issue:45375|issue:45759", "issue:44561|issue:46399", + "issue:45292|issue:45701", "issue:44877|issue:45375", "issue:44987|issue:45459", - "issue:43232|issue:44336", "issue:44995|issue:45200", "issue:44466|issue:45335", - "issue:45987|issue:46082", - "issue:45397|issue:45656", "issue:44479|issue:46170", - "issue:42886|issue:45229", - "issue:45137|issue:45600", - "issue:43475|issue:43540", - "issue:44483|issue:46202", - "issue:44060|issue:44077", - "issue:43653|issue:44162", - "issue:44995|issue:45588", - "issue:44368|issue:44977", - "issue:44368|issue:45563", - "issue:43065|issue:43408", - "issue:44188|issue:46202", - "issue:44877|issue:46170", - "issue:43931|issue:44368", - "issue:45584|issue:46170", - "issue:43065|issue:46123", - "issue:43673|issue:44734", - "issue:44297|issue:44623", - "issue:44361|issue:44749" + "issue:45397|issue:45656" ], - "prepared_review_unit_hash": "cf104261faedbda3eebf0c9f7b3455075a2cd67db61c7b25de2909488d206f73", + "prepared_review_unit_hash": "dd0288808b2252f111977153a953e315918e453b0dbb14056bf41b331cb960b9", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36129, - "estimated_input_tokens": 9033, - "estimated_eval_tokens": 18322 + "soft_pair_count": 30, + "serialized_chars": 36687, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft duplicate edges look safe to merge; the candidate pairs mostly share only broad subsystem vocabulary, not the same concrete bug. There are no PRs in this cluster.", - "confidence": 0.86, - "canonical_issue_reason": "No single canonical issue stands out across these largely unrelated reports; if forced to pick one representative issue, 44479 is the broadest concrete regression.", + "summary": "This cluster is mostly a similarity cluster, not a true duplicate set: the issues span unrelated import errors, model-specific bugs, config validation, performance, and docs. I did not find any soft pair that clearly describes the same underlying bug/change.", + "confidence": 0.34, + "canonical_issue_reason": "No strong canonical duplicate exists. If one representative is needed, issue 45003 is the broadest core-library bug and has the most central-looking weak similarity links, but the cluster is heterogeneous overall.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44479 is the strongest standalone report: a concrete version-specific regression affecting multiple Qwen video model variants.", + "best_issue_reason": "Issue 45003 is the best single representative only because it is a concrete internal utility bug with broader applicability than the model-specific reports; however, it is not a true duplicate hub.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45292", - "right": "issue:45701", + "left": "issue:36246", + "right": "issue:42757", "accept": false, - "reason": "Resize-token-embeddings/output-embeddings propagation is unrelated to a generic tokenization behavior change." + "reason": "Both are import errors, but they involve different missing symbols from different packages and unrelated code paths." }, { - "left": "issue:44561", - "right": "issue:46399", + "left": "issue:36246", + "right": "issue:44351", "accept": false, - "reason": "Removing a helper for trust_remote_code models is a different failure mode than a missing rotary-pos-emb decorator in Qwen3.5 kernels." + "reason": "Both mention import failures in transformers, but the missing classes are different and the reports point to different modules." }, { - "left": "issue:44877", - "right": "issue:45375", + "left": "issue:44336", + "right": "issue:44485", "accept": false, - "reason": "Both involve strict config handling, but one is a load failure for granite_speech and the other is a missing config field being dropped; not the same bug." + "reason": "ANSI-code output handling and GLM-5 RoPE implementation are unrelated bugs." }, { - "left": "issue:44987", - "right": "issue:45459", + "left": "issue:45003", + "right": "issue:45561", "accept": false, - "reason": "Model loading failure and tokenizer decode-error masking are different code paths and symptoms." + "reason": "One is an unsafe sys.modules access in modeling_utils; the other is a pytest-xdist file race in testing utilities." }, { - "left": "issue:43232", - "right": "issue:44336", + "left": "issue:44336", + "right": "issue:46424", "accept": false, - "reason": "Generation kwargs handling after sync_gpus is unrelated to ANSI coloring in loading reports." + "reason": "Loading-report terminal coloring and cache-length behavior are unrelated." }, { - "left": "issue:44995", - "right": "issue:45200", + "left": "issue:44779", + "right": "issue:45292", "accept": false, - "reason": "A stale cache on second forward pass is not the same as defaulting mm_token_type_ids for text-only fine-tuning." + "reason": "Tokenizer correctness and resize_token_embeddings output synchronization are different model behaviors." }, { - "left": "issue:44466", - "right": "issue:45335", + "left": "issue:43475", + "right": "issue:46369", "accept": false, - "reason": "Both touch embeddings, but one is tied-weight serialization inconsistency and the other is resize_token_embeddings not updating decoder.embed_tokens." + "reason": "Both are SAM3-related, but one is a missing attribute in vision encoder output and the other is a memory overflow problem." }, { - "left": "issue:45987", - "right": "issue:46082", + "left": "issue:44242", + "right": "issue:44857", "accept": false, - "reason": "StaticCache sequence-length return type and LlamaConfig head_dim validation are unrelated bugs." + "reason": "Load-balancing loss computation and RT-DETR AMP crashes are unrelated." }, { - "left": "issue:45397", - "right": "issue:45656", + "left": "issue:44295", + "right": "issue:44749", "accept": false, - "reason": "ZeRO3 from_pretrained loading and a double optimizer-step issue under deepspeed are different failures." + "reason": "A position_ids buffer access error is not the same bug as a slowdown after upgrading transformers." }, { - "left": "issue:44479", - "right": "issue:46170", + "left": "issue:45859", + "right": "issue:46399", "accept": false, - "reason": "A video-input regression in several Qwen variants is unrelated to a Windows/ROCm import dependency problem." + "reason": "Both are Qwen3.5/MoE-related, but one is CPU-offload incompatibility in rotary embeddings and the other is a kernel registration error." }, { - "left": "issue:42886", - "right": "issue:45229", + "left": "issue:46291", + "right": "issue:46399", "accept": false, - "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU OOM are distinct issues." + "reason": "Kernels version/revision installation failure and apply_rotary_pos_emb registration failure are distinct issues." }, { - "left": "issue:45137", - "right": "issue:45600", + "left": "issue:36246", + "right": "issue:43596", "accept": false, - "reason": "DeepSpeed ZeRO3 deque underflow is not the same as stale auto_mappings references to removed Sam3LiteText configs." + "reason": "Qwen2.5-VL import failure and DeepSpeed ZeRO3 BertModel IndexError are unrelated." }, { - "left": "issue:43475", - "right": "issue:43540", + "left": "issue:44162", + "right": "issue:45200", "accept": false, - "reason": "Missing a Sam3VisionEncoderOutput attribute and a Qwen3OmniMoe video-input ValueError are different bugs." + "reason": "ESM2 breakage and Gemma 4 token-type defaults are different model-specific bugs." }, { - "left": "issue:44483", - "right": "issue:46202", + "left": "issue:43352", + "right": "issue:45250", "accept": false, - "reason": "Chat/completions request rejection and a Gemma pipeline roundtrip token issue are unrelated." + "reason": "A model not supporting Flash Attention 2.0 yet is not the same as the generic Flash Attention 2.0 issue." }, { - "left": "issue:44060", - "right": "issue:44077", + "left": "issue:43673", + "right": "issue:44561", "accept": false, - "reason": "A tied-weights warning bug is not the same as banning optional post_init in patchtsmixer." + "reason": "Chunked-prefill cache handling and removal of is_torch_fx_available affect different code paths." }, { - "left": "issue:43653", - "right": "issue:44162", + "left": "issue:43901", + "right": "issue:45003", "accept": false, - "reason": "BigBirdTokenizer special-token decode behavior and ESM2 being broadly broken are not the same concrete defect." + "reason": "Docs for text classification pipeline and unsafe sys.modules access are unrelated." }, { - "left": "issue:44995", - "right": "issue:45588", + "left": "issue:44496", + "right": "issue:45003", "accept": false, - "reason": "Stale indexer cache on second forward pass is unrelated to s_aux=None in flash_attention integration." + "reason": "Unrecognized model config and sys.modules access are different failure modes." }, { - "left": "issue:44368", - "right": "issue:44977", + "left": "issue:43257", + "right": "issue:45137", "accept": false, - "reason": "A tie_word_embeddings warning is unrelated to Qwen3.5 flash-attention generation failure." + "reason": "Both involve DeepSpeed, but one is missing MOE weight conversion and the other is an empty-deque IndexError." }, { - "left": "issue:44368", - "right": "issue:45563", + "left": "issue:43262", + "right": "issue:44246", "accept": false, - "reason": "A config warning and a stale warning in paged generate are different problems." + "reason": "Audio chat-template sampling rate defaults and import latency are unrelated." }, { - "left": "issue:43065", - "right": "issue:43408", + "left": "issue:45292", + "right": "issue:45584", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and a wrong-model-type warning for sam3_tracker are different issues." + "reason": "Embedding resize propagation and Whisper empty transcription handling are different bugs." }, { - "left": "issue:44188", - "right": "issue:46202", + "left": "issue:45412", + "right": "issue:46123", "accept": false, - "reason": "torch.compile attention-kernel divergence is unrelated to the Gemma pipeline roundtrip symbol failure." + "reason": "RT-DETR memory cleanup and mask pipeline batch-finalization are unrelated." }, { - "left": "issue:44877", - "right": "issue:46170", + "left": "issue:45375", + "right": "issue:45759", "accept": false, - "reason": "Strict config loading for granite_speech has nothing to do with Windows ROCm / torch.distributed import failures." + "reason": "Both touch Qwen multimodal config loading, but one is strict-field dropping and the other is text_config unwrapping; they are not the same concrete bug." }, { - "left": "issue:43931", - "right": "issue:44368", + "left": "issue:44561", + "right": "issue:46399", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and a tied-embeddings warning are different loading/config bugs." + "reason": "A removed helper breaking trust_remote_code is unrelated to a kernel registration error." }, { - "left": "issue:45584", - "right": "issue:46170", + "left": "issue:45292", + "right": "issue:45701", "accept": false, - "reason": "Whisper empty-transcription generation failure is unrelated to the Windows/ROCm import issue." + "reason": "Embedding resize synchronization and tokenizer version drift are different problems." }, { - "left": "issue:43673", - "right": "issue:44734", + "left": "issue:44877", + "right": "issue:45375", "accept": false, - "reason": "Chunked-prefill cache absence and /v1/responses KV-cache indexing are different cache bugs in different paths." + "reason": "Both mention strict config handling, but they affect different model configs and different symptoms." }, { - "left": "issue:44297", - "right": "issue:44623", + "left": "issue:44987", + "right": "issue:45459", "accept": false, - "reason": "Tokenizer save_pretrained writing the wrong tokenizer_class is not the same as processor.save_pretrained omitting files." + "reason": "Loading physical-intelligence/fast and hidden tokenizer errors without protobuf are separate import/loading issues." }, { - "left": "issue:44361", - "right": "issue:44749", + "left": "issue:44995", + "right": "issue:45200", "accept": false, - "reason": "MLukeTokenizer AttributeError and the tokenization slowdown regression are unrelated." + "reason": "A stale indexer cache on second forward and Gemma 4 token-type defaults are unrelated." }, { - "left": "issue:43065", - "right": "issue:46123", + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Serialization of tied lm_head weights and t5gemma embedding resize behavior are different bugs." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Qwen VL video regression and Windows/ROCm torch.distributed import issues are unrelated." + }, + { + "left": "issue:45397", + "right": "issue:45656", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d is unrelated to MaskGenerationPipeline dropping the final partial batch." + "reason": "Gemma-4 ZeRO3 loading and double optimizer step with deepspeed are different DeepSpeed-related bugs." } ] }, @@ -12102,101 +12100,102 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ - "issue:28282", - "issue:30990", - "issue:41628", - "issue:42371", - "issue:42673", + "issue:42886", "issue:42994", + "issue:43065", "issue:43232", + "issue:43408", "issue:43475", "issue:43482", - "issue:43525", + "issue:43540", "issue:43643", "issue:43646", + "issue:43653", "issue:43673", "issue:43701", - "issue:43704", - "issue:43723", - "issue:43784", - "issue:43827", - "issue:43873", + "issue:43931", "issue:44038", + "issue:44060", + "issue:44077", "issue:44162", "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44483", "issue:44623", "issue:44749", "issue:44869", "issue:44871", - "issue:44933", - "issue:44945", - "issue:44964", - "issue:45081", - "issue:45092", - "issue:45103", + "issue:44877", + "issue:44977", + "issue:44995", "issue:45137", + "issue:45229", "issue:45230", - "issue:45310", "issue:45356", - "issue:45375", - "issue:45381", "issue:45397", - "issue:45406", - "issue:45478", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", "issue:45656", - "issue:46077", + "issue:45987", + "issue:46082", "issue:46123", "issue:46132", + "issue:46170", "issue:46202", - "issue:46326", - "issue:46399" + "issue:46421" ], "soft_pairs": [ + "issue:45987|issue:46082", + "issue:45137|issue:45600", + "issue:43475|issue:43540", + "issue:44483|issue:46202", + "issue:43653|issue:44162", + "issue:42886|issue:45229", + "issue:44060|issue:44077", + "issue:44995|issue:45588", + "issue:44368|issue:44977", + "issue:43065|issue:43408", + "issue:44368|issue:45563", + "issue:44188|issue:46202", + "issue:44877|issue:46170", + "issue:43931|issue:44368", + "issue:45584|issue:46170", + "issue:44188|issue:46421", + "issue:43065|issue:46123", + "issue:44297|issue:44623", + "issue:44361|issue:44749", "issue:44869|issue:45356", - "issue:43646|issue:44162", "issue:44623|issue:45656", + "issue:43646|issue:44162", "issue:44749|issue:45656", + "issue:43701|issue:44749", "issue:43232|issue:43673", "issue:44871|issue:46132", - "issue:43701|issue:44749", - "issue:43673|issue:43827", "issue:45230|issue:45397", "issue:42994|issue:43482", "issue:44038|issue:44623", - "issue:43475|issue:43643", - "issue:28282|issue:30990", - "issue:43475|issue:45381", - "issue:45406|issue:46202", - "issue:41628|issue:44933", - "issue:42371|issue:45092", - "issue:45375|issue:45478", - "issue:42994|issue:43525", - "issue:43723|issue:46326", - "issue:43784|issue:46399", - "issue:45310|issue:45375", - "issue:45081|issue:45103", - "issue:43873|issue:44945", - "issue:46077|issue:46123", - "issue:42673|issue:43704", - "issue:43232|issue:44188", - "issue:44964|issue:45137" + "issue:43475|issue:43643" ], - "prepared_review_unit_hash": "8ff39a99789b2df7334a78ebcf42415117188820e02d385ebd9d0622284a7551", + "prepared_review_unit_hash": "66d33a4200971214ecd412efa7bfe321a85cf110a6625ff8c8d6f4dd349d7294", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 35929, + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35931, "estimated_input_tokens": 8983, "estimated_eval_tokens": 18222 }, @@ -12205,190 +12204,199 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issue set, with only a few clear duplicates in the Qwen3.5 MoE and Qwen3ForCausalLM reports. No PRs are present.", - "confidence": 0.9, - "canonical_issue_reason": "issue:42673 is the clearest canonical issue: it has a concrete, specific bug report and an exact duplicate in issue:43704 with the same title and failure mode.", + "summary": "The cluster is mostly a grab bag of unrelated closed issues; the soft pairs only show superficial embedding similarity, not the same underlying bug. No PRs are present.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43646 is the broadest framework-level regression in the set ('Transformers 5.0.0 breaks custom model initialization') and is the closest thing to an umbrella issue; the others are mostly model-, tokenizer-, or subsystem-specific.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42673 is the strongest overall representative because it is specific, actionable, and has the cleanest duplicate match among the issues provided.", + "best_issue_reason": "issue:43646 is the best representative issue for this cluster because it captures a general v5 breakage theme, while the rest are too fragmented to serve as a better canonical duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44869", - "right": "issue:45356", + "left": "issue:45987", + "right": "issue:46082", "accept": false, - "reason": "Both involve tokenization/decoding, but they describe different bugs in different models and code paths." + "reason": "StaticCache return-type bug vs LlamaConfig head_dim validation; different code paths." }, { - "left": "issue:43646", - "right": "issue:44162", + "left": "issue:45137", + "right": "issue:45600", "accept": false, - "reason": "The ESM2 report is model-specific and not the same missing-fields/remote-code issue." + "reason": "DeepSpeed ZeRO3 deque error vs stale auto_mappings references; unrelated failures." }, { - "left": "issue:44623", - "right": "issue:45656", + "left": "issue:43475", + "right": "issue:43540", "accept": false, - "reason": "Processor save/export failure and double optimizer-step under deepspeed are unrelated." + "reason": "SAM3 encoder missing attribute vs Qwen3Omni video input ValueError; not the same bug." }, { - "left": "issue:44749", - "right": "issue:45656", + "left": "issue:44483", + "right": "issue:46202", "accept": false, - "reason": "A data filtering slowdown is not the same bug as optimizer stepping twice." + "reason": "Chat/completions request handling vs Gemma pipeline roundtrip; different behaviors." }, { - "left": "issue:43232", - "right": "issue:43673", + "left": "issue:43653", + "right": "issue:44162", "accept": false, - "reason": "Both are generation-related, but they affect different mechanisms: sync_gpus kwargs update vs chunked_prefill cache handling." + "reason": "Tokenizer special-token registration issue vs ESM2 regression; different subsystems." }, { - "left": "issue:44871", - "right": "issue:46132", + "left": "issue:42886", + "right": "issue:45229", "accept": false, - "reason": "EOS config mismatch and AttentionInterface.register behavior changes are different problems." + "reason": "Offline cache tokenizer load failure vs Gemma4 multi-GPU CUDA OOM; unrelated." }, { - "left": "issue:43701", - "right": "issue:44749", + "left": "issue:44060", + "right": "issue:44077", "accept": false, - "reason": "Checkpoint key mismatch is unrelated to the filtering performance regression." + "reason": "Tied-weights warning bug vs patchtsmixer post_init policy change; not the same problem." }, { - "left": "issue:43673", - "right": "issue:43827", + "left": "issue:44995", + "right": "issue:45588", "accept": false, - "reason": "Generation cache bug and stale docs mentioning pipeline() are not the same underlying issue." + "reason": "Stale cache on second forward vs flash_attention None-aux crash; different code paths." }, { - "left": "issue:45230", - "right": "issue:45397", + "left": "issue:44368", + "right": "issue:44977", "accept": false, - "reason": "issue:45230 is too generic; there is no evidence it matches the Gemma-4 ZeRO3 loading bug." + "reason": "Tied-embeddings warning vs flash-attention generation issue; not the same bug." }, { - "left": "issue:42994", - "right": "issue:43482", + "left": "issue:43065", + "right": "issue:43408", "accept": false, - "reason": "Saving quantized models and loading a GGUF model are opposite operations with different failure modes." + "reason": "Dummy Conv2d in SAM3 decoder vs model-type warning for sam3_video/sam3_tracker; distinct issues." }, { - "left": "issue:44038", - "right": "issue:44623", + "left": "issue:44368", + "right": "issue:45563", "accept": false, - "reason": "Qwen3-VL-Moe regression and missing processor save files are unrelated." + "reason": "Different warnings in different generation/save paths; no shared failure mode." }, { - "left": "issue:43475", - "right": "issue:43643", + "left": "issue:44188", + "right": "issue:46202", "accept": false, - "reason": "A SAM 3 vision encoder attribute error is not the same as missing fields in AutoConfig with trust_remote_code." + "reason": "torch.compile kernel divergence vs Gemma pipeline token roundtrip; unrelated." }, { - "left": "issue:28282", - "right": "issue:30990", + "left": "issue:44877", + "right": "issue:46170", "accept": false, - "reason": "Missing PyTorch dependency and sentence-transformers hanging on load are different root causes." + "reason": "Strict config loading failure vs Windows/ROCm import dependency error; different root causes." }, { - "left": "issue:43475", - "right": "issue:45381", + "left": "issue:43931", + "right": "issue:44368", "accept": false, - "reason": "These affect different model families and different failure surfaces." + "reason": "Weight-shape mismatch in Qwen3-VL-30B vs tied-embeddings warning; unrelated." }, { - "left": "issue:45406", - "right": "issue:46202", + "left": "issue:45584", + "right": "issue:46170", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and Gemma pipeline roundtrip failure are related only broadly, not the same bug." + "reason": "Whisper empty-transcription generation bug vs Windows ROCm import/type error; different areas." }, { - "left": "issue:41628", - "right": "issue:44933", + "left": "issue:44188", + "right": "issue:46421", "accept": false, - "reason": "Both are import-related, but they reference different missing symbols/modules and do not look like the same bug." + "reason": "Attention-kernel divergence under compile vs cache lazy-init recompilation; different defects." }, { - "left": "issue:42371", - "right": "issue:45092", + "left": "issue:43065", + "right": "issue:46123", "accept": false, - "reason": "TF32 API guidance and InternVL2 meta-initialization incompatibility are unrelated." + "reason": "SAM3 Conv2d placeholder vs MaskGenerationPipeline partial-batch dropping results; not the same bug." }, { - "left": "issue:45375", - "right": "issue:45478", - "accept": true, - "reason": "Both describe the same Qwen3.5 MoE from_pretrained failure in the same release range; the config-field omission likely explains the generic loading error." + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "tokenizer.save_pretrained metadata mismatch vs processor.save_pretrained missing files; same family only, not same bug." }, { - "left": "issue:42994", - "right": "issue:43525", + "left": "issue:44361", + "right": "issue:44749", "accept": false, - "reason": "Quantized saving failure and Llama4Config pad_token_id missing are different bugs." + "reason": "MLukeTokenizer AttributeError vs global slowdown regression after upgrade; unrelated." }, { - "left": "issue:43723", - "right": "issue:46326", + "left": "issue:44869", + "right": "issue:45356", "accept": false, - "reason": "Tokenizer loading in v5 and a mistral_common continue_final_message bug are different code paths." + "reason": "Whisper timestamp decode crash vs Kimi-K2.5 tokenizer codec regression; different concrete failures." }, { - "left": "issue:43784", - "right": "issue:46399", + "left": "issue:44623", + "right": "issue:45656", "accept": false, - "reason": "Sentence-transformers import NameError and Qwen3.5 kernel registration failure are unrelated." + "reason": "Missing save_pretrained files vs optimizer step called twice; unrelated." }, { - "left": "issue:45310", - "right": "issue:45375", - "accept": true, - "reason": "issue:45310 is the generic from_pretrained error for the same Qwen3.5 MoE problem that issue:45375 narrows to a config-field omission." + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom model initialization breakage vs ESM2-specific regression; too broad to be the same bug." }, { - "left": "issue:45081", - "right": "issue:45103", + "left": "issue:44749", + "right": "issue:45656", "accept": false, - "reason": "A Mistral tokenizer regex patch crash and an auto_docstring annotations crash are unrelated." + "reason": "Preprocessing slowdown vs double optimizer step under deepspeed; different behavior." }, { - "left": "issue:43873", - "right": "issue:44945", + "left": "issue:43701", + "right": "issue:44749", "accept": false, - "reason": "Quantization/offloading behavior and pipeline-parallel output corruption are different bugs." + "reason": "Checkpoint key mismatch vs preprocessing slowdown; no shared underlying change." }, { - "left": "issue:46077", - "right": "issue:46123", + "left": "issue:43232", + "right": "issue:43673", "accept": false, - "reason": "Bidirectional-attention masking and partial-batch result dropping are unrelated." + "reason": "Generation kwargs update after sync_gpus vs missing cache during chunked prefill; related area but distinct bug." }, { - "left": "issue:42673", - "right": "issue:43704", - "accept": true, - "reason": "Exact same VRAM leak report for Qwen3ForCausalLM in multiple dataloader threads; this is a clear duplicate." + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma EOS config inconsistency vs AttentionInterface.register behavior change; not the same issue." }, { - "left": "issue:43232", - "right": "issue:44188", + "left": "issue:45230", + "right": "issue:45397", "accept": false, - "reason": "Generation kwargs handling and torch.compile attention-kernel divergence are different issues." + "reason": "Generic bug report vs gemma-4 zero3 loading problem; insufficient overlap and different specifics." }, { - "left": "issue:44964", - "right": "issue:45137", + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure vs Qwen2.5-GGUF loading failure; different model-loading paths." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe v5 bug vs processor.save_pretrained missing files; unrelated." + }, + { + "left": "issue:43475", + "right": "issue:43643", "accept": false, - "reason": "A model loading regression and a DeepSpeed ZeRO3 deque IndexError are unrelated." + "reason": "SAM3 video encoder attribute error vs AutoConfig trust_remote_code field loss; different code paths." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: the only clearly supported duplicate/near-duplicate clusters are the Qwen3.5 MoE loading issues and the Qwen3ForCausalLM VRAM leak issues, and the response stays conservative elsewhere. No PRs are present, matching 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 }, @@ -12403,305 +12411,306 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ - "issue:30333", + "issue:28282", "issue:30990", - "issue:33290", - "issue:35141", "issue:36010", "issue:41628", "issue:42175", - "issue:42222", + "issue:42371", + "issue:42673", "issue:42831", - "issue:42898", "issue:42994", "issue:43232", - "issue:43278", "issue:43377", - "issue:43479", + "issue:43475", "issue:43504", + "issue:43525", "issue:43526", - "issue:43673", + "issue:43704", "issue:43723", "issue:43784", "issue:43824", - "issue:43976", - "issue:44016", + "issue:43873", "issue:44038", - "issue:44079", "issue:44188", "issue:44222", "issue:44242", "issue:44351", "issue:44496", - "issue:44568", - "issue:44623", - "issue:44734", - "issue:44779", "issue:44792", - "issue:44869", "issue:44933", "issue:44945", + "issue:44964", + "issue:45081", + "issue:45092", "issue:45103", "issue:45137", - "issue:45200", "issue:45310", - "issue:45356", + "issue:45375", + "issue:45381", + "issue:45406", + "issue:45478", "issue:45561", "issue:45712", "issue:45753", - "issue:45812", - "issue:46231" + "issue:46077", + "issue:46123", + "issue:46202", + "issue:46326", + "issue:46399" ], "soft_pairs": [ + "issue:28282|issue:30990", + "issue:43475|issue:45381", + "issue:45406|issue:46202", + "issue:41628|issue:44933", + "issue:42371|issue:45092", + "issue:45375|issue:45478", + "issue:43723|issue:46326", + "issue:42994|issue:43525", + "issue:43784|issue:46399", + "issue:45310|issue:45375", + "issue:45081|issue:45103", + "issue:43873|issue:44945", + "issue:46077|issue:46123", + "issue:42673|issue:43704", "issue:41628|issue:44351", - "issue:44945|issue:45753", + "issue:43232|issue:44188", + "issue:44964|issue:45137", "issue:44222|issue:45310", + "issue:44945|issue:45753", "issue:42994|issue:43824", "issue:44038|issue:45310", "issue:36010|issue:41628", "issue:43232|issue:44792", "issue:42831|issue:44222", - "issue:43673|issue:44188", - "issue:44496|issue:45561", "issue:42175|issue:45712", + "issue:44496|issue:45561", "issue:43526|issue:45561", "issue:43504|issue:45561", "issue:42175|issue:44933", - "issue:43377|issue:44242", - "issue:42898|issue:44568", - "issue:43278|issue:45137", - "issue:43976|issue:44933", - "issue:30990|issue:35141", - "issue:44623|issue:45310", - "issue:42222|issue:45200", - "issue:30333|issue:33290", - "issue:43479|issue:44869", - "issue:43479|issue:44079", - "issue:42898|issue:45812", - "issue:43784|issue:43824", - "issue:43723|issue:44568", - "issue:42898|issue:44779", - "issue:42898|issue:45356", - "issue:44734|issue:45103", - "issue:44016|issue:46231" + "issue:43377|issue:44242" ], - "prepared_review_unit_hash": "b5d3688143c953dda8b6c8394fca5b39e805d57bdbf2eb501970f5007b461cec", + "prepared_review_unit_hash": "f6ec3ca0d9a8af03e4cbda2ed4e9b4e7a3e540074a99a999759be964bb136e40", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36936, - "estimated_input_tokens": 9234, - "estimated_eval_tokens": 18724 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36057, + "estimated_input_tokens": 9015, + "estimated_eval_tokens": 18286 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mixed bag of closed Hugging Face Transformers issues: tokenizer regressions, model/import failures, multimodal model bugs, and a few unrelated training/serving/tooling problems. The soft pairs look superficially similar in a few broad areas, but none appear to be the same underlying bug or change strongly enough to merge.", - "confidence": 0.71, - "canonical_issue_reason": "issue:42898 is the broadest representative of the repeated v5 tokenizer-behavior regressions in the set, and it is generic enough to anchor the tokenizer-related subset better than the very model-specific reports.", + "summary": "This cluster is mostly a heterogeneous mix of unrelated issues. The only clear duplicate pair in the supplied soft edges is the Qwen3ForCausalLM VRAM leak report (42673/43704), with 42673 the best canonical issue. A few other pairs are superficially similar (imports, Qwen/MoE, FP8, masking) but describe different code paths or symptoms.", + "confidence": 0.9, + "canonical_issue_reason": "issue:42673 is the oldest and most-discussed instance of the only clearly duplicated bug in the set, and it has the stronger duplicate signal versus issue:43704.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42898 best captures the cluster\u2019s most recurring theme: tokenizer behavior changes/regressions in v5, with a concrete symptom and broad applicability.", + "best_issue_reason": "issue:42673 is the best representative of the duplicate cluster because it is the earliest, most active report of the same VRAM leak bug and has an exact-title duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41628", - "right": "issue:44351", + "left": "issue:28282", + "right": "issue:30990", "accept": false, - "reason": "Both are import errors, but for different exported symbols and likely different compatibility breaks. They do not look like the same code-path failure." + "reason": "Different failures: missing PyTorch backend import vs Sentence Transformers hanging while loading." }, { - "left": "issue:44945", - "right": "issue:45753", + "left": "issue:43475", + "right": "issue:45381", "accept": false, - "reason": "One is incorrect pipeline-parallel LLM output; the other is an infinite loop on a specific image for Qwen3_5. Different models, modalities, and failure modes." + "reason": "Both are multimodal bugs, but they affect different models and different failure modes." }, { - "left": "issue:44222", - "right": "issue:45310", + "left": "issue:45406", + "right": "issue:46202", "accept": false, - "reason": "Both mention Qwen3 MoE/FP8, but one is save_pretrained and the other is from_pretrained. Not the same concrete bug." + "reason": "Gemma tokenizer/processor problems, but one is a missing attribute and the other is a roundtrip tokenization issue." }, { - "left": "issue:42994", - "right": "issue:43824", + "left": "issue:41628", + "right": "issue:44933", "accept": false, - "reason": "Quantized model saving and a missing Qwen2.5-VL class import are unrelated issues. No shared code path is evident." + "reason": "Both are import/export problems, but they target different symbols and do not share the same code path." }, { - "left": "issue:44038", - "right": "issue:45310", + "left": "issue:42371", + "right": "issue:45092", "accept": false, - "reason": "Both concern Qwen MoE models, but they target different versions and different failure points. Too broad to treat as one bug." + "reason": "Packaging/backend install issue vs remote-code meta-init incompatibility; unrelated." }, { - "left": "issue:36010", - "right": "issue:41628", + "left": "issue:45375", + "right": "issue:45478", "accept": false, - "reason": "Both are import regressions, but they are for different missing names and likely separate compatibility fixes." + "reason": "Related Qwen3.5 MoE area, but one is a dropped config field and the other is a from_pretrained load error." }, { - "left": "issue:43232", - "right": "issue:44792", + "left": "issue:43723", + "right": "issue:46326", "accept": false, - "reason": "GenerationMixin/sync_gpus generation-state handling is unrelated to a janus image-generation test failure. Not the same bug." + "reason": "Tokenizer loading regression vs continue_final_message behavior for mistral_common; different bugs." }, { - "left": "issue:42831", - "right": "issue:44222", + "left": "issue:42994", + "right": "issue:43525", "accept": false, - "reason": "Both mention FP8, but one is an accuracy problem and the other is model saving. Different symptoms and likely different causes." + "reason": "Quantized model saving failure vs missing config attribute; different problems." }, { - "left": "issue:43673", - "right": "issue:44188", + "left": "issue:43784", + "right": "issue:46399", "accept": false, - "reason": "Both are generation-related, but one is missing cache behavior in chunked prefill and the other is divergent attention kernels under torch.compile. Distinct code paths." + "reason": "Sentence-transformers import breakage vs Qwen3.5 kernel registration crash; unrelated." }, { - "left": "issue:44496", - "right": "issue:45561", + "left": "issue:45310", + "right": "issue:45375", "accept": false, - "reason": "A model-type/config loading error and a pytest-xdist file race are unrelated." + "reason": "Same model family, but different bug types: load failure vs missing vision-config field." }, { - "left": "issue:42175", - "right": "issue:45712", + "left": "issue:45081", + "right": "issue:45103", "accept": false, - "reason": "One is dependency/backend packaging for tensorflow; the other is leftover dummy classes leaking into imports. Different maintenance issues." + "reason": "Tokenizer regex crash vs auto-docstring crash from future annotations; unrelated." }, { - "left": "issue:43526", - "right": "issue:45561", + "left": "issue:43873", + "right": "issue:44945", "accept": false, - "reason": "A label-reduction bug in BeitImageProcessorFast and a testing-utils race are unrelated." + "reason": "Quantization/offloading behavior vs pipeline-parallel output corruption; different code paths." }, { - "left": "issue:43504", - "right": "issue:45561", + "left": "issue:46077", + "right": "issue:46123", "accept": false, - "reason": "A pretrained-model legacy-field load failure does not match a captured_info.txt race in tests." + "reason": "Both involve masking, but one is Gemma4 attention masking and the other is pipeline batching state." }, { - "left": "issue:42175", - "right": "issue:44933", - "accept": false, - "reason": "Packaging/backend dependency resolution and a missing import from image_utils are both import-adjacent, but not the same issue." + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same VRAM leak bug in Qwen3ForCausalLM multi-dataloader threads; 42673 is the stronger canonical report." }, { - "left": "issue:43377", - "right": "issue:44242", + "left": "issue:41628", + "right": "issue:44351", "accept": false, - "reason": "Missing padding-mask support in MIMI encoding and missing load-balancing loss when router logits are off are different model-internal bugs." + "reason": "Both are import errors, but they involve different missing symbols and different regressions." }, { - "left": "issue:42898", - "right": "issue:44568", + "left": "issue:43232", + "right": "issue:44188", "accept": false, - "reason": "Both involve tokenizer regressions in v5, but one is cleanup spacing behavior and the other is add_special_tokens/BOS-EOS handling. Different fixes." + "reason": "Generation kwargs/sync_gpus bug vs torch.compile attention-kernel divergence; not the same issue." }, { - "left": "issue:43278", + "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "A dtype change between train/eval and a DeepSpeed ZeRO3 deque underflow are unrelated." + "reason": "Phi-4 multimodal load failure vs DeepSpeed ZeRO3 deque error; unrelated." }, { - "left": "issue:43976", - "right": "issue:44933", + "left": "issue:44222", + "right": "issue:45310", "accept": false, - "reason": "Python version compatibility and an image_utils import failure do not share a concrete underlying bug." + "reason": "FP8 save_pretrained bug vs Qwen3.5 from_pretrained load error; different failures." }, { - "left": "issue:30990", - "right": "issue:35141", + "left": "issue:44945", + "right": "issue:45753", "accept": false, - "reason": "A loading hang in sentence-transformers and a post_init embedding reinitialization bug are unrelated." + "reason": "Pipeline-parallel correctness issue vs image-specific infinite loop; distinct symptoms and code paths." }, { - "left": "issue:44623", - "right": "issue:45310", + "left": "issue:42994", + "right": "issue:43824", "accept": false, - "reason": "Missing files on processor.save_pretrained and a Qwen3.5 MoE from_pretrained error are both save/load related, but not the same failure path." + "reason": "Quantized saving vs missing model import; unrelated." }, { - "left": "issue:42222", - "right": "issue:45200", + "left": "issue:44038", + "right": "issue:45310", "accept": false, - "reason": "Both touch multimodal/vision-adjacent models, but one is a broken VitPose model family and the other is a Gemma 4 token-type-id defaulting bug." + "reason": "Both touch Qwen MoE and transformers v5, but they are different model variants and different failure modes." }, { - "left": "issue:30333", - "right": "issue:33290", + "left": "issue:36010", + "right": "issue:41628", "accept": false, - "reason": "MLflow reporting for failed jobs and an OOM in deepspeed+adafactor are completely separate training issues." + "reason": "Both are import errors, but they are for different classes and likely different API regressions." }, { - "left": "issue:43479", - "right": "issue:44869", + "left": "issue:43232", + "right": "issue:44792", "accept": false, - "reason": "A multimodal config defaulting bug and a Whisper timestamp decode crash are unrelated." + "reason": "Generation state bug vs janus image-generation test failure; unrelated." }, { - "left": "issue:43479", - "right": "issue:44079", + "left": "issue:42831", + "right": "issue:44222", "accept": false, - "reason": "Default config initialization and ModelOutput key assignment are different library utilities with different failure modes." + "reason": "Both mention FP8, but one is an accuracy issue and the other is model saving." }, { - "left": "issue:42898", - "right": "issue:45812", + "left": "issue:42175", + "right": "issue:45712", "accept": false, - "reason": "Both are tokenizer regressions in v5, but one is cleanup-space behavior and the other is wrong token IDs for Granite models. Similar theme, not the same bug." + "reason": "Backend-install packaging issue vs leftover dummy class leakage; different code paths." }, { - "left": "issue:43784", - "right": "issue:43824", + "left": "issue:44496", + "right": "issue:45561", "accept": false, - "reason": "Both are import errors, but one is a sentence-transformers NameError on nn and the other is a missing Qwen2.5-VL export. Distinct issues." + "reason": "Model config loading failure vs xdist testing race; unrelated." }, { - "left": "issue:43723", - "right": "issue:44933", + "left": "issue:43526", + "right": "issue:45561", "accept": false, - "reason": "Tokenizer loading failure and a missing image_utils import are different loading/import problems." + "reason": "Beit image-processor label bug vs testing race; unrelated." }, { - "left": "issue:44016", - "right": "issue:46231", + "left": "issue:43504", + "right": "issue:45561", "accept": false, - "reason": "A notebook syntax error and a documentation parameter-name mismatch are not the same bug." + "reason": "Beit pretrained-loading legacy-field bug vs testing race; unrelated." }, { - "left": "issue:44734", - "right": "issue:45103", + "left": "issue:42175", + "right": "issue:44933", "accept": false, - "reason": "KV-cache continuation tensor indexing in serve and an auto-docstring AttributeError with future annotations are unrelated internals issues." + "reason": "Tensorflow-backend install issue vs image_utils import mismatch; unrelated." }, { - "left": "issue:44568", - "right": "issue:46231", + "left": "issue:43377", + "right": "issue:44242", "accept": false, - "reason": "Tokenizer special-token behavior and a docs-only parameter name mismatch are unrelated." + "reason": "Different model behavior bugs: MIMI batching discrepancy vs MoE load-balancing loss omission." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet. The canonical choice of issue:42673 over issue:43704 is supported by the earlier creation date, higher discussion activity, and identical title. The soft-edge verdicts stay conservative overall and do not overstate duplicates where the packet only shows superficial similarity." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -12716,318 +12725,320 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:22355", + "issue:27957", "issue:28282", "issue:30333", + "issue:30990", + "issue:33290", "issue:34689", "issue:35141", "issue:36331", - "issue:41628", - "issue:42757", + "issue:42222", "issue:42831", "issue:42898", - "issue:43122", - "issue:43232", - "issue:43475", + "issue:43278", "issue:43479", "issue:43644", "issue:43723", + "issue:43784", + "issue:43824", "issue:43874", + "issue:43976", + "issue:44016", "issue:44060", "issue:44075", - "issue:44190", - "issue:44206", + "issue:44079", "issue:44222", - "issue:44263", "issue:44295", "issue:44297", - "issue:44373", "issue:44448", - "issue:44484", - "issue:44485", + "issue:44568", "issue:44623", + "issue:44734", + "issue:44779", + "issue:44869", "issue:44933", - "issue:44977", - "issue:45071", "issue:45092", "issue:45103", "issue:45125", + "issue:45137", + "issue:45200", "issue:45245", "issue:45310", - "issue:45381", + "issue:45356", "issue:45405", "issue:45656", "issue:45712", - "issue:45834", - "issue:46082", + "issue:45812", "issue:46132", - "issue:46209", - "issue:46231", - "issue:46305" + "issue:46231" ], "soft_pairs": [ + "issue:42898|issue:44568", + "issue:43976|issue:44933", + "issue:43278|issue:45137", + "issue:30990|issue:35141", + "issue:44623|issue:45310", + "issue:42222|issue:45200", + "issue:30333|issue:33290", + "issue:43479|issue:44869", + "issue:43479|issue:44079", + "issue:42898|issue:45812", + "issue:43784|issue:43824", + "issue:43723|issue:44568", + "issue:42898|issue:44779", + "issue:42898|issue:45356", + "issue:44016|issue:46231", + "issue:44734|issue:45103", "issue:43874|issue:45092", "issue:44295|issue:45656", "issue:44075|issue:46231", - "issue:44060|issue:45125", "issue:42898|issue:44448", + "issue:44060|issue:45125", "issue:45405|issue:45712", - "issue:44448|issue:46132", "issue:44222|issue:44297", "issue:44933|issue:45712", "issue:44297|issue:45656", - "issue:30333|issue:34689", + "issue:44448|issue:46132", "issue:44222|issue:44623", + "issue:27957|issue:30333", + "issue:30333|issue:34689", "issue:43479|issue:43644", "issue:42831|issue:44623", - "issue:45245|issue:46132", "issue:30333|issue:36331", "issue:30333|issue:35141", - "issue:28282|issue:30333", "issue:22355|issue:28282", - "issue:43122|issue:46209", - "issue:45071|issue:46082", - "issue:44297|issue:45310", - "issue:44190|issue:46082", - "issue:43475|issue:46305", - "issue:41628|issue:42757", - "issue:43723|issue:44977", - "issue:44206|issue:45103", - "issue:45381|issue:46305", - "issue:44222|issue:44484", - "issue:43232|issue:45834", - "issue:44295|issue:44623", - "issue:44263|issue:45092", - "issue:44373|issue:44485" + "issue:45245|issue:46132" ], - "prepared_review_unit_hash": "9a8a0234310d3d6206b200042c0601c4c8a0dd5e666bf300c118e7a7b3c67b40", + "prepared_review_unit_hash": "d25ada5a1b30b975823c85751b13ded869c2afcf22d78031d8a602dc03134946", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37366, - "estimated_input_tokens": 9342, - "estimated_eval_tokens": 18940 + "soft_pair_count": 35, + "serialized_chars": 38199, + "estimated_input_tokens": 9550, + "estimated_eval_tokens": 19356 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are mostly unrelated Transformers regressions and API complaints. The soft links are driven by superficial similarity (version breakage, import errors, save_pretrained, multimodal/model-loading failures), but they do not describe 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 set is heterogeneous, and the soft pairs only share broad symptom categories rather than a single underlying defect.", + "summary": "The soft pairs are mostly unrelated issues that only share broad themes like tokenization, save/load, or deepspeed. I would reject all proposed duplicates; there is no clear same-bug pair here.", + "confidence": 0.92, + "canonical_issue_reason": "If a single issue must represent the set, 42898 is the broadest and most generic tokenizer-regression report among the candidates, but the cluster is not actually a duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a safe global representative for duplicate triage here; choosing one would be arbitrary because the cluster does not center on one defect.", + "best_issue_reason": "42898 is the strongest general representative of the tokenizer-regression subgroup: concrete, user-visible, and broader than the more model-specific reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43874", - "right": "issue:45092", + "left": "issue:42898", + "right": "issue:44568", "accept": false, - "reason": "Different failures: a missing GLM46V image-processor method vs. remote-code InternVL2 meta-init incompatibility." + "reason": "Both are tokenizer regressions, but they describe different behaviors and different affected models." }, { - "left": "issue:44295", - "right": "issue:45656", + "left": "issue:43976", + "right": "issue:44933", "accept": false, - "reason": "Unrelated bugs: buffer access error for position_ids vs. optimizer stepping twice under deepspeed." + "reason": "Python-version incompatibility and a missing import are unrelated failures." }, { - "left": "issue:44075", - "right": "issue:46231", + "left": "issue:43278", + "right": "issue:45137", "accept": false, - "reason": "One is a runtime optimizer-argument bug; the other is a documentation parameter-name mismatch." + "reason": "Embedding dtype drift and a DeepSpeed deque IndexError are separate bugs." }, { - "left": "issue:44060", - "right": "issue:45125", + "left": "issue:30990", + "right": "issue:35141", "accept": false, - "reason": "Different Qwen issues: incorrect tied-weights warning vs. missing tensor-parallel plan attribute." + "reason": "Loading hangs and embedding reinitialization are different code paths and symptoms." }, { - "left": "issue:42898", - "right": "issue:44448", + "left": "issue:44623", + "right": "issue:45310", "accept": false, - "reason": "Both involve v4/v5 behavior changes, but one is tokenization cleanup semantics and the other is Pegasus output drift." + "reason": "Processor save-pretrained missing files is unrelated to Qwen3.5 MoE from_pretrained failure." }, { - "left": "issue:45405", - "right": "issue:45712", + "left": "issue:42222", + "right": "issue:45200", "accept": false, - "reason": "PEFT version pinning and leftover dummy PT objects are unrelated maintenance issues." + "reason": "Broken vitpose models and Gemma 4 multimodal token defaults are different issues." }, { - "left": "issue:44448", - "right": "issue:46132", + "left": "issue:30333", + "right": "issue:33290", "accept": false, - "reason": "Different code paths: Pegasus output differences vs. AttentionInterface.register changing callback behavior." + "reason": "MLflow reporting after failure and Adafactor OOM are unrelated." }, { - "left": "issue:44222", - "right": "issue:44297", + "left": "issue:43479", + "right": "issue:44869", "accept": false, - "reason": "Both touch save_pretrained, but one is FP8 MoE serialization and the other is tokenizer metadata mismatch." + "reason": "Config default initialization and Whisper timestamp decoding are unrelated." }, { - "left": "issue:44933", - "right": "issue:45712", + "left": "issue:43479", + "right": "issue:44079", "accept": false, - "reason": "Import/export cleanup in image_utils is not the same as dummy class leakage in dummy_pt_objects.py." + "reason": "Multimodal config defaults and ModelOutput key assignment are different subsystems." }, { - "left": "issue:44297", - "right": "issue:45656", + "left": "issue:42898", + "right": "issue:45812", "accept": false, - "reason": "Tokenizer config serialization has nothing to do with deepspeed optimizer stepping." + "reason": "Both are v5 tokenizer regressions, but the affected models and concrete bugs differ." }, { - "left": "issue:30333", - "right": "issue:34689", + "left": "issue:43784", + "right": "issue:43824", "accept": false, - "reason": "MLflow reporting on failed training jobs is unrelated to Llama model loading regression." + "reason": "Importing sentence-transformers and missing Qwen2.5-VL export are unrelated import errors." }, { - "left": "issue:44222", - "right": "issue:44623", + "left": "issue:43723", + "right": "issue:44568", "accept": false, - "reason": "Both are save_pretrained-related, but they affect different artifacts and failure modes (model FP8 vs processor files)." + "reason": "Tokenizer loading failure and missing BOS/EOS addition are not the same bug." }, { - "left": "issue:43479", - "right": "issue:43644", + "left": "issue:42898", + "right": "issue:44779", "accept": false, - "reason": "Multimodal config initialization bug is unrelated to non-persistent buffer corruption." + "reason": "Different tokenizer regressions affecting different models and tokenization logic." }, { - "left": "issue:42831", - "right": "issue:44623", + "left": "issue:42898", + "right": "issue:45356", "accept": false, - "reason": "FP8 accuracy degradation and missing processor save files are separate problems." + "reason": "Different tokenizer regression symptoms; one is generic behavior change, the other codec/regex handling." }, { - "left": "issue:45245", - "right": "issue:46132", + "left": "issue:44016", + "right": "issue:46231", "accept": false, - "reason": "Category-count runtime error is unrelated to AttentionInterface.register behavior changes." + "reason": "Notebook syntax error and documentation parameter typo are unrelated." }, { - "left": "issue:30333", - "right": "issue:36331", + "left": "issue:44734", + "right": "issue:45103", "accept": false, - "reason": "MLflow job-finalization bug is not the same as Trainer compute_loss signature incompatibility." + "reason": "KV-cache continuation crash and auto-docstring annotation handling are unrelated." }, { - "left": "issue:30333", - "right": "issue:35141", + "left": "issue:43874", + "right": "issue:45092", "accept": false, - "reason": "Training-job reporting and embedding reinitialization are distinct bugs in different subsystems." + "reason": "Missing image-processor method and remote-code meta-init incompatibility are different multimodal bugs." }, { - "left": "issue:28282", - "right": "issue:30333", + "left": "issue:44295", + "right": "issue:45656", "accept": false, - "reason": "Missing PyTorch import dependency is unrelated to MLflow reporting on failed jobs." + "reason": "Position-id buffer access and duplicate optimizer steps are unrelated." }, { - "left": "issue:22355", - "right": "issue:28282", + "left": "issue:44075", + "right": "issue:46231", "accept": false, - "reason": "Both are import errors, but one is missing transformers.onnx and the other is missing PyTorch; different root causes." + "reason": "Optimizer argument handling and a docstring parameter-name mismatch are unrelated." }, { - "left": "issue:43122", - "right": "issue:46209", + "left": "issue:42898", + "right": "issue:44448", "accept": false, - "reason": "Tokenizer output drift across versions is unrelated to FP8 inference failing on B200." + "reason": "Both mention v4-to-v5 output changes, but they are different tokenizer/output regressions." }, { - "left": "issue:45071", - "right": "issue:46082", + "left": "issue:44060", + "right": "issue:45125", "accept": false, - "reason": "General PretrainedConfig type checking regression is not the same as LlamaConfig head_dim validation." + "reason": "Incorrect tied-weights warning and missing tensor-parallel plan are distinct Qwen issues." }, { - "left": "issue:44297", - "right": "issue:45310", + "left": "issue:45405", + "right": "issue:45712", "accept": false, - "reason": "Tokenizer save metadata mismatch does not match the Qwen3.5 Moe from_pretrained regression." + "reason": "Dependency version bump and dummy-class cleanup are unrelated maintenance issues." }, { - "left": "issue:44190", - "right": "issue:46082", + "left": "issue:44222", + "right": "issue:44297", "accept": false, - "reason": "Dataset loading in an example script is unrelated to LlamaConfig constructor validation." + "reason": "FP8 save-pretrained and tokenizer save-pretrained bugs affect different components." }, { - "left": "issue:43475", - "right": "issue:46305", + "left": "issue:44933", + "right": "issue:45712", "accept": false, - "reason": "SAM3 video encoder attribute missing is a different bug from a Moonshine streaming mask off-by-one." + "reason": "Missing image_utils import and leftover dummy classes are not the same defect." }, { - "left": "issue:41628", - "right": "issue:42757", + "left": "issue:44297", + "right": "issue:45656", "accept": false, - "reason": "Import errors in different modules and packages; no shared underlying defect." + "reason": "Tokenizer metadata save bug and DeepSpeed optimizer stepping are unrelated." }, { - "left": "issue:43723", - "right": "issue:44977", + "left": "issue:44448", + "right": "issue:46132", "accept": false, - "reason": "Tokenizer loading failure and Qwen flash-attention generation bug are separate issues." + "reason": "Pegasus v4/v5 output regression and AttentionInterface.register behavior change are different bugs." }, { - "left": "issue:44206", - "right": "issue:45103", + "left": "issue:44222", + "right": "issue:44623", "accept": false, - "reason": "Feature-extractor argument incompatibility is unrelated to the auto_docstring annotations crash." + "reason": "Model save-pretrained FP8 bug and processor save-pretrained missing files are not the same issue." }, { - "left": "issue:45381", - "right": "issue:46305", + "left": "issue:27957", + "right": "issue:30333", "accept": false, - "reason": "Vision position-id computation and streaming sliding-window masking are different algorithms and bugs." + "reason": "Flash Attention support and MLflow reporting after failures are unrelated." }, { - "left": "issue:44222", - "right": "issue:44484", + "left": "issue:30333", + "right": "issue:34689", "accept": false, - "reason": "FP8 save_pretrained failure is unrelated to the default 50GB shard-size question." + "reason": "Training-job reporting and Llama 3.2 vision model loading are unrelated." }, { - "left": "issue:43232", - "right": "issue:45834", + "left": "issue:43479", + "right": "issue:43644", "accept": false, - "reason": "Generation kwarg update ordering is unrelated to long OCR input index errors." + "reason": "Multimodal config defaults and junk-filled buffers are different v5 initialization bugs." }, { - "left": "issue:44295", + "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "Position_ids buffer access and processor save_pretrained missing files are different serialization/runtime issues." + "reason": "FP8 accuracy issues and processor file-saving bugs are unrelated." }, { - "left": "issue:44263", - "right": "issue:45092", + "left": "issue:22355", + "right": "issue:28282", "accept": false, - "reason": "torch.split return-value handling in a MoE indexer is unrelated to remote-code meta-init incompatibility." + "reason": "Missing transformers.onnx and missing PyTorch are different optional-dependency import errors." }, { - "left": "issue:44373", - "right": "issue:44485", + "left": "issue:45245", + "right": "issue:46132", "accept": false, - "reason": "Docstring wording for position_ids is not the same as the GLM-5 RoPE implementation issue." + "reason": "Category-limit runtime error and registered-function behavior change are unrelated." } ] }, @@ -13049,283 +13060,289 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ - "issue:22355", - "issue:28218", "issue:30990", "issue:38175", - "issue:39401", "issue:41628", "issue:42757", "issue:42831", - "issue:42886", "issue:42907", - "issue:42994", - "issue:43066", + "issue:43122", "issue:43232", - "issue:43408", - "issue:43479", - "issue:43576", + "issue:43475", "issue:43644", "issue:43723", "issue:43761", - "issue:43784", - "issue:43824", - "issue:44038", + "issue:44190", + "issue:44206", + "issue:44222", "issue:44261", "issue:44263", "issue:44279", + "issue:44295", + "issue:44297", "issue:44360", - "issue:44368", "issue:44373", "issue:44483", - "issue:44568", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44734", "issue:44829", "issue:44871", - "issue:44913", + "issue:44977", "issue:44991", + "issue:45071", "issue:45092", - "issue:45278", + "issue:45103", "issue:45310", "issue:45357", + "issue:45381", "issue:45406", - "issue:45593", - "issue:45753", + "issue:45834", "issue:46002", "issue:46050", "issue:46082", + "issue:46209", "issue:46305", "issue:46328", - "issue:46369", - "issue:46402" + "issue:46421", + "issue:46424" ], "soft_pairs": [ + "issue:45071|issue:46082", + "issue:43122|issue:46209", + "issue:44297|issue:45310", + "issue:41628|issue:42757", + "issue:43475|issue:46305", + "issue:44190|issue:46082", + "issue:43723|issue:44977", + "issue:44206|issue:45103", + "issue:45381|issue:46305", + "issue:43232|issue:46424", + "issue:44734|issue:46421", + "issue:44222|issue:44484", + "issue:43232|issue:45834", + "issue:44295|issue:44623", + "issue:44263|issue:45092", + "issue:44373|issue:44485", "issue:43723|issue:44871", "issue:44483|issue:45406", "issue:44279|issue:46328", - "issue:42907|issue:46050", "issue:43723|issue:44991", - "issue:44263|issue:44360", + "issue:42907|issue:46050", "issue:30990|issue:38175", + "issue:44263|issue:44360", "issue:45092|issue:45357", "issue:43761|issue:46082", - "issue:44829|issue:46002", "issue:43644|issue:46305", + "issue:44829|issue:46002", "issue:44261|issue:45092", - "issue:42831|issue:45310", - "issue:43479|issue:44913", - "issue:41628|issue:43824", - "issue:43408|issue:46402", - "issue:42831|issue:44038", - "issue:44263|issue:44373", - "issue:39401|issue:44568", - "issue:42886|issue:42994", - "issue:42886|issue:43066", - "issue:42757|issue:45278", - "issue:44368|issue:46305", - "issue:22355|issue:28218", - "issue:42831|issue:45753", - "issue:43232|issue:45593", - "issue:43408|issue:46369", - "issue:43576|issue:43784" + "issue:42831|issue:45310" ], - "prepared_review_unit_hash": "dbab9310413d5158a08aea6e80fb5301c5ed59d8a716d5b31de490879c1417d7", + "prepared_review_unit_hash": "0fd8afdc594ca25355caf7b55b9a84e37acbd63ee89195963a967a8e057242a8", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 35841, - "estimated_input_tokens": 8961, - "estimated_eval_tokens": 18178 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35560, + "estimated_input_tokens": 8890, + "estimated_eval_tokens": 18036 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No true duplicate cluster here: the soft pairs are mostly lexical matches across unrelated tokenizer/import/model/quantization/regression reports, so I reject all of them.", - "confidence": 0.92, - "canonical_issue_reason": "No single canonical issue fits this set; the items span many unrelated bugs and failure modes, so there is no clear duplicate hub to anchor the cluster.", + "summary": "Loose cluster of mostly unrelated Transformers issues: v5 regressions, tokenizer/config loading failures, cache/serving bugs, and a few model-specific defects. No soft pair appears to be the same underlying bug or change.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45071 is the broadest and most representative item in the cluster: a clear Transformers v5 regression affecting core config behavior. It\u2019s concise, generic, and closest to the cluster\u2019s shared theme of v5 breakage.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a strong global representative for this cluster, because even the closest pairs differ by component, model, or symptom.", + "best_issue_reason": "issue:45071 is the best single issue representative because it describes a core framework regression rather than a model-specific symptom or a narrow edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43723", - "right": "issue:44871", + "left": "issue:45071", + "right": "issue:46082", "accept": false, - "reason": "Both involve tokenizer behavior, but one is a v5 AutoTokenizer load failure and the other is a Gemma eos_token_id mismatch; different bugs." + "reason": "Both are config-related v5 problems, but one is broad type-checking behavior and the other is LlamaConfig validation for head_dim; different code paths and fixes." }, { - "left": "issue:44483", - "right": "issue:45406", + "left": "issue:43122", + "right": "issue:46209", "accept": false, - "reason": "Chat/completions request handling is unrelated to a Gemma4Processor missing _tokenizer attribute in serve." + "reason": "Different subsystems and failure modes: tokenizer output changes vs FP8 inference failure on B200." }, { - "left": "issue:44279", - "right": "issue:46328", + "left": "issue:44297", + "right": "issue:45310", "accept": false, - "reason": "A dependency issue is not the same as a SequenceFeatureExtractor.pad performance regression." + "reason": "One is tokenizer save_pretrained metadata mismatch; the other is a Qwen3.5 from_pretrained model loading error. Related area, not the same bug." }, { - "left": "issue:42907", - "right": "issue:46050", + "left": "issue:41628", + "right": "issue:42757", "accept": false, - "reason": "Different quantization workflows: failed dequantized Ministrals/Devstrals saving vs sinq save/load breakage." + "reason": "Both are import errors, but for different symbols from different packages/dependencies; not the same underlying issue." }, { - "left": "issue:43723", - "right": "issue:44991", + "left": "issue:43475", + "right": "issue:46305", "accept": false, - "reason": "Both mention tokenizer loading in v5, but one is a general regression report and the other is a model-specific failure; not enough to treat as the same bug." + "reason": "Different models and different bugs: missing vision encoder attribute vs a sliding-window mask off-by-one." }, { - "left": "issue:44263", - "right": "issue:44360", + "left": "issue:44190", + "right": "issue:46082", "accept": false, - "reason": "Same broad component name, but torch.split return values and a missing ReLU are different defects." + "reason": "Unrelated: local dataset loading in an example script vs LlamaConfig head_dim validation." }, { - "left": "issue:30990", - "right": "issue:38175", + "left": "issue:43723", + "right": "issue:44977", "accept": false, - "reason": "Sentence-Transformers loading hang and SIGLIP2 zero probabilities are unrelated symptoms in different code paths." + "reason": "Tokenizer loading failure and flash-attention generation issue are unrelated." }, { - "left": "issue:45092", - "right": "issue:45357", + "left": "issue:44206", + "right": "issue:45103", "accept": false, - "reason": "InternVL2 meta-init incompatibility and Qwen3.5 visual encoder key saving are different model-specific issues." + "reason": "Different failure modes and code paths: feature extractor argument mismatch vs auto_docstring crash with future annotations." }, { - "left": "issue:43761", - "right": "issue:46082", + "left": "issue:45381", + "right": "issue:46305", "accept": false, - "reason": "CLIPVisionModel hidden_states regression and LlamaConfig head_dim validation are unrelated." + "reason": "Both mention positional behavior, but they affect different models and different mechanisms; not the same bug." }, { - "left": "issue:44829", - "right": "issue:46002", + "left": "issue:43232", + "right": "issue:46424", "accept": false, - "reason": "Degenerate FlashAttention3 training and NaNs in classification heads are different failure modes." + "reason": "Both are cache/generation related, but one concerns updating model kwargs after sync_gpus and the other cache_length semantics for quantized cache." }, { - "left": "issue:43644", - "right": "issue:46305", + "left": "issue:44734", + "right": "issue:46421", "accept": false, - "reason": "Junk non-persistent buffers and a sliding_window_mask off-by-one bug are not the same change." + "reason": "Serving KV-cache indexing crash vs cache lazy init causing recompilation are separate issues." }, { - "left": "issue:44261", - "right": "issue:45092", + "left": "issue:44222", + "right": "issue:44484", "accept": false, - "reason": "A missing rms_norm_eps in MLA q_a_layernorm is unrelated to InternVL2 meta initialization." + "reason": "Different save_pretrained topics: FP8 MoE serialization bug vs shard-size default question." }, { - "left": "issue:42831", - "right": "issue:45310", + "left": "issue:43232", + "right": "issue:45834", "accept": false, - "reason": "FineGrainedFP8 accuracy drift and Qwen3.5 MoE from_pretrained failure are different problems." + "reason": "Generation kwargs handling is unrelated to Kosmos2.5 long OCR index errors." }, { - "left": "issue:43479", - "right": "issue:44913", + "left": "issue:44295", + "right": "issue:44623", "accept": false, - "reason": "Phi4Multimodal default-config initialization and GPTNeoX rotary_pct reload behavior are unrelated." + "reason": "Buffer registration/read issue vs processor save_pretrained missing files; different components and symptoms." }, { - "left": "issue:41628", - "right": "issue:43824", + "left": "issue:44263", + "right": "issue:45092", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and likely different export paths." + "reason": "Different model code paths: DSA indexer split behavior vs remote-code checkpoint meta-init incompatibility." }, { - "left": "issue:43408", - "right": "issue:46402", + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "A sam3_video/sam3_tracker warning is not the same bug as Sam3Model text encoder weights not loading." + "reason": "A docstring mismatch is not the same as a RoPE implementation bug." }, { - "left": "issue:42831", - "right": "issue:45753", + "left": "issue:43723", + "right": "issue:44871", "accept": false, - "reason": "FP8 accuracy degradation and a Qwen3_5 infinite loop on one image do not share the same root cause." + "reason": "Tokenizer load regression and eos_token_id config mismatch are distinct problems." }, { - "left": "issue:44261", - "right": "issue:44373", + "left": "issue:44483", + "right": "issue:45406", "accept": false, - "reason": "A precision/config bug and a wrong docstring are not duplicates." + "reason": "One is an API request acceptance issue, the other a processor attribute error in serve; not the same bug." }, { - "left": "issue:39401", - "right": "issue:44568", + "left": "issue:44279", + "right": "issue:46328", "accept": false, - "reason": "Qwen3 offset_mapping issues and mdeberta-v3 BOS/EOS special-token behavior are distinct tokenizer bugs." + "reason": "Generic dependency issue vs a performance bug in SequenceFeatureExtractor.pad." }, { - "left": "issue:42886", - "right": "issue:42994", + "left": "issue:43723", + "right": "issue:44991", "accept": false, - "reason": "Offline tokenizer cache loading and quantized model saving failures are unrelated." + "reason": "Both are tokenizer-loading related, but they concern different models and likely different causes." }, { - "left": "issue:42886", - "right": "issue:43066", + "left": "issue:42907", + "right": "issue:46050", "accept": false, - "reason": "Tokenizer cache/offline loading is different from a wrong decoder type configuration in v5." + "reason": "Both involve save/load and quantization, but one is dequantized model saving and the other is sinq round-trip breakage; not the same concrete defect." }, { - "left": "issue:42757", - "right": "issue:45278", + "left": "issue:30990", + "right": "issue:38175", "accept": false, - "reason": "A specific is_offline_mode import error is not the same as a broad collection of import errors after upgrade." + "reason": "Completely different models and symptoms: loading hangs vs zero probabilities." }, { - "left": "issue:44368", - "right": "issue:46305", + "left": "issue:44263", + "right": "issue:44360", "accept": false, - "reason": "A tie_word_embeddings warning during LoRA fine-tuning is unrelated to a streaming mask off-by-one." + "reason": "Same broad area (DSA indexer) but different concrete bugs: split return behavior vs missing ReLU." }, { - "left": "issue:22355", - "right": "issue:28218", + "left": "issue:45092", + "right": "issue:45357", "accept": false, - "reason": "Missing transformers.onnx and extra tokenizer spacing after added tokens are different issues." + "reason": "Both are model-specific v5 compatibility problems, but they affect different checkpoints and different serialization/init paths." }, { - "left": "issue:42831", - "right": "issue:45753", + "left": "issue:43761", + "right": "issue:46082", "accept": false, - "reason": "Same model family is not enough; FP8 accuracy loss and an image-specific infinite loop are different bugs." + "reason": "Vision model hidden_states output bug vs LlamaConfig head_dim validation are unrelated." }, { - "left": "issue:43232", - "right": "issue:45593", + "left": "issue:43644", + "right": "issue:46305", "accept": false, - "reason": "Generation kwarg sync behavior and D-FINE auxiliary-loss behavior are unrelated." + "reason": "Non-persistent buffer initialization corruption is unrelated to sliding-window masking." }, { - "left": "issue:43408", - "right": "issue:46369", + "left": "issue:44829", + "right": "issue:46002", "accept": false, - "reason": "Sam3 model-type warning and memory overflow in sam3-litetext are different failure modes." + "reason": "Different training pathologies with different likely causes; not the same bug." }, { - "left": "issue:43576", - "right": "issue:43784", + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "RoPE/layernorm precision issue vs remote-code checkpoint meta-init incompatibility are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:45310", "accept": false, - "reason": "Broken transformers env CLI and a sentence-transformers import NameError are not the same underlying problem." + "reason": "FP8 accuracy degradation is unrelated to Qwen3.5 from_pretrained loading failure." } ] }, @@ -13347,311 +13364,297 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:22355", + "issue:27957", "issue:28218", "issue:30990", "issue:33357", "issue:39401", "issue:41628", "issue:42371", + "issue:42757", + "issue:42831", + "issue:42886", "issue:42907", - "issue:43116", + "issue:42994", + "issue:43066", "issue:43122", "issue:43232", "issue:43408", - "issue:43452", - "issue:43475", + "issue:43479", "issue:43550", + "issue:43576", "issue:43646", "issue:43756", "issue:43784", + "issue:43824", "issue:43937", - "issue:43976", "issue:44038", "issue:44060", "issue:44117", "issue:44188", - "issue:44222", - "issue:44261", "issue:44263", "issue:44295", "issue:44297", - "issue:44360", + "issue:44368", "issue:44373", - "issue:44485", - "issue:44488", - "issue:44514", "issue:44521", - "issue:44625", + "issue:44568", "issue:44704", - "issue:44743", - "issue:44779", - "issue:44829", + "issue:44913", "issue:44945", - "issue:44964", - "issue:45200", "issue:45278", - "issue:45290", - "issue:45600", + "issue:45593", + "issue:45753", "issue:46002", "issue:46132", - "issue:46326" + "issue:46305", + "issue:46326", + "issue:46369", + "issue:46402" ], "soft_pairs": [ - "issue:43756|issue:44060", + "issue:43479|issue:44913", + "issue:41628|issue:43824", + "issue:43408|issue:46402", + "issue:42831|issue:44038", + "issue:44263|issue:44373", + "issue:39401|issue:44568", + "issue:22355|issue:27957", + "issue:42757|issue:45278", + "issue:27957|issue:30990", + "issue:42886|issue:43066", + "issue:42886|issue:42994", + "issue:44368|issue:46305", + "issue:42831|issue:45753", + "issue:22355|issue:28218", + "issue:43408|issue:46369", + "issue:43232|issue:45593", + "issue:43576|issue:43784", "issue:43646|issue:46002", + "issue:43756|issue:44060", "issue:43550|issue:46132", - "issue:44117|issue:44704", "issue:44295|issue:44297", + "issue:44117|issue:44704", "issue:30990|issue:33357", - "issue:43122|issue:44521", "issue:42371|issue:44263", - "issue:44188|issue:44373", + "issue:43122|issue:44521", "issue:42907|issue:44945", + "issue:44188|issue:44373", "issue:43122|issue:46326", - "issue:39401|issue:44488", - "issue:42907|issue:43937", - "issue:44261|issue:44373", "issue:41628|issue:45278", - "issue:43976|issue:44945", - "issue:44704|issue:45290", - "issue:43784|issue:45600", - "issue:43232|issue:44743", - "issue:42907|issue:44964", - "issue:39401|issue:44521", - "issue:43116|issue:45200", - "issue:44625|issue:44829", - "issue:43122|issue:44514", - "issue:44038|issue:44297", - "issue:44360|issue:44485", - "issue:44038|issue:44295", - "issue:43408|issue:43475", - "issue:28218|issue:30990", - "issue:42907|issue:43452", - "issue:43122|issue:44779", - "issue:44038|issue:44222" + "issue:42907|issue:43937" ], - "prepared_review_unit_hash": "95038b7e95ca5796b49fd907abb31df80628d9b8554ff4c33722b4cdeac7b080", + "prepared_review_unit_hash": "6804ae1a4f0db8c3be9bb9383fb06d40d48c0cc9a1b98ec2405f4cb9ae4b1398", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36918, - "estimated_input_tokens": 9230, - "estimated_eval_tokens": 18716 + "soft_pair_count": 30, + "serialized_chars": 36184, + "estimated_input_tokens": 9046, + "estimated_eval_tokens": 18348 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it mixes unrelated tokenizer, model-loading, generation, attention, and API-compatibility bugs. I found no soft pair that is clearly the same underlying issue, so no merges are warranted.", + "summary": "This cluster is a loose semantic grouping of mostly unrelated Transformers issues (imports, tokenizers, model-specific regressions, and training/runtime bugs). None of the proposed soft pairs look like the same underlying bug/change, so all are rejected. The broad import-regression report is the best representative issue, but only as an umbrella, not a duplicate target for the others.", "confidence": 0.84, - "canonical_issue_reason": "Issue 43122 is the broadest tokenization-regression report in the set and the most plausible anchor for the tokenizer-related subcluster, even though the overall cluster is not a true duplicate set.", + "canonical_issue_reason": "issue:45278 is the broadest and most representative umbrella for the cluster\u2019s recurring import-regression theme after upgrade, even though it does not duplicate the other issues.", "canonical_pr_reason": null, - "best_issue_reason": "43122 is the most generally useful issue to keep as the representative anchor because it describes a version-to-version tokenizer behavior change rather than a narrow model-specific symptom.", + "best_issue_reason": "issue:45278 is the most suitable cluster representative because it captures a general user-facing regression pattern and can contextualize several related import complaints.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43756", - "right": "issue:44060", + "left": "issue:43479", + "right": "issue:44913", "accept": false, - "reason": "Different models and failures: RoPE layer dropping vs tied-weights warning across Qwen3-Next. Not the same bug." + "reason": "Both concern config handling, but they hit different models and different fields; not the same bug." }, { - "left": "issue:43646", - "right": "issue:46002", + "left": "issue:41628", + "right": "issue:43824", "accept": false, - "reason": "Both mention initialization, but one is custom model init breakage and the other is NaNs in classification heads. Different symptoms and code paths." + "reason": "Both are import errors, but for different missing symbols and different code paths." }, { - "left": "issue:43550", - "right": "issue:46132", + "left": "issue:43408", + "right": "issue:46402", "accept": false, - "reason": "One is a Bamba torch.compile+SDPA failure; the other is AttentionInterface.register changing callback behavior. Too broad and not the same fix." + "reason": "Same sam3 family, but one is a wrong model-type warning and the other is missing text-encoder weights." }, { - "left": "issue:44117", - "right": "issue:44704", + "left": "issue:42831", + "right": "issue:44038", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES None handling and AutoProcessor cached_file kwargs are separate API bugs." + "reason": "Different model-specific regressions: FP8 accuracy vs Qwen3-VL-Moe / Transformers 5 compatibility." }, { - "left": "issue:44295", - "right": "issue:44297", + "left": "issue:44263", + "right": "issue:44373", "accept": false, - "reason": "Position_ids buffer access and tokenizer_class mismatch on save_pretrained are unrelated." + "reason": "Unrelated problems: tensor split return values versus a docstring issue." }, { - "left": "issue:30990", - "right": "issue:33357", + "left": "issue:39401", + "right": "issue:44568", "accept": false, - "reason": "Loading hang in sentence-transformers vs MacOS bus error on CLIP are distinct failures." + "reason": "Both are tokenizer bugs, but they affect different tokenizers and different behaviors (offset mapping vs BOS/EOS insertion)." }, { - "left": "issue:43122", - "right": "issue:44521", + "left": "issue:22355", + "right": "issue:27957", "accept": false, - "reason": "Both involve tokenization/chat templating, but one is a tokenizer-version regression and the other is assistant_mask generation for multimodal inputs." + "reason": "ONNX import failure and Flash Attention model support are unrelated." }, { - "left": "issue:42371", - "right": "issue:44263", + "left": "issue:42757", + "right": "issue:45278", "accept": false, - "reason": "TF32 settings guidance and a torch.split return-value issue are unrelated." + "reason": "There is broad import-error overlap, but one is a specific missing hub export and the other is a general post-upgrade import regression; not the same concrete bug." }, { - "left": "issue:44188", - "right": "issue:44373", + "left": "issue:27957", + "right": "issue:30990", "accept": false, - "reason": "Torch.compile attention-kernel divergence and a docstring typo are not the same bug." + "reason": "Different subsystems and symptoms: attention backend support vs loading hang." }, { - "left": "issue:42907", - "right": "issue:44945", + "left": "issue:42886", + "right": "issue:43066", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals and wrong LLM output with pipeline parallelism are different problems." + "reason": "Offline cache loading failure is unrelated to wrong tokenizer decoder type in v5." }, { - "left": "issue:43122", - "right": "issue:46326", + "left": "issue:42886", + "right": "issue:42994", "accept": false, - "reason": "Both are tokenizer-related, but 43122 is a version regression while 46326 is a specific mistral_common continue_final_message bug." + "reason": "Tokenizer cache/offline behavior and quantized model saving are different bugs." }, { - "left": "issue:39401", - "right": "issue:44488", + "left": "issue:44368", + "right": "issue:46305", "accept": false, - "reason": "Qwen3 offset_mapping errors and failing to load cjvt/sleng-bert are different tokenizer issues." + "reason": "A warning about tie_word_embeddings is unrelated to a sliding-window mask off-by-one error." }, { - "left": "issue:42907", - "right": "issue:43937", + "left": "issue:42831", + "right": "issue:45753", "accept": false, - "reason": "Dequantized save_pretrained failure and GLM-5 GenerationConfig validation are unrelated." + "reason": "Accuracy issue in FP8 is not the same as an infinite-loop multimodal inference bug." }, { - "left": "issue:44261", - "right": "issue:44373", + "left": "issue:22355", + "right": "issue:28218", "accept": false, - "reason": "MLA layernorm epsilon precision issue and a wrong docstring for position_ids do not share a code path." + "reason": "Missing transformers.onnx module and extra-space tokenization behavior are unrelated." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:43408", + "right": "issue:46369", "accept": false, - "reason": "Single missing AutoImageProcessor import vs many import errors after a version bump are not the same concrete bug." + "reason": "Both touch sam3, but one is a model-type mismatch warning and the other is a memory overflow." }, { - "left": "issue:43976", - "right": "issue:44945", + "left": "issue:43232", + "right": "issue:45593", "accept": false, - "reason": "Python version incompatibility and pipeline-parallel incorrect output are unrelated." + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss logic are different code paths." }, { - "left": "issue:44704", - "right": "issue:45290", + "left": "issue:43576", + "right": "issue:43784", "accept": false, - "reason": "cached_file kwarg forwarding and chat-template crashes on tool-call assistant messages are distinct." + "reason": "A broken env CLI command is not the same as a NameError during sentence-transformers import." }, { - "left": "issue:43784", - "right": "issue:45600", + "left": "issue:43646", + "right": "issue:46002", "accept": false, - "reason": "Import NameError in sentence-transformers and stale auto_mappings references to removed configs are different breakages." + "reason": "Custom model initialization breakage and NaNs in classification heads are distinct failures." }, { - "left": "issue:43232", - "right": "issue:44743", + "left": "issue:43756", + "right": "issue:44060", "accept": false, - "reason": "Generation kwargs sync_gpus ordering and recurrent-state reset with cache are both generation-related but not the same bug." + "reason": "Different model-specific issues: RoPE layer retention vs tied-weights warning." }, { - "left": "issue:42907", - "right": "issue:44964", + "left": "issue:43550", + "right": "issue:46132", "accept": false, - "reason": "Dequantized save_pretrained failure and multimodal model loading failure are unrelated." + "reason": "torch.compile SDPA failure is unrelated to AttentionInterface.register changing function behavior." }, { - "left": "issue:39401", - "right": "issue:44521", + "left": "issue:44295", + "right": "issue:44297", "accept": false, - "reason": "Offset mapping errors and zero assistant masks are different tokenizer/template defects." + "reason": "Reading a registered buffer and tokenizer save metadata mismatch are different bugs." }, { - "left": "issue:43116", - "right": "issue:45200", + "left": "issue:44117", + "right": "issue:44704", "accept": false, - "reason": "Empty multi-label classification results and mm_token_type_ids defaulting are separate training/inference issues." + "reason": "One is None handling in tokenizer mapping; the other is missing kwargs propagation in AutoProcessor." }, { - "left": "issue:44625", - "right": "issue:44829", + "left": "issue:30990", + "right": "issue:33357", "accept": false, - "reason": "num_labels propagation and flash_attention_3 degenerate training are unrelated." + "reason": "Loading stuck and a MacOS bus error are different symptoms with different likely causes." }, { - "left": "issue:43122", - "right": "issue:44514", + "left": "issue:42371", + "right": "issue:44263", "accept": false, - "reason": "Tokenizer version regression and Qwen2_5_VLProcessor batched apply_chat_template crash are different problems." + "reason": "TF32 configuration guidance is unrelated to torch.split return values." }, { - "left": "issue:44038", - "right": "issue:44297", + "left": "issue:43122", + "right": "issue:44521", "accept": false, - "reason": "Qwen3-VL-Moe loading issue and tokenizer save_pretrained class mismatch are unrelated." + "reason": "Different tokenizer/template bugs: tokenization drift vs assistant mask generation for multimodal inputs." }, { - "left": "issue:44360", - "right": "issue:44485", + "left": "issue:42907", + "right": "issue:44945", "accept": false, - "reason": "A missing ReLU in the DSA indexer is not related to GLM-5 RoPE implementation." + "reason": "Quantized model saving failure and pipeline-parallel wrong output are unrelated runtime issues." }, { - "left": "issue:44038", - "right": "issue:44295", + "left": "issue:44188", + "right": "issue:44373", "accept": false, - "reason": "Qwen3-VL-Moe loading and position_ids buffer access are different bugs." + "reason": "Compiler-dependent attention-kernel divergence is not the same as a docstring problem." }, { - "left": "issue:43408", - "right": "issue:43475", + "left": "issue:43122", + "right": "issue:46326", "accept": false, - "reason": "SAM3 model-type mismatch warning and missing fpn_position_embeddings are both SAM3-adjacent but not the same underlying issue." + "reason": "Both involve tokenization, but one is tokenization drift across versions and the other is continue_final_message for mistral_common tokenizers." }, { - "left": "issue:28218", - "right": "issue:30990", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "Extra space after added token and a Sentence Transformers loading hang are unrelated." + "reason": "Both mention import errors, but they are not the same concrete missing import or code path." }, { "left": "issue:42907", - "right": "issue:43452", - "accept": false, - "reason": "Saving dequantized models and gguf_file breaking from_pretrained paths are distinct." - }, - { - "left": "issue:43122", - "right": "issue:44779", - "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models/behavior and do not look like one fixable code-path bug." - }, - { - "left": "issue:44038", - "right": "issue:44222", + "right": "issue:43937", "accept": false, - "reason": "Qwen3-VL-Moe loading and FP8 save_pretrained for moe are separate issues." + "reason": "Saving dequantized models and invalid GenerationConfig are different failure modes." } ] }, @@ -13673,282 +13676,295 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:22355", "issue:29127", "issue:29159", "issue:36296", - "issue:38617", + "issue:39401", "issue:42175", - "issue:42222", - "issue:42371", "issue:42673", "issue:42757", - "issue:43064", + "issue:42907", "issue:43066", "issue:43116", + "issue:43122", "issue:43278", - "issue:43441", + "issue:43408", + "issue:43452", + "issue:43475", "issue:43493", - "issue:43611", - "issue:43673", + "issue:43784", "issue:43824", - "issue:43825", - "issue:43874", - "issue:43937", "issue:43976", "issue:43992", "issue:44038", - "issue:44075", - "issue:44188", + "issue:44222", + "issue:44261", + "issue:44295", + "issue:44297", "issue:44351", - "issue:44355", + "issue:44360", + "issue:44373", "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", "issue:44704", "issue:44779", - "issue:44936", + "issue:44829", "issue:44938", "issue:44945", + "issue:44964", "issue:45072", - "issue:45083", + "issue:45200", "issue:45216", - "issue:45278", - "issue:45357", - "issue:45446", + "issue:45290", "issue:45478", - "issue:45563", - "issue:45753", - "issue:46002", - "issue:46231", - "issue:46328" + "issue:45600", + "issue:45753" ], "soft_pairs": [ + "issue:39401|issue:44488", + "issue:44261|issue:44373", + "issue:43976|issue:44945", + "issue:44704|issue:45290", + "issue:43784|issue:45600", + "issue:42907|issue:44964", + "issue:39401|issue:44521", + "issue:43116|issue:45200", + "issue:44038|issue:44297", + "issue:43122|issue:44514", + "issue:44625|issue:44829", + "issue:44360|issue:44485", + "issue:44038|issue:44295", + "issue:43408|issue:43475", + "issue:42907|issue:43452", + "issue:43122|issue:44779", + "issue:44038|issue:44222", "issue:45216|issue:45478", - "issue:36296|issue:43116", "issue:43824|issue:44351", - "issue:22355|issue:29159", + "issue:36296|issue:43116", "issue:42757|issue:44351", + "issue:22355|issue:29159", "issue:43278|issue:44945", "issue:42175|issue:44938", "issue:43493|issue:45753", - "issue:22355|issue:29127", "issue:43066|issue:44779", "issue:42673|issue:45072", - "issue:43992|issue:44704", - "issue:38617|issue:45278", - "issue:44355|issue:46231", - "issue:43937|issue:44485", - "issue:43611|issue:46002", - "issue:44188|issue:46328", - "issue:43116|issue:43493", - "issue:42222|issue:43441", - "issue:44075|issue:44936", - "issue:42371|issue:43874", - "issue:45357|issue:45478", - "issue:44038|issue:45478", - "issue:43976|issue:45446", - "issue:43116|issue:43825", - "issue:45083|issue:45446", - "issue:43064|issue:43673", - "issue:43937|issue:45563" + "issue:22355|issue:29127", + "issue:43992|issue:44704" ], - "prepared_review_unit_hash": "ff4d87a26a633f953b80f1014b3465397bc9d34cf0748b0f0673a2a9cb7a8154", + "prepared_review_unit_hash": "0cd1e82f7b2798d8db01e6155f1f0ceafb3286877dcd29c680dfb6fcaddf4a2d", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35130, - "estimated_input_tokens": 8783, - "estimated_eval_tokens": 17822 + "soft_pair_count": 29, + "serialized_chars": 35476, + "estimated_input_tokens": 8869, + "estimated_eval_tokens": 17994 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated issue reports with a few superficial overlaps around imports, Qwen3.5, and tokenizer/model-v5 regressions. No soft pair looks like the same concrete bug or change, so all soft edges are rejected.", - "confidence": 0.87, - "canonical_issue_reason": "issue:45278 is the broadest representative item in this set, covering the general v5 import-breakage theme that most closely matches the cluster\u2019s strongest recurring motif.", + "summary": "The set is mostly a collection of unrelated issues across tokenizer regressions, model-loading bugs, training/inference problems, and packaging/import errors. I found no strong duplicate pair with enough evidence to merge; the best anchor issue is the broad tokenization-regression report in 43122.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 43122 is the clearest broad regression report: it defines a version boundary, a reproducible tokenizer behavior change, and is more general than the model-specific or symptom-specific reports around it.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best single issue to anchor the cluster because it is broad, clearly user-facing, and more representative than the highly specific model- or API-edge reports.", + "best_issue_reason": "43122 is the strongest standalone issue to use as an anchor because it is concrete, versioned, and likely to attract follow-up analysis without being overly niche.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45216", - "right": "issue:45478", + "left": "issue:39401", + "right": "issue:44488", "accept": false, - "reason": "Both involve Qwen3.5 save/load behavior, but one is incorrect save_pretrained checkpoint keys and the other is a from_pretrained load error; different code paths and likely different fixes." + "reason": "Both are about Qwen-related text behavior, but one is an offset-mapping bug and the other is a model-loading failure; different code paths." }, { - "left": "issue:36296", - "right": "issue:43116", + "left": "issue:44261", + "right": "issue:44373", "accept": false, - "reason": "Tensor-parallel training and multi-label example-script output are unrelated bugs in different workflows." + "reason": "One is a numerical/config bug in MLA layernorm, the other is a docstring issue for position_ids; not the same defect." }, { - "left": "issue:43824", - "right": "issue:44351", + "left": "issue:43976", + "right": "issue:44945", "accept": false, - "reason": "Both are import errors, but for different symbols in different modules; no evidence they share the same root cause." + "reason": "Python compatibility/install failure vs incorrect output under pipeline parallelism; unrelated symptoms and fixes." }, { - "left": "issue:22355", - "right": "issue:29159", + "left": "issue:44704", + "right": "issue:45290", "accept": false, - "reason": "One is an onnx module import failure, the other is tokenizer slow/fast inconsistency; unrelated problems." + "reason": "AutoProcessor kwargs forwarding and chat-template tool-call handling are distinct APIs with different failure modes." }, { - "left": "issue:42757", - "right": "issue:44351", + "left": "issue:43784", + "right": "issue:45600", "accept": false, - "reason": "Different missing imports from different packages and APIs; not the same underlying bug." + "reason": "Import-time NameError in sentence-transformers integration vs CI break from stale auto-mapping references; not the same bug." }, { - "left": "issue:43278", - "right": "issue:44945", + "left": "issue:42907", + "right": "issue:44964", "accept": false, - "reason": "BF16-to-FP32 dtype drift during evaluation is a different issue from incorrect output under pipeline parallelism." + "reason": "Save/dequantization failure for Ministral/Devstral is unrelated to loading Phi-4 multimodal." }, { - "left": "issue:42175", - "right": "issue:44938", + "left": "issue:39401", + "right": "issue:44521", "accept": false, - "reason": "Packaging/backend dependency issue versus Python-version load failure; different failure modes." + "reason": "Both touch Qwen3.5 tokenization/chat templating, but offset_mapping and assistant_masks are different behaviors and code paths." }, { - "left": "issue:43493", - "right": "issue:45753", + "left": "issue:43116", + "right": "issue:45200", "accept": false, - "reason": "Different models and symptoms: SigLIP2 implementation mismatch vs Qwen3.5 image-specific infinite loop." + "reason": "Example script multi-label output bug vs Gemma 4 multimodal token-type defaulting; unrelated functionality." }, { - "left": "issue:22355", - "right": "issue:29127", + "left": "issue:44038", + "right": "issue:44297", "accept": false, - "reason": "onnx import failure and LayoutLMv3 error-message clarity are unrelated." + "reason": "Qwen3-VL-Moe loading bug and tokenizer_config save-pretrained mismatch are separate issues." }, { - "left": "issue:43066", - "right": "issue:44779", + "left": "issue:43122", + "right": "issue:44514", "accept": false, - "reason": "Tokenizer decoder-type regression and Deepseek tokenizer incorrect results are both tokenizer-related, but they are model-specific and not clearly the same bug." + "reason": "Versioned tokenizer regression vs batched chat-template crash with padding=False; different failure surfaces." }, { - "left": "issue:42673", - "right": "issue:45072", + "left": "issue:44625", + "right": "issue:44829", "accept": false, - "reason": "VRAM leak in multi-threaded Qwen3ForCausalLM inference is unrelated to dtype mismatch in bfloat16 inference." + "reason": "Config propagation for Qwen3.5 labels is unrelated to flash_attention_3 training degeneration." }, { - "left": "issue:43992", - "right": "issue:44704", + "left": "issue:44360", + "right": "issue:44485", "accept": false, - "reason": "Both involve from_pretrained, but one misses a model weight and the other drops kwargs in AutoProcessor; different concrete code paths." + "reason": "A missing ReLU in a DSA indexer is unrelated to a GLM-5 RoPE implementation discussion." }, { - "left": "issue:38617", - "right": "issue:45278", + "left": "issue:44038", + "right": "issue:44295", "accept": false, - "reason": "A specific missing import symbol is not the same bug as a broad report of many import errors after upgrade." + "reason": "Model-load compatibility issue vs a position_ids buffer-read error; no shared concrete bug path." }, { - "left": "issue:44355", - "right": "issue:46231", + "left": "issue:43408", + "right": "issue:43475", "accept": false, - "reason": "Running compiled Python files and a documentation parameter-name mismatch are unrelated." + "reason": "Both involve SAM 3 Video, but one is a config-type warning and the other is a missing attribute in encoder output; too different to merge." }, { - "left": "issue:43937", - "right": "issue:44485", + "left": "issue:42907", + "right": "issue:43452", "accept": false, - "reason": "GenerationConfig validation failure and GLM-5 RoPE implementation concerns are different bugs." + "reason": "Dequantized save failure and gguf_file load failure are different serialization paths." }, { - "left": "issue:43611", - "right": "issue:46002", + "left": "issue:43122", + "right": "issue:44779", "accept": false, - "reason": "base_model_prefix loading regression and NaNs in classification heads are different failure classes." + "reason": "Both mention tokenizer regressions in v5, but Deepseek-specific incorrect outputs are not the same concrete bug as the generic slow/fast-tokenizer inconsistency." }, { - "left": "issue:44188", - "right": "issue:46328", + "left": "issue:44038", + "right": "issue:44222", "accept": false, - "reason": "Attention-kernel divergence under torch.compile and SequenceFeatureExtractor.pad inefficiency are unrelated." + "reason": "Qwen3-VL-Moe loading regression and FP8 save_pretrained bug are unrelated model lifecycle failures." }, { - "left": "issue:43116", - "right": "issue:43493", + "left": "issue:45216", + "right": "issue:45478", "accept": false, - "reason": "Example-script multi-label output issue and SigLIP2 implementation mismatch do not share a concrete code path." + "reason": "Both concern Qwen3.5 and 5.4.0, but save_pretrained corruption and from_pretrained errors are different code paths and not clearly the same defect." }, { - "left": "issue:42222", - "right": "issue:43441", + "left": "issue:43824", + "right": "issue:44351", "accept": false, - "reason": "Broken VitPose models and FlashAttention failure for Ministral-3 are unrelated model-specific regressions." + "reason": "Both are import errors, but they involve different missing symbols from different modules and likely different causes." }, { - "left": "issue:44075", - "right": "issue:44936", + "left": "issue:36296", + "right": "issue:43116", "accept": false, - "reason": "Unused SGD args and train-then-evaluate failure are different training-loop bugs." + "reason": "Tensor-parallel training bug is unrelated to example-script multi-label classification behavior." }, { - "left": "issue:42371", - "right": "issue:43874", + "left": "issue:42757", + "right": "issue:44351", "accept": false, - "reason": "TF32 API guidance and a missing image-processor method are unrelated." + "reason": "Missing huggingface_hub symbol and missing transformers symbol are separate dependency/API breakages." }, { - "left": "issue:45357", - "right": "issue:45478", + "left": "issue:22355", + "right": "issue:29159", "accept": false, - "reason": "Both are Qwen3.5 regressions, but one is incorrect save_pretrained visual-encoder keys and the other is a from_pretrained error; not mergeable as one fix." + "reason": "Module import failure for transformers.onnx is unrelated to tokenizer slow/fast inconsistency." }, { - "left": "issue:44038", - "right": "issue:45478", + "left": "issue:43278", + "right": "issue:44945", "accept": false, - "reason": "Different Qwen3 variants and different symptoms; no clear shared underlying bug." + "reason": "Dtype drift between training and evaluate is unrelated to incorrect outputs under pipeline parallelism." }, { - "left": "issue:43976", - "right": "issue:45446", + "left": "issue:42175", + "right": "issue:44938", "accept": false, - "reason": "Python-version compatibility and an incorrect PyTorch version check are distinct compatibility bugs." + "reason": "Packaging extras missing TensorFlow and Python 3.14 load failure are different environment issues." }, { - "left": "issue:43116", - "right": "issue:43825", + "left": "issue:43493", + "right": "issue:45753", "accept": false, - "reason": "Example-script output regression and pipeline translation-task messaging are unrelated." + "reason": "SigLIP2 implementation mismatch and Qwen3_5 infinite-loop-on-image are different model-specific problems." }, { - "left": "issue:45083", - "right": "issue:45446", + "left": "issue:43066", + "right": "issue:44779", "accept": false, - "reason": "A model helper-function bug and a version-check bug in flex_attention are unrelated." + "reason": "Wrong tokenizer decoder type and Deepseek tokenization regression may both touch tokenizers, but they are not the same concrete failure." }, { - "left": "issue:43064", - "right": "issue:43673", + "left": "issue:42673", + "right": "issue:45072", "accept": false, - "reason": "Optimizer-state corruption under FSDP2/PEFT and missing GenerationMixin cache in chunked_prefill are different training/runtime bugs." + "reason": "VRAM leak in multi-threaded Qwen3ForCausalLM use is unrelated to dtype mismatches in bfloat16 inference." }, { - "left": "issue:43937", - "right": "issue:45563", + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Import error for transformers.onnx and a LayoutLMv3 error-message clarity issue are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype change and pipeline-parallel wrong output are distinct training/inference bugs." + }, + { + "left": "issue:43992", + "right": "issue:44704", "accept": false, - "reason": "Invalid generation config and stale num_return_sequences warning are separate generation-path issues." + "reason": "Missing embed_tokens.weight on load is unrelated to AutoProcessor kwargs forwarding." } ] }, @@ -13970,304 +13986,626 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:33453", "issue:38617", "issue:39401", - "issue:41628", "issue:42222", "issue:42371", - "issue:42490", - "issue:42617", "issue:42673", + "issue:43064", + "issue:43116", "issue:43122", - "issue:43408", - "issue:43421", + "issue:43441", "issue:43450", - "issue:43582", + "issue:43493", + "issue:43611", "issue:43673", - "issue:43742", + "issue:43825", "issue:43827", + "issue:43874", "issue:43906", "issue:43937", + "issue:43976", "issue:43986", + "issue:44038", + "issue:44075", "issue:44117", - "issue:44220", - "issue:44261", + "issue:44188", "issue:44291", + "issue:44355", "issue:44380", - "issue:44410", "issue:44466", - "issue:44484", + "issue:44485", "issue:44779", - "issue:44829", - "issue:44877", + "issue:44936", "issue:44977", "issue:45071", "issue:45081", - "issue:45216", - "issue:45310", - "issue:45341", - "issue:45362", + "issue:45083", + "issue:45278", + "issue:45357", "issue:45405", "issue:45446", - "issue:45561", + "issue:45478", + "issue:45563", "issue:45588", "issue:46002", "issue:46077", - "issue:46178", "issue:46231", "issue:46298", "issue:46328" ], "soft_pairs": [ + "issue:38617|issue:45278", + "issue:43937|issue:44485", + "issue:43611|issue:46002", + "issue:44188|issue:46328", + "issue:44355|issue:46231", + "issue:43116|issue:43493", + "issue:42222|issue:43441", + "issue:42371|issue:43874", + "issue:44075|issue:44936", + "issue:45357|issue:45478", + "issue:44038|issue:45478", + "issue:43976|issue:45446", + "issue:43116|issue:43825", + "issue:45083|issue:45446", + "issue:43064|issue:43673", "issue:33453|issue:39401", - "issue:44380|issue:46077", - "issue:44117|issue:45588", + "issue:43937|issue:45563", "issue:39401|issue:44779", + "issue:44117|issue:45588", + "issue:44380|issue:46077", "issue:42673|issue:43906", "issue:43827|issue:45446", "issue:43450|issue:43986", - "issue:43937|issue:44977", "issue:43673|issue:46328", "issue:44291|issue:45071", "issue:44466|issue:46002", + "issue:43937|issue:44977", "issue:45405|issue:45446", "issue:44117|issue:45081", - "issue:43122|issue:46298", - "issue:45341|issue:45561", - "issue:44410|issue:45310", - "issue:45341|issue:46328", - "issue:38617|issue:41628", - "issue:42222|issue:46178", - "issue:44829|issue:45216", - "issue:42617|issue:45341", - "issue:45362|issue:46328", - "issue:42673|issue:43582", - "issue:43408|issue:43742", - "issue:39401|issue:43421", - "issue:42490|issue:42673", - "issue:44877|issue:46231", - "issue:42371|issue:44261", - "issue:44484|issue:45310", - "issue:38617|issue:44220" + "issue:43122|issue:46298" ], - "prepared_review_unit_hash": "ea71fdab5b8deb5887c44a877f5d220909554671caded1dc84eb507574f29ca3", + "prepared_review_unit_hash": "8e7ad74bfe1e0056e9cdc5fa01c3c248cf6b3886ec704be0788872e4d99cdba2", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36574, - "estimated_input_tokens": 9144, - "estimated_eval_tokens": 18544 + "serialized_chars": 36562, + "estimated_input_tokens": 9141, + "estimated_eval_tokens": 18538 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of distinct Transformers bug reports rather than true duplicates. The only strong thematic overlap is around tokenizer regressions, but the soft-similarity pairs still differ in model, subsystem, or failure mode. No PRs are present.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 43122 is the best anchor: it describes a broad, cross-version tokenization regression with a clear before/after comparison, making it more representative than the narrower model-specific tokenizer reports.", + "summary": "The cluster is a grab-bag of mostly unrelated Transformers bugs/regressions. The broadest representative is the v5 import-regression report, while the soft edge candidates are mostly similar only at the subsystem/model-family level and should remain separate.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45278 is the most general and representative bug report in this set: a broad post-upgrade import-regression affecting multiple symbols, rather than a narrow model- or feature-specific failure.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43122 is the strongest global issue candidate because it is the most general and reusable regression report among the set, whereas the others are mostly narrow model- or API-specific bugs.", + "best_issue_reason": "issue:45278 best captures the cluster\u2019s dominant theme of v5 upgrade breakage and is broad enough to stand in for the general regression class.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Both are import errors after a version upgrade, but 38617 is a specific missing symbol while 45278 is a broader multi-import regression; they do not look like the same concrete bug." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM-5, but one is a GenerationConfig validation failure and the other is a RoPE implementation discussion; different failures and code paths." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Loading with base_model_prefix and NaNs in classification heads are unrelated symptoms with different root causes." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile and a padding performance issue are unrelated." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled Python-file errors and a documentation parameter-name mismatch are not the same underlying bug." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example-script multi-label output failure and SigLIP2 implementation discrepancy are different model/problem domains." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "VitPose loading breakage and Ministral FlashAttention failure are separate model-specific regressions." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance and a missing image-processor method are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer-argument handling and trainer.evaluate() failing after train() are distinct training bugs." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both concern Qwen3.5/MoE, but one is save_pretrained key serialization and the other is from_pretrained loading failure; not the same concrete defect." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Related model family, but these are different Qwen variants and different failure modes." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python-version compatibility and an incorrect PyTorch version check are unrelated." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "An example-script classification result bug is unrelated to a pipeline error-message/docs issue." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Feature-extraction length behavior and a flex_attention import/version check are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Optimizer-state corruption under FSDP2/PEFT and missing GenerationMixin cache in chunked prefill are different code paths." + }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Both involve tokenization/loading, but one is a generic tokenizer-loading regression and the other is a Qwen3 offset_mapping bug; different concrete failure modes." + "reason": "Both are tokenizer-related regressions, but one is loading and the other is offset_mapping correctness; not the same bug." }, { - "left": "issue:44380", - "right": "issue:46077", + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig invalidity and a stale warning about num_return_sequences are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:44779", "accept": false, - "reason": "Attention-scaling under SDPA/FlashAttention for GPT2 is a different code path from Gemma4 building causal masks for bidirectional attention." + "reason": "Different tokenizer regressions in different model families; no sign of the same underlying defect." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None and a video processor crash without torchvision are unrelated subsystems and bugs." + "reason": "Tokenizer mapping fallback handling and a flash_attention crash on s_aux=None are unrelated." }, { - "left": "issue:39401", - "right": "issue:44779", + "left": "issue:44380", + "right": "issue:46077", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different models and outputs: offset_mapping corruption vs incorrect Deepseek tokenization." + "reason": "Both involve attention behavior, but one is backend scaling under SDPA/FlashAttention and the other is Gemma4 bidirectional masking; different concrete bugs." }, { "left": "issue:42673", "right": "issue:43906", "accept": false, - "reason": "A VRAM leak in Qwen3ForCausalLM across dataloader threads is not the same bug as an isolated reproduction of an earlier issue." + "reason": "VRAM leakage in multi-threaded dataloader use and an isolated reproduction of another issue are not enough to treat as the same bug." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs still referencing pipeline() and a PyTorch version check for AuxRequest import are unrelated." + "reason": "Outdated docs referencing pipeline() and a PyTorch version check bug are unrelated." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "Batched video processor shape handling and an AutoProcessor torchvision import crash are different defects." - }, - { - "left": "issue:43937", - "right": "issue:44977", - "accept": false, - "reason": "GLM-5 invalid GenerationConfig and Qwen3.5 flash-attention generation failure are different model/config paths." + "reason": "Batched video processor shape handling and a missing torchvision crash are distinct video-processing issues." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Chunked-prefill cache handling and a SequenceFeatureExtractor.pad performance issue are unrelated." + "reason": "Generation cache missing in chunked prefill and a padding performance inefficiency are unrelated." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "Unexpected _is_hf_initialized argument during init_empty_weights is a different bug from v5.4.0 PreTrainedConfig type checking." + "reason": "init_empty_weights constructor-arg breakage and PretrainedConfig type-checking regression are both v5 regressions but not the same code-path problem." }, { "left": "issue:44466", "right": "issue:46002", "accept": false, - "reason": "lm_head.weight serialization inconsistency and NaNs in classification heads are different failure classes." + "reason": "lm_head.weight serialization inconsistency and NaNs in classification heads are unrelated." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GenerationConfig validation and a flash-attention generation failure in a different model are separate issues." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "A PEFT minimum-version bump and a flex_attention version check are unrelated dependency issues." + "reason": "A dependency version pinning problem and an import-version check bug are unrelated." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "A tokenizer mapping None bug and a Mistral regex patch AttributeError are both tokenizer-related but not the same underlying issue." + "reason": "Tokenizer mapping None-handling and a Mistral regex patch crash are both tokenizer-loading adjacent, but they are different failures and should not be merged." }, { "left": "issue:43122", "right": "issue:46298", "accept": false, - "reason": "Tokenization behavior regression and a noisy logging message are not duplicates." - }, - { - "left": "issue:45341", - "right": "issue:45561", - "accept": false, - "reason": "A testing_utils bug and a pytest-xdist race in captured_info.txt are distinct; the latter is a specific concurrency issue." - }, + "reason": "Tokenizer output changes and a noisy logging message are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 46, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42490", + "issue:42617", + "issue:42673", + "issue:42831", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43937", + "issue:44075", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44877", + "issue:44933", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45561", + "issue:45656", + "issue:46178", + "issue:46231", + "issue:46328" + ], + "soft_pairs": [ + "issue:44410|issue:45310", + "issue:45341|issue:45561", + "issue:38617|issue:41628", + "issue:45341|issue:46328", + "issue:42222|issue:46178", + "issue:44829|issue:45216", + "issue:42617|issue:45341", + "issue:45362|issue:46328", + "issue:42673|issue:43582", + "issue:43408|issue:43742", + "issue:39401|issue:43421", + "issue:42490|issue:42673", + "issue:44877|issue:46231", + "issue:42371|issue:44261", + "issue:44484|issue:45310", + "issue:38617|issue:44220", + "issue:43937|issue:44496", + "issue:44933|issue:45278", + "issue:43066|issue:43122", + "issue:42222|issue:45542", + "issue:44410|issue:45478", + "issue:44483|issue:45464", + "issue:42371|issue:42831", + "issue:43012|issue:43577", + "issue:44075|issue:45656", + "issue:42175|issue:42222", + "issue:43097|issue:44493", + "issue:43673|issue:45341", + "issue:41628|issue:45071", + "issue:44263|issue:44485", + "issue:45071|issue:45542" + ], + "prepared_review_unit_hash": "28929648be1f9614d2836d7689ec2cadbe958fd16249b2ce19a36f11804e1e48", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36530, + "estimated_input_tokens": 9133, + "estimated_eval_tokens": 18522 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly false positives across unrelated regressions, import errors, model-specific bugs, and docs/perf issues. Only the tokenizer pair 43066/43122 looks like the same v5 tokenizer regression.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 43066 is the clearest root-cause report for the tokenizer v5 regression; 43122 reads like the same symptom reported from a different angle.", + "canonical_pr_reason": null, + "best_issue_reason": "43066 is the best single issue to keep: it is specific, actionable, and can subsume the duplicate symptom in 43122.", + "best_pr_reason": null, + "soft_edge_verdicts": [ { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing Qwen3next projections in layer 0 and Qwen3.5 Moe from_pretrained failure are related only by model family, not by the same concrete bug." + "reason": "Different model families and failures: missing projections in qwen3next vs Qwen3.5 MoE from_pretrained error." }, { "left": "issue:45341", - "right": "issue:46328", + "right": "issue:45561", "accept": false, - "reason": "A testing utility bug and a SequenceFeatureExtractor.pad inefficiency are unrelated." + "reason": "Both mention testing utils, but one is a vague bug report and the other is a specific xdist race on captured_info.txt; not clearly the same bug." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but for different missing symbols and different code paths." + "reason": "Both are import errors, but for different symbols and code paths." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "Testing utils bug vs a padding performance issue in SequenceFeatureExtractor; unrelated." }, { "left": "issue:42222", "right": "issue:46178", "accept": false, - "reason": "Vitpose model breakage and a BACKENDS_MAPPING tensorflow_text error are unrelated." + "reason": "VitPose model breakage and tensorflow_text backend mapping are unrelated." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "FlashAttention-3 degenerate training and Qwen3.5 save_pretrained serialization are different bugs." + "reason": "Flash-attention training degeneration and Qwen3.5 save_pretrained checkpoint corruption are different bugs." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "A 3d_parallel.py execution failure and a testing_utils bug are unrelated." + "reason": "A script launch failure is unrelated to a testing_utils bug." }, { "left": "issue:45362", "right": "issue:46328", "accept": false, - "reason": "Qwen3.5 chat crashing and a SequenceFeatureExtractor.pad inefficiency are unrelated." + "reason": "Qwen3.5 chat crash vs padding performance; no shared codepath." }, { "left": "issue:42673", "right": "issue:43582", "accept": false, - "reason": "A VRAM leak in Qwen3ForCausalLM and an Apple Silicon TypeError in caching_allocator_warmup are unrelated." + "reason": "VRAM leak in dataloader threads and an Apple Silicon warmup TypeError are separate issues." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "A sam3_video/sam3_tracker warning and a MobileLLM-125M key error are different issues." + "reason": "Different loading warnings/errors for different model types." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Both concern tokenizers, but one is wrong offset_mapping and the other is post-processor not updating after runtime special-token changes." + "reason": "Tokenizer offset mapping bug vs runtime special-token post-processor update; related area but not the same defect." }, { "left": "issue:42490", "right": "issue:42673", "accept": false, - "reason": "Local-path trailing-slash import warning and a multi-thread VRAM leak are unrelated." + "reason": "Local path slash import warning and VRAM leak are unrelated." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Strict config blocking granite_speech loading and a docs parameter-name mismatch are not the same bug." + "reason": "Strict config loading vs a documentation parameter-name mismatch." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 behavior settings and an MLA q_a_layernorm precision/config issue are unrelated." + "reason": "TF32 API warning and MLA rms_norm_eps precision issue are unrelated." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "A question about max_shard_size and a Qwen3.5 Moe loading error are unrelated." + "reason": "Shard-size question vs Qwen3.5 MoE loading error; different concerns." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "Different import/runtime failures in unrelated modules; no shared concrete code path." + "reason": "ImportError for a missing config_utils symbol vs an fbank feature-extraction bug." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GenerationConfig validation failure and unrecognized model loading are different issues." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing image_utils import is not the same as a broad many-import-errors regression." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": true, + "reason": "Both point to the same v5 tokenizer regression: changed decoder/type behavior leading to different tokenization." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose model breakage and missing TensorFlow backend after tensorboard install are unrelated." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Same model family, but different concrete failures: missing projections vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Different chat/completions API failures and likely different request/streaming paths." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 behavior warning and FineGrainedFP8 accuracy regression are unrelated." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both involve dtype/precision, but one is a compile warning and the other is incorrect model dtype propagation on load." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "Optimizer args ignored and optimizer step called twice are distinct optimizer bugs." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow backend installation issue vs broken VitPose models; unrelated." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed tie_embeddings API and unexpected key warnings from position ids are different regressions." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing in chunked_prefill is unrelated to a testing_utils bug." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Import error for AutoImageProcessor vs PretrainedConfig type-checking regression; not the same bug." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "torch.split return-value issue in GlmMoeDsaIndexer vs GLM-5 RoPE implementation; different codepaths." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "PretrainedConfig type-checking regression and undefined tf backend error are unrelated." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The clustering is mostly conservative and grounded in the packet. The only accepted pair, 43066/43122, is plausibly the same tokenizer-v5 regression based on the titles, and the other pair judgments stay on the cautious side of false unless there is stronger body-level evidence not shown here." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -14275,85 +14613,70 @@ "label": "issue", "component_index": 1, "component_count": 19, - "review_unit_index": 46, + "review_unit_index": 47, "review_unit_count": 53, "status": "reviewed", "reason": null, "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ - "issue:41628", - "issue:42175", - "issue:42222", - "issue:42371", - "issue:42831", - "issue:43012", - "issue:43066", + "issue:27957", + "issue:38617", + "issue:42757", + "issue:42890", + "issue:43064", "issue:43097", - "issue:43122", "issue:43335", + "issue:43352", "issue:43441", "issue:43452", - "issue:43577", + "issue:43502", + "issue:43531", "issue:43673", + "issue:43873", "issue:43874", "issue:43931", - "issue:43937", "issue:43986", + "issue:43992", "issue:44038", - "issue:44075", + "issue:44062", "issue:44117", "issue:44261", "issue:44263", + "issue:44315", + "issue:44351", "issue:44360", "issue:44410", - "issue:44483", "issue:44484", "issue:44485", - "issue:44493", - "issue:44496", "issue:44829", "issue:44933", "issue:44987", "issue:45003", "issue:45071", - "issue:45278", "issue:45341", "issue:45372", + "issue:45375", "issue:45440", - "issue:45464", "issue:45478", - "issue:45542", - "issue:45656", + "issue:45588", "issue:45753", + "issue:46050", "issue:46123", "issue:46172", "issue:46231", + "issue:46298", "issue:46369" ], "soft_pairs": [ - "issue:43937|issue:44496", - "issue:44933|issue:45278", - "issue:43066|issue:43122", - "issue:42222|issue:45542", - "issue:44483|issue:45464", - "issue:44410|issue:45478", - "issue:42371|issue:42831", - "issue:44075|issue:45656", - "issue:43012|issue:43577", - "issue:42175|issue:42222", - "issue:43097|issue:44493", - "issue:43673|issue:45341", - "issue:41628|issue:45071", - "issue:44263|issue:44485", - "issue:45071|issue:45542", + "issue:27957|issue:43352", "issue:43874|issue:44261", - "issue:44263|issue:46172", - "issue:44360|issue:46172", "issue:44484|issue:45478", + "issue:44263|issue:46172", "issue:43931|issue:44410", + "issue:44360|issue:46172", "issue:43441|issue:44829", "issue:43931|issue:44038", "issue:43335|issue:45071", @@ -14367,240 +14690,247 @@ "issue:43986|issue:44484", "issue:43874|issue:44263", "issue:43986|issue:45372", - "issue:45440|issue:46172" + "issue:45440|issue:46172", + "issue:43531|issue:45341", + "issue:43097|issue:46050", + "issue:44261|issue:44485", + "issue:43097|issue:45588", + "issue:43064|issue:44315", + "issue:38617|issue:42757", + "issue:44062|issue:44117", + "issue:44410|issue:45375", + "issue:44038|issue:44410", + "issue:43502|issue:43992", + "issue:42890|issue:46298", + "issue:38617|issue:44351", + "issue:43097|issue:43873" ], - "prepared_review_unit_hash": "06b6ada1cf801703c5cc8ce3cd97ea637c2a9cabceffafd8416ddfa2fefb1cd7", + "prepared_review_unit_hash": "153be6571f80e530131df75892e1404c769ee256f50b6b994c369558a13da3b9", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37534, - "estimated_input_tokens": 9384, - "estimated_eval_tokens": 19024 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36772, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is a loose collection of Transformers v5 regression reports and model-specific bugs, with only shallow title similarity between many pairs. The broadest representative issue is the umbrella report about widespread import breakage after upgrading to 5.5.0; there is no PR in this set.", - "confidence": 0.39, - "canonical_issue_reason": "issue:45278 is the broadest umbrella: it captures multiple import failures after a version upgrade, which is closer to the overall cluster theme than the more narrowly scoped model/backend-specific reports.", + "summary": "This cluster is mostly soft-similarity false positives: the paired issues generally share broad themes like FlashAttention, import errors, or model-loading failures, but not the same concrete bug or code path. I do not see a true duplicate pair here.", + "confidence": 0.93, + "canonical_issue_reason": "No single issue is a clear canonical duplicate target: the items span unrelated bugs across different models and subsystems. If a representative issue is needed, it should be a broad, active bug report rather than one of the narrow model-specific cases.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best standalone issue for the cluster because it is the most general and representative of the import/API breakage pattern seen across the set.", + "best_issue_reason": "issue:43873 is the strongest standalone issue in the set: it is open, has substantial discussion/inbound references, and describes a substantive, broadly relevant bug rather than a narrow one-off failure.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43937", - "right": "issue:44496", - "accept": false, - "reason": "Both are model-loading/config problems, but one is a GenerationConfig validation failure and the other is an unrecognized-model/model_type issue; different concrete bugs." - }, - { - "left": "issue:44933", - "right": "issue:45278", + "left": "issue:27957", + "right": "issue:43352", "accept": false, - "reason": "Both mention import failures, but 44933 is a specific missing import from image_utils while 45278 is a broad upgrade-related import regression report." + "reason": "Both mention Flash Attention 2, but they concern different models and likely different model-compatibility checks." }, { - "left": "issue:43066", - "right": "issue:43122", + "left": "issue:43874", + "right": "issue:44261", "accept": false, - "reason": "Both involve tokenizer behavior changes, but one is about an incorrect decoder type and the other about changed tokenization output across versions; not the same bug." + "reason": "Different problems: a missing image-patch method in GLM46V versus an MLA layernorm config precision issue." }, { - "left": "issue:42222", - "right": "issue:45542", + "left": "issue:44484", + "right": "issue:45478", "accept": false, - "reason": "One is a ViTPose model breakage; the other is a backend-installation problem with TensorFlow/tensorboard. Different subsystems and causes." + "reason": "Unrelated: shard-size documentation/question versus a Qwen3.5 MoE loading error." }, { - "left": "issue:44483", - "right": "issue:45464", + "left": "issue:44263", + "right": "issue:46172", "accept": false, - "reason": "Both are chat/completions failures, but they affect different endpoints/symptoms and do not look like the same code-path defect." + "reason": "Different code paths and models: GlmMoeDsaIndexer split handling versus DeepSeek comb transpose logic." }, { - "left": "issue:44410", - "right": "issue:45478", + "left": "issue:43931", + "right": "issue:44410", "accept": false, - "reason": "Different Qwen-family model failures: missing projections in one layer versus a from_pretrained error for a MoE model. Not the same underlying issue." + "reason": "Both are Qwen-family model bugs, but one is a weight-shape mismatch and the other is missing projections in layer 0." }, { - "left": "issue:42371", - "right": "issue:42831", + "left": "issue:44360", + "right": "issue:46172", "accept": false, - "reason": "Unrelated precision topics: TF32 configuration guidance versus a FineGrainedFP8 accuracy issue." + "reason": "DSA indexer ReLU concern is not the same bug as the DeepSeek-v4 comb dtype/transpose issue." }, { - "left": "issue:44075", - "right": "issue:45656", + "left": "issue:43441", + "right": "issue:44829", "accept": false, - "reason": "Both concern optimizers, but one is about SGD arguments being ignored and the other about double optimizer steps under DeepSpeed; different bugs." + "reason": "Both involve FlashAttention, but one is a model-specific failure and the other is degenerate training with flash_attention_3." }, { - "left": "issue:43012", - "right": "issue:43577", + "left": "issue:43931", + "right": "issue:44038", "accept": false, - "reason": "Both mention dtype/precision, but one is a warning during bfloat16 compilation and the other is a loading bug where model/qformer dtype stays float32." + "reason": "Same general Qwen/VL area, but different failures: shape mismatch versus a Transformers v5 compatibility bug." }, { - "left": "issue:42175", - "right": "issue:42222", + "left": "issue:43335", + "right": "issue:45071", "accept": false, - "reason": "Backend dependency installation issue versus a broken model family; no shared concrete failure." + "reason": "SwitchTransformers sparse-layer construction and PretrainedConfig type checking are unrelated bugs." }, { "left": "issue:43097", - "right": "issue:44493", + "right": "issue:43452", "accept": false, - "reason": "One is about a removed config flag, the other about unexpected position-id keys. Similar release-era regressions, but not the same defect." + "reason": "An API-removal/documentation issue is not the same as gguf_file breaking tokenizer/model loading." }, { - "left": "issue:43673", + "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "A generation-cache bug under chunked prefill versus a small testing_utils bug; unrelated code paths." + "reason": "Different files and symptoms: unsafe sys.modules access versus a testing_utils bug." }, { - "left": "issue:41628", - "right": "issue:45071", + "left": "issue:43673", + "right": "issue:46123", "accept": false, - "reason": "ImportError for AutoImageProcessor is unrelated to a PretrainedConfig type-checking regression." + "reason": "Both affect generation/pipeline behavior, but the cache-missing chunked_prefill bug is distinct from the final-batch is_last bug." }, { - "left": "issue:44263", - "right": "issue:44485", + "left": "issue:43986", + "right": "issue:45341", "accept": false, - "reason": "Both are GLM-related, but one is about a torch.split/indexer bug and the other about RoPE implementation; different concrete defects." + "reason": "A video AutoProcessor import crash is unrelated to the testing_utils issue." }, { - "left": "issue:45071", - "right": "issue:45542", + "left": "issue:44933", + "right": "issue:46231", "accept": false, - "reason": "A type-checking regression in PretrainedConfig is unrelated to missing TensorFlow backend installation behavior." + "reason": "Nonexistent image_utils import and a doc parameter-name mismatch are different classes of issues." }, { - "left": "issue:43874", - "right": "issue:44261", + "left": "issue:45753", + "right": "issue:46369", "accept": false, - "reason": "One is an image-processor method missing from GLM46V; the other is an MLA layernorm epsilon/config precision mismatch." + "reason": "Different models and failures: Qwen3_5 infinite loop on one image versus SAM3-lite memory overflow." }, { - "left": "issue:44263", - "right": "issue:46172", + "left": "issue:44117", + "right": "issue:44987", "accept": false, - "reason": "Different model families and different tensor-shape/dtype issues; not mergeable as one fix." + "reason": "Tokenizer-mapping None handling is not the same as the physical-intelligence/fast loading failure." }, { - "left": "issue:44360", - "right": "issue:46172", + "left": "issue:43986", + "right": "issue:44484", "accept": false, - "reason": "A missing ReLU in the DSA indexer is unrelated to the DeepSeek V4 tensor operation report." + "reason": "Completely different topics: torchvision-dependent processor loading versus shard-size behavior." }, { - "left": "issue:44484", - "right": "issue:45478", + "left": "issue:43874", + "right": "issue:44263", "accept": false, - "reason": "Shard-size policy question versus a model-loading regression; no shared bug." + "reason": "Different GLM46V vs GlmMoeDsaIndexer failures; same broad area only by naming." }, { - "left": "issue:43931", - "right": "issue:44410", + "left": "issue:43986", + "right": "issue:45372", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch versus missing attention projections in qwen3next; different model internals and symptoms." + "reason": "Both are loading/import failures, but they involve different optional dependencies and different processors/models." }, { - "left": "issue:43441", - "right": "issue:44829", + "left": "issue:45440", + "right": "issue:46172", "accept": false, - "reason": "FlashAttention failure on Ministral-3 versus degenerate training with flash_attention_3; different code paths and outcomes." + "reason": "Native DeepseekV3 divergence is a model-implementation mismatch, not the DeepSeek-v4 comb transpose bug." }, { - "left": "issue:43931", - "right": "issue:44038", + "left": "issue:43531", + "right": "issue:45341", "accept": false, - "reason": "Both involve Qwen3-VL/Qwen3-VL-Moe, but one is a weight-shape mismatch and the other is a separate from_pretrained bug; not the same issue." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the testing_utils bug." }, { - "left": "issue:43335", - "right": "issue:45071", + "left": "issue:43097", + "right": "issue:46050", "accept": false, - "reason": "SwitchTransformers sparse-layer construction bug versus a PretrainedConfig typing regression; unrelated." + "reason": "A v5 API removal issue and a quantization save/load regression are different bugs." }, { - "left": "issue:43097", - "right": "issue:43452", + "left": "issue:44261", + "right": "issue:44485", "accept": false, - "reason": "A removed config attribute is unrelated to gguf_file loading failures in tokenizer/model loaders." + "reason": "Both are model-specific implementation discussions, but the missing rms_norm_eps bug is distinct from GLM-5 RoPE implementation." }, { - "left": "issue:45003", - "right": "issue:45341", + "left": "issue:43097", + "right": "issue:45588", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils versus a testing utility bug; different modules and failure modes." + "reason": "Tie-embeddings removal and flash_attention.py crashing on sink-less models are unrelated." }, { - "left": "issue:43673", - "right": "issue:46123", + "left": "issue:43064", + "right": "issue:44315", "accept": false, - "reason": "Generation cache handling versus MaskGenerationPipeline batch-finalization logic; different pipeline layers and bugs." + "reason": "FSDP2/PEFT optimizer-state corruption is a different trainer bug from Liger Kernel not being applied with model_init." }, { - "left": "issue:43986", - "right": "issue:45341", + "left": "issue:38617", + "right": "issue:42757", "accept": false, - "reason": "AutoProcessor/torchvision crash versus a small testing_utils bug; not the same underlying defect." + "reason": "Both are ImportError reports, but they are for different missing symbols in different packages." }, { - "left": "issue:44933", - "right": "issue:46231", + "left": "issue:44062", + "right": "issue:44117", "accept": false, - "reason": "A missing import from image_utils is unrelated to documentation naming mismatch in set_training()." + "reason": "AddedToken special-argument conflict and tokenizer-mapping None handling are separate tokenizer bugs." }, { - "left": "issue:45753", - "right": "issue:46369", + "left": "issue:44410", + "right": "issue:45375", "accept": false, - "reason": "Infinite-loop generation on a specific image versus memory overflow in sam3-litetext; distinct symptoms and likely distinct causes." + "reason": "Different Qwen3-family config/model issues: missing projections versus a silently dropped config field." }, { - "left": "issue:44117", - "right": "issue:44987", + "left": "issue:44038", + "right": "issue:44410", "accept": false, - "reason": "Both involve tokenizer loading, but one is a mapping-NONE assumption and the other is a package-specific load failure; too different to merge." + "reason": "Both involve Qwen3 variants, but one is a Transformers v5 loading break and the other is missing layer projections." }, { - "left": "issue:43986", - "right": "issue:44484", + "left": "issue:43502", + "right": "issue:43992", "accept": false, - "reason": "Processor-loading crash without torchvision is unrelated to a save_pretrained shard-size question." + "reason": "local_files_only network leakage and missing embed_tokens.weight are different from load-path bugs." }, { - "left": "issue:43874", - "right": "issue:44263", + "left": "issue:42890", + "right": "issue:46298", "accept": false, - "reason": "Different GLM model components and different missing-method vs tensor-return bugs; not the same code path." + "reason": "Test flakiness from missing set_seed() is unrelated to a noisy logging message." }, { - "left": "issue:43986", - "right": "issue:45372", + "left": "issue:38617", + "right": "issue:44351", "accept": false, - "reason": "Missing torchvision for video model loading and missing ReasoningEffort import for Gemma 4 are both import-related, but they affect different dependencies and processors." + "reason": "Two different import failures; they do not point to the same missing symbol or code path." }, { - "left": "issue:45440", - "right": "issue:46172", + "left": "issue:43097", + "right": "issue:43873", "accept": false, - "reason": "Both are DeepSeek-family reports, but one is about divergence from remote implementation and the other about a specific tensor transpose/dtype expression." + "reason": "Tie-embeddings removal and quantization offloading problems are different subsystems and bugs." } ] }, @@ -14615,18 +14945,19 @@ "label": "issue", "component_index": 1, "component_count": 19, - "review_unit_index": 47, + "review_unit_index": 48, "review_unit_count": 53, "status": "reviewed", "reason": null, "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:36683", "issue:38617", + "issue:42490", "issue:42757", - "issue:42890", "issue:42907", "issue:43012", "issue:43064", @@ -14635,58 +14966,46 @@ "issue:43208", "issue:43352", "issue:43502", - "issue:43531", + "issue:43526", "issue:43550", + "issue:43582", "issue:43824", "issue:43825", - "issue:43873", "issue:43906", "issue:43976", "issue:43986", "issue:43992", "issue:44038", - "issue:44062", "issue:44117", "issue:44162", - "issue:44261", - "issue:44315", - "issue:44351", + "issue:44222", + "issue:44230", "issue:44368", "issue:44380", - "issue:44410", + "issue:44483", "issue:44484", - "issue:44485", + "issue:44514", "issue:44655", + "issue:44683", "issue:44704", + "issue:44936", "issue:44938", "issue:45103", - "issue:45341", - "issue:45375", + "issue:45478", "issue:45479", - "issue:45588", + "issue:45600", "issue:45712", "issue:45753", - "issue:46050", "issue:46132", "issue:46172", "issue:46178", - "issue:46298" + "issue:46298", + "issue:46365", + "issue:46421" ], "soft_pairs": [ - "issue:43531|issue:45341", - "issue:44261|issue:44485", - "issue:43097|issue:46050", - "issue:43097|issue:45588", - "issue:43064|issue:44315", - "issue:38617|issue:42757", - "issue:44062|issue:44117", - "issue:44410|issue:45375", - "issue:44038|issue:44410", - "issue:43502|issue:43992", - "issue:42890|issue:46298", - "issue:38617|issue:44351", - "issue:43097|issue:43873", "issue:44704|issue:45103", + "issue:44380|issue:46421", "issue:43986|issue:44117", "issue:43824|issue:43986", "issue:44938|issue:46178", @@ -14694,239 +15013,250 @@ "issue:43208|issue:43550", "issue:43992|issue:44484", "issue:43012|issue:43550", - "issue:43825|issue:44655", "issue:43012|issue:44368", + "issue:43825|issue:44655", "issue:45712|issue:46298", "issue:42907|issue:44655", "issue:42907|issue:45753", - "issue:43352|issue:44380", "issue:43097|issue:46298", + "issue:43352|issue:44380", "issue:43352|issue:43976", "issue:44162|issue:46172", "issue:44380|issue:45479", "issue:44380|issue:46132", - "issue:43064|issue:44380" + "issue:43064|issue:44380", + "issue:44483|issue:44514", + "issue:44222|issue:45478", + "issue:36683|issue:43582", + "issue:43208|issue:43526", + "issue:44936|issue:46365", + "issue:42490|issue:43012", + "issue:43012|issue:44683", + "issue:42757|issue:43824", + "issue:38617|issue:43502", + "issue:45600|issue:46298", + "issue:44038|issue:44230" ], - "prepared_review_unit_hash": "d88f6b7b57277c6f70615e346e2d86050172e60a8bd6ae060425958c875f40ae", + "prepared_review_unit_hash": "c132f81ef0451e713aba86b97a079d4487e77844d69ddb1229c8902e47073b7b", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 33, - "serialized_chars": 36870, - "estimated_input_tokens": 9218, - "estimated_eval_tokens": 18692 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37093, + "estimated_input_tokens": 9274, + "estimated_eval_tokens": 18804 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is over-broad and mostly contains unrelated closed issues with only loose theme overlap around Transformers v5 regressions, model-specific bugs, and import/load failures. None of the suggested soft pairs looks like a true duplicate or same underlying fix.", - "confidence": 0.91, - "canonical_issue_reason": "If a hub must be chosen, issue 43097 is the broadest v5 API/regression complaint and has inbound references, but it is not a reliable duplicate canonical for the rest of this mixed cluster.", + "summary": "The soft pairs are overwhelmingly false positives: they span unrelated import errors, model-specific config regressions, backend/compile issues, and save/load behaviors. I don\u2019t see any pair that clearly shares the same underlying bug or change.", + "confidence": 0.93, + "canonical_issue_reason": "Issue #43824 is the cleanest high-signal regression report here: a specific ImportError, a clear user-facing failure mode, and multiple inbound references. It is more concrete and representative than the other candidates.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43097 is the closest representative of the largest apparent theme (v5 behavior/regression), but the cluster is too heterogeneous for a strong canonical issue choice.", + "best_issue_reason": "#43824 is the strongest single issue to anchor this set because it has a precise failure, visible impact, and enough discussion/references to be actionable.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43531", - "right": "issue:45341", + "left": "issue:44704", + "right": "issue:45103", "accept": false, - "reason": "Unrelated problems: sliding_window behavior in Qwen3-MoE vs a small testing_utils bug." + "reason": "Different failures: kwargs propagation into cached_file vs an auto-docstring crash under future annotations." }, { - "left": "issue:44261", - "right": "issue:44485", + "left": "issue:44380", + "right": "issue:46421", "accept": false, - "reason": "Both are model/config bugs, but they concern different models and different code paths (rms_norm_eps precision vs RoPE implementation)." + "reason": "Both touch model/runtime behavior, but one is SDPA/FlashAttention scaling semantics and the other is cache lazy initialization/recompilation." }, { - "left": "issue:43097", - "right": "issue:46050", + "left": "issue:43986", + "right": "issue:44117", "accept": false, - "reason": "Different underlying issues: a removed v5 API flag vs quantized model save/load corruption." + "reason": "One is an AutoProcessor crash without torchvision; the other is TOKENIZER_MAPPING_NAMES returning None. Different code paths." }, { - "left": "issue:43097", - "right": "issue:45588", + "left": "issue:43824", + "right": "issue:43986", "accept": false, - "reason": "One is a v5 API removal, the other is a flash_attention integration crash on sink-less models; not the same fix." + "reason": "Both are import/loading-related, but one is a missing model class export and the other is a video-model AutoProcessor crash tied to torchvision." }, { - "left": "issue:43064", - "right": "issue:44315", + "left": "issue:44938", + "right": "issue:46178", "accept": false, - "reason": "Distinct training bugs: optimizer state mismatch under FSDP2/PEFT vs Liger kernel not being applied with model_init." + "reason": "Different platform/backend issues: Python 3.14 import failure vs an invalid BACKENDS_MAPPING entry for tensorflow_text." }, { - "left": "issue:38617", - "right": "issue:42757", + "left": "issue:43066", + "right": "issue:43906", "accept": false, - "reason": "Both are import errors, but they are different missing symbols from different modules and likely separate regressions." + "reason": "No evidence they share the same bug; 43906 is only a reproduction issue, not the same tokenizer-decoder regression." }, { - "left": "issue:44062", - "right": "issue:44117", + "left": "issue:43208", + "right": "issue:43550", "accept": false, - "reason": "Tokenizer-related, but different failures: AddedToken keyword conflict vs TOKENIZER_MAPPING_NAMES returning None." + "reason": "xLSTM training bugs and Bamba torch.compile/SDPA failure are unrelated model-specific issues." }, { - "left": "issue:44410", - "right": "issue:45375", + "left": "issue:43992", + "right": "issue:44484", "accept": false, - "reason": "Both touch Qwen-family models, but one is a missing projection bug and the other is a missing config field; different fixes." + "reason": "Missing embed_tokens weights during from_pretrained is unrelated to the max_shard_size default in save_pretrained." }, { - "left": "issue:44038", - "right": "issue:44410", + "left": "issue:43012", + "right": "issue:43550", "accept": false, - "reason": "Same broad family, but different model variants and failure modes; not the same concrete bug." + "reason": "A PyTorch precision warning on bfloat16 is not the same as a compile-time SDPA failure." }, { - "left": "issue:43502", - "right": "issue:43992", + "left": "issue:43012", + "right": "issue:44368", "accept": false, - "reason": "Loading/API isolation bug vs missing encoder weight in UMT5; unrelated code paths." + "reason": "A warning about float32 precision is unrelated to tie_word_embeddings/tie_embeddings warnings in Qwen3.5 fine-tuning." }, { - "left": "issue:42890", - "right": "issue:46298", + "left": "issue:43825", + "right": "issue:44655", "accept": false, - "reason": "A flaky integration test due to missing set_seed is unrelated to a noisy logging message." + "reason": "One is a misleading pipeline task message; the other is failure to save Pipeline objects. Different bugs." }, { - "left": "issue:38617", - "right": "issue:44351", + "left": "issue:45712", + "right": "issue:46298", "accept": false, - "reason": "Different missing imports ('layer_type_validation' vs 'HybridCache'); both are import regressions but not the same bug." + "reason": "Left is leftover dummy class cleanup; right is noisy logging. Same area of hygiene, not same defect." }, { - "left": "issue:43097", - "right": "issue:43873", + "left": "issue:42907", + "right": "issue:44655", "accept": false, - "reason": "Unrelated: removal of a v5 config flag vs offloading not working with quantization." + "reason": "Saving dequantized Ministrals/Devstrals and saving Pipeline objects are different save_pretrained failures." }, { - "left": "issue:44704", - "right": "issue:45103", + "left": "issue:42907", + "right": "issue:45753", "accept": false, - "reason": "AutoProcessor kwargs propagation bug and auto_docstring annotation crash are separate tooling issues." + "reason": "A dequantized save failure is unrelated to an infinite loop on a specific image input." }, { - "left": "issue:43986", - "right": "issue:44117", + "left": "issue:43097", + "right": "issue:46298", "accept": false, - "reason": "Different failure modes: missing torchvision dependency crash vs tokenizer mapping assumption bug." + "reason": "Removed tie_embeddings_and_encoder_decoder behavior is a functional API change, not a generic noisy logging issue." }, { - "left": "issue:43824", - "right": "issue:43986", + "left": "issue:43352", + "right": "issue:44380", "accept": false, - "reason": "Both are import/load problems, but one is a missing exported class and the other is a missing optional dependency; separate fixes." + "reason": "Nemotron FlashAttention-2 support and GPT2 attention scaling under SDPA/FlashAttention are different backend/model problems." }, { - "left": "issue:44938", - "right": "issue:46178", + "left": "issue:43352", + "right": "issue:43976", "accept": false, - "reason": "Different compatibility issues: Python 3.14 load failure vs missing BACKENDS_MAPPING entry for tensorflow_text." + "reason": "FlashAttention support for a model is unrelated to Python version compatibility in Transformers 5.1.0." }, { - "left": "issue:43066", - "right": "issue:43906", + "left": "issue:44162", + "right": "issue:46172", "accept": false, - "reason": "Not the same bug: tokenizer decoder type mismatch vs an isolated reproduction of a different issue." + "reason": "ESM2 breakage and a deepseek-v4 tensor expression bug are different model-specific regressions." }, { - "left": "issue:43208", - "right": "issue:43550", + "left": "issue:44380", + "right": "issue:45479", "accept": false, - "reason": "Different models and errors: xLSTM training blockers vs Bamba torch.compile/SDPA failure." + "reason": "Attention scaling under backend dispatch is unrelated to a zero-loss classification bug." }, { - "left": "issue:43992", - "right": "issue:44484", + "left": "issue:44380", + "right": "issue:46132", "accept": false, - "reason": "Model weight loading bug vs a question about the 50GB shard size default; not duplicates." + "reason": "GPT2 scaling under SDPA/FlashAttention is not the same as AttentionInterface.register changing registered-function behavior." }, { - "left": "issue:43012", - "right": "issue:43550", + "left": "issue:43064", + "right": "issue:44380", "accept": false, - "reason": "A precision warning during compile is not the same as a runtime torch.compile/SDPA failure." + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to attention scaling being ignored under backend kernels." }, { - "left": "issue:43825", - "right": "issue:44655", + "left": "issue:44483", + "right": "issue:44514", "accept": false, - "reason": "Both involve pipeline APIs, but one is an incorrect task-support message and the other is pipeline serialization; different issues." + "reason": "Chat/completions request rejection and Qwen2_5_VLProcessor batching crash are different API surfaces and failures." }, { - "left": "issue:43012", - "right": "issue:44368", + "left": "issue:44222", + "right": "issue:45478", "accept": false, - "reason": "Both mention warnings, but the warnings come from different causes and code paths." + "reason": "Both mention Qwen3.5 MoE, but one is FP8 save_pretrained and the other is from_pretrained loading; not the same concrete bug." }, { - "left": "issue:45712", - "right": "issue:46298", + "left": "issue:36683", + "right": "issue:43582", "accept": false, - "reason": "Left is about dummy class cleanup/leakage; right is only about a noisy logging message." + "reason": "Gemma3Config missing vocab_size and Apple Silicon allocator warmup TypeError are unrelated." }, { - "left": "issue:42907", - "right": "issue:44655", + "left": "issue:43208", + "right": "issue:43526", "accept": false, - "reason": "Saving dequantized models and saving Pipeline objects are different serialization problems." + "reason": "xLSTM training blockers and BeitImageProcessorFast reduce_labels returning one label are unrelated." }, { - "left": "issue:42907", - "right": "issue:45753", + "left": "issue:44936", + "right": "issue:46365", "accept": false, - "reason": "Completely unrelated: dequantized save/load failure vs Qwen3_5 infinite-loop behavior on a specific image." + "reason": "trainer.evaluate() after train() failing and inability to import Trainer with datasets installed are different trainer/import issues." }, { - "left": "issue:43352", - "right": "issue:44380", + "left": "issue:42490", + "right": "issue:43012", "accept": false, - "reason": "Both are attention/backend-related, but one is unsupported Flash Attention 2.0 in a model and the other is GPT2 attention scaling being ignored." + "reason": "Trailing-slash local model path import warning is unrelated to a bfloat16 precision warning." }, { - "left": "issue:43097", - "right": "issue:43976", + "left": "issue:43012", + "right": "issue:44683", "accept": false, - "reason": "A v5 API removal issue is unrelated to a Python version compatibility failure." + "reason": "Precision warning on bfloat16 and compiled flex_attention failing on torch>=2.9 are different bugs." }, { - "left": "issue:44162", - "right": "issue:46172", + "left": "issue:42757", + "right": "issue:43824", "accept": false, - "reason": "Different models and defects: ESM2 breakage vs a DeepSeek v4 transpose/dtype bug." + "reason": "Missing is_offline_mode and missing Qwen2_5_VLForConditionalGeneration are unrelated import errors." }, { - "left": "issue:44380", - "right": "issue:45479", + "left": "issue:38617", + "right": "issue:43502", "accept": false, - "reason": "Attention backend scaling bug and zero-loss classification bug are unrelated." + "reason": "layer_type_validation import removal and API requests despite local_files_only=True are different regressions." }, { - "left": "issue:44380", - "right": "issue:46132", + "left": "issue:45600", + "right": "issue:46298", "accept": false, - "reason": "One is about attention scaling under backends; the other is about AttentionInterface.register changing function behavior." + "reason": "Broken auto_mappings CI references and noisy logging message are not the same issue." }, { - "left": "issue:43064", - "right": "issue:44380", + "left": "issue:44038", + "right": "issue:44230", "accept": false, - "reason": "Training-state corruption under FSDP2/PEFT is unrelated to GPT2 attention scaling under SDPA/FlashAttention." + "reason": "Qwen3-VL-Moe bug in Transformers 5.0 and FP8 inference support for Qwen3.5 MoE are different requests/failures." } ] }, @@ -14941,304 +15271,325 @@ "label": "issue", "component_index": 1, "component_count": 19, - "review_unit_index": 48, + "review_unit_index": 49, "review_unit_count": 53, "status": "reviewed", "reason": null, "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:36010", "issue:36683", - "issue:38617", - "issue:42490", + "issue:42175", "issue:42757", "issue:42994", - "issue:43012", + "issue:43010", + "issue:43064", "issue:43066", - "issue:43208", "issue:43299", "issue:43352", "issue:43475", - "issue:43502", - "issue:43526", - "issue:43582", + "issue:43525", "issue:43611", "issue:43618", "issue:43756", "issue:43761", - "issue:43824", "issue:43867", "issue:43901", "issue:43937", "issue:43976", - "issue:44038", + "issue:43992", + "issue:44075", "issue:44077", - "issue:44222", - "issue:44230", + "issue:44242", + "issue:44355", "issue:44380", "issue:44483", "issue:44485", - "issue:44514", + "issue:44492", "issue:44625", "issue:44655", - "issue:44683", - "issue:44936", + "issue:44855", "issue:44938", "issue:45216", "issue:45276", + "issue:45278", "issue:45290", - "issue:45478", - "issue:45600", + "issue:45310", + "issue:45375", + "issue:45584", + "issue:45593", "issue:45698", "issue:46077", + "issue:46082", "issue:46172", + "issue:46202", "issue:46209", - "issue:46298", "issue:46365" ], "soft_pairs": [ - "issue:44222|issue:45478", - "issue:44483|issue:44514", - "issue:36683|issue:43582", - "issue:43208|issue:43526", - "issue:44936|issue:46365", - "issue:42490|issue:43012", - "issue:43012|issue:44683", - "issue:42757|issue:43824", - "issue:38617|issue:43502", - "issue:45600|issue:46298", - "issue:43618|issue:44380", - "issue:44038|issue:44230", "issue:45698|issue:46172", + "issue:43618|issue:44380", "issue:43618|issue:43761", "issue:44483|issue:45290", "issue:43299|issue:43867", "issue:43756|issue:44485", "issue:36683|issue:43475", - "issue:43066|issue:46209", "issue:42994|issue:44655", + "issue:43066|issue:46209", "issue:45276|issue:46077", - "issue:44077|issue:44380", "issue:43901|issue:44655", "issue:36010|issue:42757", - "issue:43352|issue:43937", "issue:43611|issue:44655", + "issue:43352|issue:43937", "issue:36683|issue:44077", "issue:43867|issue:45216", "issue:44625|issue:46077", - "issue:43976|issue:44938" + "issue:43976|issue:44938", + "issue:45584|issue:46365", + "issue:42175|issue:46365", + "issue:43352|issue:45593", + "issue:46077|issue:46082", + "issue:36683|issue:43525", + "issue:43618|issue:45216", + "issue:44077|issue:46365", + "issue:36683|issue:44242", + "issue:44075|issue:46365", + "issue:43992|issue:45310", + "issue:43064|issue:44492", + "issue:43010|issue:45278", + "issue:44355|issue:44855", + "issue:43867|issue:46202", + "issue:43867|issue:45375" ], - "prepared_review_unit_hash": "64078d5cf0c7af92254becec724d0959dbb8fdc7eb858038a24374e75b78df8e", + "prepared_review_unit_hash": "49e00c70d9d71d2393921f79d237f13bd3eeb6af3403af66826955288f8c5c60", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36137, - "estimated_input_tokens": 9035, - "estimated_eval_tokens": 18326 + "soft_pair_count": 33, + "serialized_chars": 37169, + "estimated_input_tokens": 9293, + "estimated_eval_tokens": 18842 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated Transformers bugs/regressions spanning import errors, model-specific loading issues, attention/compile problems, and documentation mismatches. I don\u2019t see any soft pair that clearly refers to the same underlying defect, so all candidate merges should be rejected.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 45698 is the broadest and most active runtime regression here, with the highest discussion/inbound activity around save_pretrained/from_pretrained behavior.", + "summary": "The cluster is mostly heterogeneous: many items are Transformer v5 regressions, but they touch different models, configs, and code paths. I did not find any soft pair that looks like the same underlying bug/change.", + "confidence": 0.29, + "canonical_issue_reason": "Issue 45698 is the broadest and most actively discussed remaining issue, centered on save/load behavior for custom modules after `save_pretrained`/`from_pretrained`. It is the best anchor only for that subset, not for the whole mixed cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45698 is the best representative issue in this set because it is a concrete end-user bug, broadly scoped, and has the strongest evidence of impact and follow-up.", + "best_issue_reason": "45698 is the strongest standalone issue in the set: open, high discussion activity, and a concrete reproducible save/load regression with clear user impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44222", - "right": "issue:45478", + "left": "issue:45698", + "right": "issue:46172", "accept": false, - "reason": "Related to Qwen3.5 MoE save/load, but one is FP8 save_pretrained and the other is a from_pretrained error; not the same concrete bug." + "reason": "Different bugs: custom module reload after saving vs a DeepSeek-V4 tensor transpose/dtype issue." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Both mention attention, but one is missing CLIPOutput attentions and the other ignores GPT-2 attention scaling under SDPA/FlashAttention; different code paths." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Same model family, but separate regressions: CLIPOutput.attentions not assigned vs CLIPVisionModel.hidden_states becoming None." }, { "left": "issue:44483", - "right": "issue:44514", + "right": "issue:45290", "accept": false, - "reason": "Different layers and failures: chat/completions request validation vs apply_chat_template batching crash." + "reason": "API request validation failure vs chat template tokenization crash; not the same bug." }, { - "left": "issue:36683", - "right": "issue:43582", + "left": "issue:43299", + "right": "issue:43867", "accept": false, - "reason": "Gemma3 config attribute bug vs Apple Silicon allocator TypeError; unrelated subsystems." + "reason": "Qwen3VL MoE loading failure and a generic sorted-state_dict load error are different root causes." }, { - "left": "issue:43208", - "right": "issue:43526", + "left": "issue:43756", + "right": "issue:44485", "accept": false, - "reason": "xLSTM training bugs vs BeitImageProcessorFast label reduction bug; no shared code path." + "reason": "Smollm3 RoPE layer-count mismatch and GLM-5 RoPE implementation discussion are not the same concrete defect." }, { - "left": "issue:44936", - "right": "issue:46365", + "left": "issue:36683", + "right": "issue:43475", "accept": false, - "reason": "Trainer lifecycle failure after train vs importing Trainer breaking until datasets is uninstalled; different issues." + "reason": "Different missing attributes on different objects/models: `Gemma3Config.vocab_size` vs `Sam3VisionEncoderOutput.fpn_position_embeddings`." }, { - "left": "issue:42490", - "right": "issue:43012", + "left": "issue:42994", + "right": "issue:44655", "accept": false, - "reason": "Local-path import warning vs PyTorch precision warning when compiling bfloat16; not the same bug." + "reason": "Quantized model saving failure and pipeline `save_pretrained` failure are related only broadly to saving, but not the same bug." }, { - "left": "issue:43012", - "right": "issue:44683", + "left": "issue:43066", + "right": "issue:46209", "accept": false, - "reason": "One is a warning, the other is a compiled flex_attention failure on newer torch; different failure modes." + "reason": "Tokenizer decoder-type regression vs FP8 inference failure on B200; unrelated symptoms and subsystems." }, { - "left": "issue:42757", - "right": "issue:43824", + "left": "issue:45276", + "right": "issue:46077", "accept": false, - "reason": "Both are import errors, but for different symbols and code paths; not a duplicate." + "reason": "Both are Gemma4 issues, but one is resize_token_embeddings propagation and the other is bidirectional attention-mask construction." }, { - "left": "issue:38617", - "right": "issue:43502", + "left": "issue:43901", + "right": "issue:44655", "accept": false, - "reason": "Missing configuration_utils import vs local_files_only still causing network requests; unrelated." + "reason": "Documentation mismatch for `return_all_scores` is not the same as a runtime pipeline serialization bug." }, { - "left": "issue:45600", - "right": "issue:46298", + "left": "issue:36010", + "right": "issue:42757", "accept": false, - "reason": "auto_mappings CI breakage vs noisy logging message; no shared underlying defect." + "reason": "Both are import errors, but for different symbols from different packages (`transformers.generation` vs `huggingface_hub`)." }, { - "left": "issue:43618", - "right": "issue:44380", + "left": "issue:43611", + "right": "issue:44655", "accept": false, - "reason": "CLIP attentions regression vs GPT2 attention scaling ignored under SDPA/FlashAttention; different models and behavior." + "reason": "Different loading/saving problems: `base_model_prefix` handling vs pipeline serialization." }, { - "left": "issue:44038", - "right": "issue:44230", + "left": "issue:43352", + "right": "issue:43937", "accept": false, - "reason": "Both involve Qwen3-VL/MoE, but one is a general v5 bug and the other is FP8 inference support; not the same change." + "reason": "Flash Attention support gating for Nemotron is unrelated to GLM-5 generation-config validation." }, { - "left": "issue:45698", - "right": "issue:46172", + "left": "issue:36683", + "right": "issue:44077", "accept": false, - "reason": "Save/load custom-module resolution bug vs DeepSeek-V4 dtype/transpose issue; unrelated." + "reason": "Missing config attribute vs a restriction on optional `post_init`; not the same failure." }, { - "left": "issue:43618", - "right": "issue:43761", + "left": "issue:43867", + "right": "issue:45216", "accept": false, - "reason": "Both CLIP regressions, but one drops attentions and the other drops hidden_states; insufficient evidence of one underlying bug." + "reason": "Both involve model loading/saving, but one is state_dict ordering and the other is incorrect Qwen3.5 checkpoint contents after `save_pretrained`." }, { - "left": "issue:44483", - "right": "issue:45290", + "left": "issue:44625", + "right": "issue:46077", "accept": false, - "reason": "API request rejection vs chat template tokenization crash with tool calls; different code paths." + "reason": "`num_labels` propagation in Qwen3.5 is unrelated to Gemma4 attention-mask behavior." }, { - "left": "issue:43299", - "right": "issue:43867", + "left": "issue:43976", + "right": "issue:44938", "accept": false, - "reason": "Qwen3VL-Moe loading regression vs generic sorted-state_dict load error; not the same defect." + "reason": "Different Python-version compatibility bugs on different interpreter versions and with different causes." }, { - "left": "issue:43756", - "right": "issue:44485", + "left": "issue:45584", + "right": "issue:46365", "accept": false, - "reason": "Different models and different RoPE-related behaviors; too broad to merge." + "reason": "Whisper generation after special-token alignment is unrelated to `Trainer` import being blocked by `datasets`." }, { - "left": "issue:36683", - "right": "issue:43475", + "left": "issue:42175", + "right": "issue:46365", "accept": false, - "reason": "Gemma3Config vocab_size regression vs SAM3 vision encoder output missing field; unrelated." + "reason": "Dependency/backend installation problem vs import conflict with `datasets`; not the same underlying bug." }, { - "left": "issue:43066", - "right": "issue:46209", + "left": "issue:43352", + "right": "issue:45593", "accept": false, - "reason": "Tokenizer decoder type regression vs FP8 inference on B200; unrelated." + "reason": "Flash Attention unsupported warning and missing auxiliary losses in D-FINE are unrelated." }, { - "left": "issue:42994", - "right": "issue:44655", + "left": "issue:46077", + "right": "issue:46082", "accept": false, - "reason": "Quantized model saving failure vs pipeline save_pretrained failure; both save-related, but not the same concrete bug." + "reason": "Gemma4 attention-mask bug vs LlamaConfig head-dim validation are different model/config problems." }, { - "left": "issue:45276", - "right": "issue:46077", + "left": "issue:36683", + "right": "issue:43525", "accept": false, - "reason": "Different Gemma4 bugs: token embedding resize propagation vs bidirectional attention mask construction." + "reason": "Both are missing config attributes, but they affect different models and different fields; too broad to merge." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions missing and Qwen3.5 saved checkpoint corruption are unrelated." }, { "left": "issue:44077", - "right": "issue:44380", + "right": "issue:46365", "accept": false, - "reason": "Model config validation vs attention scaling backend regression; unrelated." + "reason": "`patchtsmixer` init-policy change is unrelated to a `Trainer` import conflict caused by `datasets`." }, { - "left": "issue:43901", - "right": "issue:44655", + "left": "issue:36683", + "right": "issue:44242", "accept": false, - "reason": "Documentation mismatch for return_all_scores vs pipeline serialization failure; different scope." + "reason": "Missing `vocab_size` and missing load-balancing loss are different bugs in different subsystems." }, { - "left": "issue:36010", - "right": "issue:42757", + "left": "issue:44075", + "right": "issue:46365", "accept": false, - "reason": "Both are import errors, but for different missing symbols from different modules; not duplicates." + "reason": "Optimizer argument handling has nothing to do with `Trainer` import resolution." }, { - "left": "issue:43352", - "right": "issue:43937", + "left": "issue:43992", + "right": "issue:45310", "accept": false, - "reason": "Flash Attention support error for Nemotron vs invalid GenerationConfig for GLM-5; distinct model-specific issues." + "reason": "Different models and different errors: UMT5 missing weights vs Qwen3.5 MoE `from_pretrained` failure." }, { - "left": "issue:43611", - "right": "issue:44655", + "left": "issue:43064", + "right": "issue:44492", "accept": false, - "reason": "base_model_prefix loading regression vs pipeline save_pretrained failure; separate save/load paths." + "reason": "Distributed optimizer-state bug and a typo in cache strategies are unrelated." }, { - "left": "issue:36683", - "right": "issue:44077", + "left": "issue:43010", + "right": "issue:45278", "accept": false, - "reason": "Gemma3 config attribute bug vs patchtsmixer post_init validation; no shared bug." + "reason": "`@torch.no_grad` decoration for cache updates is not the same as broad import failures after upgrade." }, { - "left": "issue:43867", - "right": "issue:45216", + "left": "issue:44355", + "right": "issue:44855", "accept": false, - "reason": "Generic load failure with sorted state_dict vs Qwen3.5 checkpoint saving regression; related theme only, not same bug." + "reason": "Compiled-file import errors and a Python 3.13 decorator/comment parsing regression are different root causes." }, { - "left": "issue:44625", - "right": "issue:46077", + "left": "issue:43867", + "right": "issue:46202", "accept": false, - "reason": "Qwen3.5 num_labels propagation issue vs Gemma4 bidirectional attention mask issue; unrelated." + "reason": "Generic sorted-state_dict load failure vs Gemma pipeline roundtrip on the underscore token are unrelated." }, { - "left": "issue:43976", - "right": "issue:44938", + "left": "issue:43867", + "right": "issue:45375", "accept": false, - "reason": "Python version incompatibility reports with different version constraints and likely different causes; not the same issue." + "reason": "State_dict ordering issue and missing `deepstack_visual_indexes` config field are different bugs." } ] }, @@ -15253,325 +15604,311 @@ "label": "issue", "component_index": 1, "component_count": 19, - "review_unit_index": 49, + "review_unit_index": 50, "review_unit_count": 53, "status": "reviewed", "reason": null, "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:28282", + "issue:30333", "issue:33453", "issue:36296", "issue:36683", "issue:41669", - "issue:42175", + "issue:42548", "issue:42890", "issue:43010", "issue:43064", - "issue:43352", + "issue:43065", + "issue:43232", "issue:43502", - "issue:43525", "issue:43575", + "issue:43576", "issue:43618", + "issue:43723", "issue:43867", - "issue:43931", "issue:43976", "issue:43992", "issue:44062", - "issue:44075", "issue:44077", - "issue:44242", + "issue:44112", + "issue:44222", "issue:44246", + "issue:44351", "issue:44355", "issue:44360", - "issue:44492", + "issue:44387", + "issue:44485", + "issue:44556", "issue:44625", - "issue:44855", + "issue:44749", "issue:44857", "issue:44938", "issue:44964", "issue:45071", - "issue:45216", - "issue:45278", - "issue:45310", + "issue:45125", "issue:45357", "issue:45375", "issue:45440", "issue:45478", "issue:45561", - "issue:45584", - "issue:45593", + "issue:45600", "issue:45850", - "issue:46077", - "issue:46082", - "issue:46202", + "issue:46132", "issue:46285", "issue:46365" ], "soft_pairs": [ - "issue:45584|issue:46365", - "issue:46077|issue:46082", - "issue:42175|issue:46365", - "issue:43352|issue:45593", - "issue:36683|issue:43525", - "issue:43618|issue:45216", - "issue:44077|issue:46365", - "issue:43352|issue:43931", - "issue:36683|issue:44242", - "issue:44075|issue:46365", - "issue:43992|issue:45310", - "issue:43064|issue:44492", - "issue:43010|issue:45278", - "issue:44355|issue:44855", - "issue:43867|issue:46202", - "issue:43867|issue:45375", - "issue:44938|issue:45071", "issue:41669|issue:44246", + "issue:44938|issue:45071", "issue:36683|issue:43064", "issue:42890|issue:45561", "issue:43010|issue:45850", "issue:44625|issue:46365", "issue:43992|issue:45478", "issue:43867|issue:45357", - "issue:36296|issue:43575", "issue:43010|issue:43976", + "issue:36296|issue:43575", "issue:43618|issue:44077", "issue:28282|issue:33453", "issue:43618|issue:46285", "issue:43618|issue:44857", "issue:44360|issue:45440", "issue:43502|issue:44062", - "issue:42890|issue:44964" + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:43723|issue:44556", + "issue:44387|issue:44556", + "issue:44222|issue:45125", + "issue:43010|issue:43065", + "issue:43576|issue:46132", + "issue:43502|issue:44485", + "issue:42548|issue:44351", + "issue:45125|issue:45478", + "issue:44355|issue:44749", + "issue:30333|issue:33453", + "issue:43010|issue:43232", + "issue:45125|issue:45375" ], - "prepared_review_unit_hash": "e4ef1dd5ab49c47ff4c14b12fad688f16fea43b774db3b5668d79ad86f2d2f00", + "prepared_review_unit_hash": "c02bf1e946eac8e1896227e8dae61b7a841a41067e6edafb5af497350b3c06a9", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37193, - "estimated_input_tokens": 9299, - "estimated_eval_tokens": 18854 + "soft_pair_count": 31, + "serialized_chars": 36541, + "estimated_input_tokens": 9136, + "estimated_eval_tokens": 18528 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No pair here looks like the same underlying bug/change. The items mix unrelated import, config, model-loading, training, and test regressions, so this cluster is too heterogeneous for duplicate merging.", - "confidence": 0.91, - "canonical_issue_reason": "issue:45278 is the closest umbrella for the broad import-regression / post-upgrade breakage theme, but the cluster is still too mixed to treat it as a true duplicate anchor.", + "summary": "This cluster is a loose mix of unrelated Transformers v5 regression reports. The soft-edge pairs share broad themes like import/loading, quantization, tensor parallelism, or test failures, but none look like the same concrete bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "issue:42548 is the strongest representative because it describes a broad, user-facing core import/export break in the package, which is more central than the narrower model- or test-specific reports here.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best overall representative because it is the broadest and most central of the import-error/regression reports, though it still does not unify the whole cluster.", - "best_pr_reason": null, - "soft_edge_verdicts": [ - { - "left": "issue:45584", - "right": "issue:46365", - "accept": false, - "reason": "Empty transcription failure vs Trainer import blocked by datasets; unrelated." - }, - { - "left": "issue:46077", - "right": "issue:46082", - "accept": false, - "reason": "Gemma4 attention-mask bug vs LlamaConfig head_dim validation; different code paths." - }, + "best_issue_reason": "issue:42548 is the best single issue to anchor the cluster since it reflects a high-impact API regression and is easy to recognize as a canonical core-package failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ { - "left": "issue:42175", - "right": "issue:46365", + "left": "issue:41669", + "right": "issue:44246", "accept": false, - "reason": "Missing TensorFlow backend in torch install vs Trainer import issue; unrelated." + "reason": "Both concern import-related performance, but one is about `import *` slowing import time and the other about intermittent slow `import transformers`; not the same bug." }, { - "left": "issue:43352", - "right": "issue:45593", + "left": "issue:44938", + "right": "issue:45071", "accept": false, - "reason": "FlashAttention2 support error vs D-FINE auxiliary-loss behavior; different models." + "reason": "Python 3.14 load failure and `PretrainedConfig` type-checking break are different compatibility issues." }, { "left": "issue:36683", - "right": "issue:43525", + "right": "issue:43064", "accept": false, - "reason": "Gemma3 vocab_size missing vs Llama4 pad_token_id missing; different configs/models." + "reason": "A missing `vocab_size` on `Gemma3Config` and a FSDP2/PEFT optimizer-state mismatch are unrelated problems." }, { - "left": "issue:43618", - "right": "issue:45216", + "left": "issue:42890", + "right": "issue:45561", "accept": false, - "reason": "CLIPOutput attentions regression vs Qwen3.5 save_pretrained key corruption; unrelated." + "reason": "Seed-dependent integration flakiness and a pytest-xdist race in testing utils are different test failures." }, { - "left": "issue:44077", - "right": "issue:46365", + "left": "issue:43010", + "right": "issue:45850", "accept": false, - "reason": "patchtsmixer post_init policy vs Trainer import problem; unrelated." + "reason": "Decorating cache/layer `update()` with `torch.no_grad` is a specific autograd fix; \u201cbreaks too many models\u201d is far broader and not the same defect." }, { - "left": "issue:43352", - "right": "issue:43931", + "left": "issue:44625", + "right": "issue:46365", "accept": false, - "reason": "Nemotron FlashAttention2 unsupported vs Qwen3-VL weight-shape mismatch; unrelated." + "reason": "Missing `num_labels` propagation in Qwen3.5 is a model config bug; inability to import `Trainer` until uninstalling `datasets` is an import dependency issue." }, { - "left": "issue:36683", - "right": "issue:44242", + "left": "issue:43992", + "right": "issue:45478", "accept": false, - "reason": "Gemma3 config attr bug vs router-logits load-balancing loss wiring; unrelated." + "reason": "`UMT5Encoder` missing `embed_tokens.weight` and Qwen3.5 MoE `from_pretrained` error are distinct model-loading failures." }, { - "left": "issue:44075", - "right": "issue:46365", + "left": "issue:43867", + "right": "issue:45357", "accept": false, - "reason": "SGD args ignored vs Trainer import issue; unrelated." + "reason": "A load error when state dict keys are sorted and incorrect visual encoder keys saved by Qwen3.5 are not the same bug." }, { - "left": "issue:43992", - "right": "issue:45310", + "left": "issue:43010", + "right": "issue:43976", "accept": false, - "reason": "UMT5Encoder missing embed_tokens.weight vs Qwen3.5 MoE from_pretrained error; different bugs." + "reason": "`torch.no_grad` on cache updates and Python-version support in 5.1.0 are unrelated." }, { - "left": "issue:43064", - "right": "issue:44492", + "left": "issue:36296", + "right": "issue:43575", "accept": false, - "reason": "FSDP2 optimizer-state corruption vs cache-strategy typo; unrelated." + "reason": "Tensor-parallel training bug and OOM when loading a specific Qwen2 model with TP are related at most broadly, but not the same concrete issue." }, { - "left": "issue:43010", - "right": "issue:45278", + "left": "issue:43618", + "right": "issue:44077", "accept": false, - "reason": "Cache update no_grad bug vs broad import-regression report; not the same issue." + "reason": "Missing `CLIPOutput.attentions` assignment and `patchtsmixer` post_init policy are separate model API changes." }, { - "left": "issue:44355", - "right": "issue:44855", + "left": "issue:28282", + "right": "issue:33453", "accept": false, - "reason": "Compiled-file import errors vs Python 3.13 JIT comment parsing; unrelated." + "reason": "PyTorch-missing import error for `AutoModel` and tokenizer-loading regression are different code paths." }, { - "left": "issue:43867", - "right": "issue:46202", + "left": "issue:43618", + "right": "issue:46285", "accept": false, - "reason": "Sorted state_dict load failure vs Gemma roundtrip token issue; different failures." + "reason": "Both mention CLIP-related structures, but one is an output field regression and the other is inconsistent model class structure; not the same bug." }, { - "left": "issue:43867", - "right": "issue:45375", + "left": "issue:43618", + "right": "issue:44857", "accept": false, - "reason": "Sorted state_dict load failure vs missing Qwen3.5 vision config field; unrelated." + "reason": "CLIP attentions regression and `LwDetrImageLoss` AMP crash are unrelated model bugs." }, { - "left": "issue:44938", - "right": "issue:45071", + "left": "issue:44360", + "right": "issue:45440", "accept": false, - "reason": "Python 3.14 load failure vs PretrainedConfig type-check regression; unrelated." + "reason": "A discussion about a DSA indexer activation function and a DeepSeekV3 implementation divergence are different topics." }, { - "left": "issue:41669", - "right": "issue:44246", + "left": "issue:43502", + "right": "issue:44062", "accept": false, - "reason": "Import-* slowdown vs missing load-balancing loss; different problems." + "reason": "`local_files_only=True` still making network requests and `AddedToken()` duplicate `special` kwarg are unrelated tokenization/loading issues." }, { - "left": "issue:36683", - "right": "issue:43064", + "left": "issue:42890", + "right": "issue:44964", "accept": false, - "reason": "Gemma3 config attr bug vs FSDP2 optimizer-state bug; unrelated." + "reason": "Missing `set_seed()` in tests and failure to load a specific multimodal model are unrelated." }, { "left": "issue:42890", - "right": "issue:45561", + "right": "issue:44112", "accept": false, - "reason": "Missing set_seed in tests vs xdist race on captured_info.txt; different test issues." + "reason": "A flaky test due to missing seeds and a CI stale device-override failure are different test issues." }, { - "left": "issue:43010", - "right": "issue:45850", + "left": "issue:42890", + "right": "issue:45600", "accept": false, - "reason": "Cache no_grad bug vs generic 'breaks too many models' complaint; not the same bug." + "reason": "Seed-related integration test instability and auto_mappings references to removed configs are unrelated." }, { - "left": "issue:44625", - "right": "issue:46365", + "left": "issue:43723", + "right": "issue:44556", "accept": false, - "reason": "Qwen3.5 num_labels propagation vs Trainer import issue; unrelated." + "reason": "Tokenizer loading regression and checkpoint reload failure after upgrade are different loading problems." }, { - "left": "issue:43992", - "right": "issue:45478", + "left": "issue:44387", + "right": "issue:44556", "accept": false, - "reason": "UMT5Encoder embedding-weight miss vs Qwen3.5 MoE from_pretrained error; different model-loading bugs." + "reason": "Int4 quantization memory growth/OOM and checkpoint reload failure are not the same bug." }, { - "left": "issue:43867", - "right": "issue:45357", + "left": "issue:44222", + "right": "issue:45125", "accept": false, - "reason": "Sorted state_dict load failure vs wrong visual-encoder keys in save_pretrained; different paths." + "reason": "FP8 `save_pretrained` behavior for MoE and missing `_tp_plan` for tensor parallelism are separate features/bugs." }, { - "left": "issue:36296", - "right": "issue:43575", + "left": "issue:43010", + "right": "issue:43065", "accept": false, - "reason": "Tensor-parallel training bug vs OOM loading a large Qwen2 model; different failure modes." + "reason": "`torch.no_grad` decoration for cache updates and a dummy `nn.Conv2d` in Sam3PixelDecoder are unrelated." }, { - "left": "issue:43010", - "right": "issue:43976", + "left": "issue:43576", + "right": "issue:46132", "accept": false, - "reason": "Cache update no_grad bug vs Python-version compatibility issue; unrelated." + "reason": "A broken `transformers env` command and `AttentionInterface.register` behavior change are distinct issues." }, { - "left": "issue:43618", - "right": "issue:44077", + "left": "issue:43502", + "right": "issue:44485", "accept": false, - "reason": "CLIPOutput attentions regression vs patchtsmixer post_init policy; unrelated." + "reason": "Unexpected network requests with `local_files_only=True` and GLM-5 RoPE implementation are unrelated." }, { - "left": "issue:28282", - "right": "issue:33453", + "left": "issue:42548", + "right": "issue:44351", "accept": false, - "reason": "PyTorch-missing ImportError vs tokenizer loading regression; unrelated." + "reason": "Both are import failures, but one is `PreTrainedModel` export/import and the other is `HybridCache`; different symbols and likely different causes." }, { - "left": "issue:43618", - "right": "issue:46285", + "left": "issue:45125", + "right": "issue:45478", "accept": false, - "reason": "Both are CLIP-related, but one is missing attentions and the other is structural inconsistency; not the same bug." + "reason": "Missing `_tp_plan` and Qwen3.5 MoE `from_pretrained` failure are related only by model family, not the same concrete bug." }, { - "left": "issue:43618", - "right": "issue:44857", + "left": "issue:44355", + "right": "issue:44749", "accept": false, - "reason": "CLIPOutput attentions regression vs LwDetr AMP crash; unrelated." + "reason": "Compiled Python file errors and a slowdown when filtering data after upgrading are not the same issue." }, { - "left": "issue:44360", - "right": "issue:45440", + "left": "issue:30333", + "right": "issue:33453", "accept": false, - "reason": "DSA indexer ReLU discussion vs DeepseekV3MoE divergence; unrelated." + "reason": "MLflow reporting on failed training jobs and tokenizer loading regression are unrelated." }, { - "left": "issue:43502", - "right": "issue:44062", + "left": "issue:43010", + "right": "issue:43232", "accept": false, - "reason": "local_files_only still triggers API calls vs AddedToken duplicate kwarg; unrelated." + "reason": "Cache/layer update autograd decoration and `_update_model_kwargs_for_generation` after `sync_gpus` are different generation-path bugs." }, { - "left": "issue:42890", - "right": "issue:44964", + "left": "issue:45125", + "right": "issue:45375", "accept": false, - "reason": "Missing set_seed in tests vs Phi-4 multimodal load failure; unrelated." + "reason": "Missing tensor-parallel plan and a missing `deepstack_visual_indexes` field dropped by strict config handling are different config/model issues." } ] }, @@ -15586,15 +15923,16 @@ "label": "issue", "component_index": 1, "component_count": 19, - "review_unit_index": 50, + "review_unit_index": 51, "review_unit_count": 53, "status": "reviewed", "reason": null, "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ + "issue:27957", "issue:30333", "issue:30990", "issue:31795", @@ -15606,59 +15944,44 @@ "issue:42548", "issue:42757", "issue:42831", - "issue:42890", - "issue:43010", - "issue:43065", - "issue:43232", "issue:43299", "issue:43502", + "issue:43526", "issue:43576", - "issue:43723", + "issue:43673", "issue:43824", "issue:43827", "issue:43874", "issue:43901", - "issue:44112", + "issue:44038", "issue:44222", "issue:44230", + "issue:44273", "issue:44315", - "issue:44351", "issue:44355", "issue:44360", - "issue:44387", + "issue:44393", "issue:44485", + "issue:44492", "issue:44556", + "issue:44623", "issue:44655", - "issue:44749", "issue:44908", "issue:44938", "issue:44961", "issue:45071", "issue:45125", - "issue:45375", + "issue:45230", + "issue:45362", "issue:45468", - "issue:45478", - "issue:45600", - "issue:46132", + "issue:45542", + "issue:45850", + "issue:46095", "issue:46143", "issue:46144", "issue:46209" ], "soft_pairs": [ - "issue:42890|issue:44112", - "issue:42890|issue:45600", - "issue:43723|issue:44556", - "issue:44387|issue:44556", - "issue:44222|issue:45125", - "issue:43010|issue:43065", - "issue:43576|issue:46132", - "issue:43502|issue:44485", - "issue:42548|issue:44351", - "issue:45125|issue:45478", - "issue:44355|issue:44749", - "issue:30333|issue:33453", - "issue:43010|issue:43232", - "issue:45125|issue:45375", "issue:30990|issue:31795", "issue:43502|issue:44222", "issue:43299|issue:44556", @@ -15670,277 +15993,255 @@ "issue:44355|issue:44655", "issue:44315|issue:44360", "issue:33453|issue:34689", + "issue:27957|issue:33453", "issue:42757|issue:43576", - "issue:33453|issue:36296", "issue:43576|issue:45071", "issue:43576|issue:43824", "issue:31795|issue:38617", "issue:44961|issue:45468", - "issue:43576|issue:44938", - "issue:44222|issue:44230", - "issue:46143|issue:46144", - "issue:41628|issue:42548", - "issue:44485|issue:44961", - "issue:44908|issue:44961", - "issue:36296|issue:43901", - "issue:44315|issue:46209" - ], - "prepared_review_unit_hash": "1f43ab0273228566f69f6329d1735bb2d4e4300df1580a8cde4868c8d58e66f6", - "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 - }, - "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 39, - "serialized_chars": 38764, - "estimated_input_tokens": 9691, - "estimated_eval_tokens": 19638 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": true, - "analyst_result": { - "summary": "No strong duplicate cluster here: the items are mostly unrelated Transformers issues that only share broad themes like v5 regressions, loading/import problems, or FP8/MoE work. I found no soft pair that looks like the same underlying bug/change, so all soft edges should be rejected.", - "confidence": 0.79, - "canonical_issue_reason": "issue:44556 is the most representative broad user-facing regression report in this set, centered on v5 checkpoint/model reload breakage after upgrade.", - "canonical_pr_reason": null, - "best_issue_reason": "issue:44556 best captures the cluster\u2019s dominant pattern of v5 loading regressions and is the broadest representative issue among the candidates.", - "best_pr_reason": null, - "soft_edge_verdicts": [ - { - "left": "issue:42890", - "right": "issue:44112", - "accept": false, - "reason": "Both are CI/test failures, but one is a missing set_seed flake and the other is a stale device-override test; different bugs and code paths." - }, - { - "left": "issue:42890", - "right": "issue:45600", - "accept": false, - "reason": "Test flakiness in Sam HQ vs broken auto_mappings references to removed configs; unrelated symptoms and fixes." - }, - { - "left": "issue:43723", - "right": "issue:44556", - "accept": false, - "reason": "Both involve loading regressions in v5, but one is tokenizer loading and the other is checkpoint/model reload after upgrade; not the same concrete bug." - }, - { - "left": "issue:44387", - "right": "issue:44556", - "accept": false, - "reason": "OOM under int4 quantization is a memory/regression issue, while 44556 is a model reload failure; different failure modes." - }, - { - "left": "issue:44222", - "right": "issue:45125", - "accept": false, - "reason": "FP8 save_pretrained for MoE and missing _tp_plan for tensor parallelism are separate feature gaps in different code paths." - }, - { - "left": "issue:43010", - "right": "issue:43065", - "accept": false, - "reason": "One is about decorating cache/layer update with no_grad, the other about a dummy Conv2d in Sam3PixelDecoder; unrelated." - }, - { - "left": "issue:43576", - "right": "issue:46132", - "accept": false, - "reason": "Broken env CLI and AttentionInterface.register behavior change are different subsystems and different bugs." - }, - { - "left": "issue:43502", - "right": "issue:44485", - "accept": false, - "reason": "local_files_only network leakage and GLM-5 RoPE implementation are unrelated." - }, - { - "left": "issue:42548", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but for different missing symbols from different modules; not the same export regression." - }, - { - "left": "issue:45125", - "right": "issue:45478", - "accept": false, - "reason": "Both touch Qwen3.5 MoE loading, but one is missing tensor-parallel plan metadata and the other is a from_pretrained regression; not one bug." - }, - { - "left": "issue:44355", - "right": "issue:44749", - "accept": false, - "reason": "Compiled Python file errors and a data-processing slowdown are unrelated." - }, - { - "left": "issue:30333", - "right": "issue:33453", - "accept": false, - "reason": "MLFlow reporting on training failure vs tokenizer loading regression; no shared underlying bug." - }, - { - "left": "issue:43010", - "right": "issue:43232", - "accept": false, - "reason": "Cache update no_grad decoration and generation kwargs after sync_gpus are different code paths." - }, - { - "left": "issue:45125", - "right": "issue:45375", - "accept": false, - "reason": "Missing _tp_plan and missing deepstack_visual_indexes are different model/config issues." - }, + "issue:44222|issue:44230", + "issue:43576|issue:44938", + "issue:41628|issue:42548", + "issue:46143|issue:46144", + "issue:44908|issue:44961", + "issue:44485|issue:44961", + "issue:36296|issue:43901", + "issue:44315|issue:46209", + "issue:36296|issue:43526", + "issue:43673|issue:44492", + "issue:45362|issue:45850", + "issue:44623|issue:45230", + "issue:44273|issue:45542", + "issue:44038|issue:45125", + "issue:44230|issue:46095", + "issue:44393|issue:45071" + ], + "prepared_review_unit_hash": "1153933e6143ed68598457792aa529f41851a6def46cf4d1a4cc52fdd092e015", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 36712, + "estimated_input_tokens": 9178, + "estimated_eval_tokens": 18612 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues, with a few broad v5-regression reports mixed in. None of the soft pairs look like true duplicates under a strict same-bug/same-change standard.", + "confidence": 0.58, + "canonical_issue_reason": "issue:45850 is the broadest umbrella in the set: it explicitly summarizes the general v5 breakage theme and can absorb the various model-loading/regression complaints better than the narrower, model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the best global representative for the cluster because it is the only high-level issue that spans the repeated 'v5 breaks X' pattern rather than a single model or subsystem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Sentence Transformers loading hangs and documentation confusion about forward args are not the same bug." + "reason": "Sentence Transformers loading hangs and forward-docs confusion are unrelated problems." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "Unexpected API requests with local_files_only and FP8 save_pretrained MoE are unrelated." + "reason": "local_files_only network leakage and FP8 moe save_pretrained are different code paths." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are v5 model-loading regressions, but they affect different model families and likely different root causes." + "reason": "Both are load regressions, but they affect different models and likely different failure points." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "A network-call bug and an unrelated placeholder issue title; no duplicate evidence." + "reason": "API requests despite local_files_only and a unrelated racoon-named issue are not the same bug." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLFlow reporting failure vs tensor-parallel training bug are unrelated." + "reason": "MLflow reporting on failure and tensor-parallel training are unrelated." }, { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "FP8 accuracy issues and a missing image-processor method AttributeError are different problems." + "reason": "FineGrainedFP8 accuracy and a missing image-patch helper in GLM46V are distinct defects." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-file errors and PretrainedConfig type-checking regressions are unrelated." + "reason": "Compiled Python file errors and PretrainedConfig type checking are different regressions." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing pipeline() after removals is unrelated to the other issue." + "reason": "Pipeline docs cleanup and the racoon issue share no underlying change." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Compiled Python file errors and save_pretrained pipeline save failure are different bugs." + "reason": "Compiled-file handling and Pipeline.save_pretrained are separate failures." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel not applied with model_init and a DSA indexer ReLU discussion are unrelated." + "reason": "Liger Kernel application via model_init and a DSA ReLU discussion are unrelated." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Both are loading regressions, but for different models and different loading failures." + "reason": "Tokenizer-loading regression and Llama 3.2 vision model loading are not the same concrete bug." }, { - "left": "issue:42757", - "right": "issue:43576", + "left": "issue:27957", + "right": "issue:33453", "accept": false, - "reason": "ImportError for is_offline_mode and a broken env command are separate issues." + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to tokenizer loading regression." }, { - "left": "issue:33453", - "right": "issue:36296", + "left": "issue:42757", + "right": "issue:43576", "accept": false, - "reason": "Tokenizer loading regression vs tensor parallel training bug; unrelated." + "reason": "missing is_offline_mode import and a broken env command are different issues." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Broken env CLI and PretrainedConfig type checking are different concerns." + "reason": "The env CLI breakage and PretrainedConfig type checking have no common code-path." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "Env command breakage and Qwen2_5_VL import failure are distinct issues." + "reason": "Broken env command and missing Qwen2.5-VL export are unrelated." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion about forward parameters and an ImportError for layer_type_validation are unrelated." + "reason": "Documentation ambiguity and a missing symbol import are not duplicates." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "No meaningful overlap; these are unrelated issue reports." - }, - { - "left": "issue:44485", - "right": "issue:44961", - "accept": false, - "reason": "GLM-5 RoPE implementation discussion and an unrelated title do not match." + "reason": "Racoon and Gemma-4 audio positional encoding are unrelated." }, { "left": "issue:44222", "right": "issue:44230", "accept": false, - "reason": "Both involve FP8/MoE, but one is save_pretrained support and the other is inference support; not the same concrete fix." + "reason": "FP8 save_pretrained for moe and FP8 inference support are different features/failures." }, { - "left": "issue:46143", - "right": "issue:46144", + "left": "issue:43576", + "right": "issue:44938", "accept": false, - "reason": "Kwargs propagation in RoFormer methods and attention-interface usage in RoFormer attention are related at a high level but not clearly the same bug or mergeable into one PR." + "reason": "env CLI breakage and Python 3.14 load failure are distinct regressions." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Different missing import symbols from different modules; same broad class only." + "reason": "Missing AutoImageProcessor and missing PreTrainedModel exports are similar in shape but not clearly the same underlying bug." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "One is kwargs not being forwarded, the other is the RoFormer attention implementation not using the attention interface." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs handling vs an unrelated issue title; no duplicate evidence." + "reason": "inverse_sqrt scheduler kwargs handling and racoon are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation and racoon have no shared failure mode." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor parallel training bug and pipeline docs mentioning return_all_scores are unrelated." + "reason": "tensor-parallel training and TextClassificationPipeline docs mismatch are unrelated." }, { "left": "issue:44315", "right": "issue:46209", "accept": false, - "reason": "Model_init/Liger Kernel application and FP8 inference on B200 are different bugs in different areas." + "reason": "Liger Kernel not applied with model_init and FP8 inference on B200 are different bugs." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "tensor-parallel training and BeitImageProcessorFast reduce_labels are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "chunked_prefill cache handling and a cache-strategies typo are not the same issue." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A single Qwen3.5-35B chat crash is narrower than the broad v5 breakage umbrella." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "processor.save_pretrained missing files and a generic bug report title are not enough to match as duplicates." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "lazy loading and a TensorFlow backend installation error are different problems." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "generic Qwen3-VL-Moe breakage and missing _tp_plan for tensor parallelism are not the same concrete defect." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support and a DeepSeek weight-conversion question are unrelated." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output issues and PretrainedConfig type checking do not share a code path." } ] }, @@ -15955,43 +16256,43 @@ "label": "issue", "component_index": 1, "component_count": 19, - "review_unit_index": 51, + "review_unit_index": 52, "review_unit_count": 53, "status": "reviewed", "reason": null, "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:31795", "issue:33453", - "issue:36296", + "issue:34689", "issue:41669", "issue:42371", "issue:42890", "issue:43352", "issue:43408", + "issue:43519", "issue:43525", - "issue:43526", "issue:43572", "issue:43576", "issue:43673", + "issue:43704", "issue:43827", "issue:43976", - "issue:44038", "issue:44077", - "issue:44230", "issue:44246", "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", @@ -16000,47 +16301,39 @@ "issue:45003", "issue:45071", "issue:45092", - "issue:45125", "issue:45230", "issue:45278", + "issue:45310", "issue:45341", - "issue:45362", + "issue:45468", "issue:45478", - "issue:45542", "issue:45753", "issue:45850", - "issue:46095", + "issue:46018", + "issue:46172", "issue:46218", "issue:46328" ], "soft_pairs": [ - "issue:36296|issue:43526", - "issue:43673|issue:44492", - "issue:45362|issue:45850", - "issue:44623|issue:45230", - "issue:44273|issue:45542", - "issue:44038|issue:45125", - "issue:44230|issue:46095", - "issue:44393|issue:45071", "issue:44273|issue:45003", "issue:43576|issue:44556", "issue:44393|issue:45753", - "issue:41669|issue:43673", "issue:43525|issue:45850", + "issue:41669|issue:43673", "issue:44556|issue:45092", "issue:43572|issue:45850", "issue:44492|issue:45850", - "issue:42371|issue:44492", "issue:45278|issue:45850", + "issue:42371|issue:44492", "issue:41669|issue:46218", "issue:42890|issue:44908", "issue:43408|issue:44315", - "issue:44246|issue:44273", "issue:44297|issue:45230", + "issue:44246|issue:44273", "issue:43352|issue:44393", "issue:44295|issue:45230", - "issue:44704|issue:44908", "issue:43576|issue:44908", + "issue:44704|issue:44908", "issue:43976|issue:44393", "issue:44908|issue:45071", "issue:41669|issue:44351", @@ -16054,282 +16347,297 @@ "issue:31795|issue:33453", "issue:44315|issue:44829", "issue:43576|issue:44273", - "issue:44655|issue:44861" + "issue:44655|issue:44861", + "issue:45230|issue:45310", + "issue:44360|issue:45468", + "issue:44273|issue:44908", + "issue:43576|issue:46218", + "issue:43519|issue:44485", + "issue:46018|issue:46172", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689" ], - "prepared_review_unit_hash": "dc230821d4727b7d599488f9483925b7d8532ea976057125d87ba0d6c4c1743f", + "prepared_review_unit_hash": "8b6312b4a650ed82ca56e8b52b95e28da0dc0d58394fca4cd0ad22e9ef59e618", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 41, - "serialized_chars": 39433, - "estimated_input_tokens": 9859, - "estimated_eval_tokens": 19974 + "soft_pair_count": 42, + "serialized_chars": 39546, + "estimated_input_tokens": 9887, + "estimated_eval_tokens": 20030 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose grab bag of unrelated Transformers issues, many around v5 regressions, imports, caching, or model-specific bugs. None of the soft pairs look like the same concrete underlying bug or change, so all should be rejected. If one issue must act as an umbrella, 45850 is the broadest, but it is too generic to be a strong canonical duplicate target.", - "confidence": 0.92, - "canonical_issue_reason": "Issue 45850 is the broadest umbrella for the many v5 breakage reports, but it is still too vague to represent a single concrete bug.", + "summary": "This cluster is a mixed set of Transformers v5-era regressions and adjacent docs/perf/model-specific bugs. The items are thematically related by version upgrades, loading/import behavior, and model-specific breakages, but there are no strong duplicate pairs with the evidence provided.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 45278 is the best representative of the cluster because it states a concrete, broad v5 upgrade regression with many import failures, which matches the cluster\u2019s dominant theme better than the narrower or more vague reports.", "canonical_pr_reason": null, - "best_issue_reason": "45850 is the most general cluster anchor, covering broad v5 regression complaints better than the narrower model-specific reports.", + "best_issue_reason": "Issue 45278 is the most suitable global representative: it is concrete, user-facing, and broadly captures the upgrade breakage pattern without being as vague as the more generic complaints.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:36296", - "right": "issue:43526", - "accept": false, - "reason": "Different problems: tensor-parallel training vs BEiT fast processor label handling." - }, - { - "left": "issue:43673", - "right": "issue:44492", - "accept": false, - "reason": "Cache-missing during chunked prefill is unrelated to a typo in cache strategy docs/code." - }, - { - "left": "issue:45362", - "right": "issue:45850", - "accept": false, - "reason": "Specific Qwen3.5 chat crash vs a broad complaint about many models breaking; not the same bug." - }, - { - "left": "issue:44623", - "right": "issue:45230", - "accept": false, - "reason": "Concrete processor save bug vs an unspecified generic bug report." - }, - { - "left": "issue:44273", - "right": "issue:45542", - "accept": false, - "reason": "Lazy loading failure and TensorFlow/tensorboard backend error are different code paths." - }, - { - "left": "issue:44038", - "right": "issue:45125", - "accept": false, - "reason": "Both involve Qwen3 MOE, but one is a general bug and the other is missing tensor-parallel plan support." - }, - { - "left": "issue:44230", - "right": "issue:46095", - "accept": false, - "reason": "FP8 inference support request is unrelated to a DeepSeek weight-conversion question." - }, - { - "left": "issue:44393", - "right": "issue:45071", - "accept": false, - "reason": "Qwen3-VL bbox output errors are unrelated to PretrainedConfig type-checking regressions." - }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading complaints and unsafe sys.modules access are related only at a very broad import level." + "reason": "Both touch lazy-loading/import behavior, but one is about lazy loading not working and the other about unsafe sys.modules access/performance; not the same bug." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Broken `transformers env` command is not the same as checkpoint reload failures after upgrade." + "reason": "One is a broken `transformers env` command, the other is checkpoint reload failure after upgrading; different surfaces and code paths." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Both are Qwen3-VL-related, but one is bbox output failure and the other is an infinite loop on a specific image." + "reason": "Both are Qwen3-VL problems, but one is 2D box hallucination/error and the other is an infinite loop on a specific image; too different to merge." }, { - "left": "issue:41669", - "right": "issue:43673", + "left": "issue:43525", + "right": "issue:45850", "accept": false, - "reason": "Import * performance regression is not the same as missing GenerationMixin cache in chunked prefill." + "reason": "A missing `pad_token_id` in `Llama4Config` is a specific config bug; the other is a broad complaint that v5 breaks too many models." }, { - "left": "issue:43525", - "right": "issue:45850", + "left": "issue:41669", + "right": "issue:43673", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is a specific config bug, not a general v5 breakage umbrella." + "reason": "Removing `import *` for import-time cost is unrelated to missing `GenerationMixin` cache during chunked prefill." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Checkpoint reload incompatibility and remote-code meta-init incompatibility are different failure modes." + "reason": "Both are upgrade incompatibilities, but they affect different models and failure modes; not the same concrete loading bug." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "StableLmConfig pad token field regression is a specific model config issue, not the same as broad v5 breakage." + "reason": "A missing `pad_token_idx` in one config is not the same as a broad statement that many models break in v5." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A cache-strategy typo is not the same underlying problem as a broad v5 regression report." + "reason": "A typo in cache strategies is a narrow docs/code issue, not the same as a general v5 breakage report." }, { - "left": "issue:42371", - "right": "issue:44492", + "left": "issue:45278", + "right": "issue:45850", "accept": false, - "reason": "TF32 API guidance and cache-strategy typo are unrelated." + "reason": "Both are broad regression reports, but they are not the same underlying failure; one is import errors on upgrade, the other is a generic complaint about many models." }, { - "left": "issue:45278", - "right": "issue:45850", + "left": "issue:42371", + "right": "issue:44492", "accept": false, - "reason": "Both are broad upgrade regressions, but one is import errors and the other is generic model breakage; too vague to merge." + "reason": "TF32 API settings and a cache-strategy typo are unrelated changes." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import-time slowdown from import * is not the same as failing to import a specific model class." + "reason": "One is about import-star cleanup and import-time cost; the other is a missing `Qwen3_5ForCausalLM` import error." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Missing set_seed in tests and inverse_sqrt scheduler kwargs are unrelated." + "reason": "Missing `set_seed()` in tests is unrelated to scheduler kwargs being ignored." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Model-type mismatch warning is unrelated to Liger Kernel not being applied with `model_init`." + "reason": "A model-type warning for SAM3 is unrelated to Liger Kernel not being applied during `model_init`." }, { - "left": "issue:44246", - "right": "issue:44273", + "left": "issue:44297", + "right": "issue:45230", "accept": false, - "reason": "Import-time slowdown and lazy loading correctness are different issues." + "reason": "Tokenizer class mismatch on save is a specific serialization bug; the other is an unspecified generic bug report." }, { - "left": "issue:44297", - "right": "issue:45230", + "left": "issue:44246", + "right": "issue:44273", "accept": false, - "reason": "Tokenizer class mismatch on save is a specific serialization bug, not a generic bug report." + "reason": "Import slowness and lazy-loading failure are related at a high level, but not clearly the same defect." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention 2 unsupported for Nemotron is unrelated to Qwen3-VL bbox hallucinations/errors." + "reason": "Flash Attention 2 support gating for Nemotron is unrelated to Qwen3-VL bounding-box errors." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "position_ids buffer access error is a concrete runtime bug, not an unspecified report." + "reason": "Reading a buffer-registered `position_ids` and a generic bug report do not describe the same issue." }, { - "left": "issue:44704", + "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "Missing kwargs to cached_file and scheduler kwargs loss are different APIs and code paths." + "reason": "A broken env command and an inverse_sqrt scheduler kwargs bug are unrelated." }, { - "left": "issue:43576", + "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "`transformers env` breakage is unrelated to scheduler argument handling." + "reason": "`AutoProcessor.from_pretrained` kwargs propagation and scheduler kwargs handling are different code paths." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version compatibility regression is unrelated to Qwen3-VL output errors." + "reason": "Python version support breakage is unrelated to Qwen3-VL 2D box output errors." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwargs bug and PretrainedConfig type-checking breakage are unrelated." + "reason": "Ignored scheduler kwargs and PretrainedConfig type checking are unrelated regressions." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import * slowdown is not the same as a missing top-level HybridCache export." + "reason": "Import-star cleanup/import cost is not the same as failing to import `HybridCache`." }, { "left": "issue:44492", "right": "issue:46328", "accept": false, - "reason": "Cache-strategy typo and SequenceFeatureExtractor.pad performance issue are different defects." + "reason": "A cache-strategy typo and a SequenceFeatureExtractor.pad performance problem are unrelated." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import performance regression is unrelated to a testing_utils bug." + "reason": "Model import cleanup/perf is unrelated to a bug in `testing_utils.py`." }, { "left": "issue:43976", "right": "issue:45850", "accept": false, - "reason": "Python support regression is not the same as generic v5 model breakage." + "reason": "Python compatibility breakage is not the same as the vague claim that too many models break." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5 failures: bbox output errors vs from_pretrained error on MOE." + "reason": "Different Qwen model families and different symptoms; not the same bug." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "A `post_init` policy issue in patchtsmixer is unrelated to `_get_tied_weight_keys` crashing on a list." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python version incompatibility and a missing Qwen3_5 import are different problems." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing `pipeline()` after removals is unrelated to scheduler kwargs." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion about forward args is unrelated to a tokenizer loading regression." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel not being applied and flash_attention_3 causing degenerate training are different implementation issues." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "A broken env command and lazy loading not working are both v5-adjacent, but not the same specific bug." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Saving Pipeline objects and tied-weight key extraction are unrelated failures." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "A generic bug report and a specific Qwen3.5 MoE from_pretrained failure are not the same issue." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer ReLU discussion is unrelated to Gemma-4 audio positional encoding." }, { - "left": "issue:44077", - "right": "issue:44861", + "left": "issue:44273", + "right": "issue:44908", "accept": false, - "reason": "Optional post_init policy and tied-weight-key AttributeError are different model/config internals." + "reason": "Lazy-loading issues and scheduler kwargs handling are unrelated." }, { - "left": "issue:43976", + "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "Python compatibility and missing model import are unrelated." + "reason": "A broken env command is unrelated to a Qwen3_5 import failure." }, { - "left": "issue:43827", - "right": "issue:44908", + "left": "issue:43519", + "right": "issue:44485", "accept": false, - "reason": "Docs still mentioning pipeline() and scheduler kwargs loss are unrelated." + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are different model-specific bugs." }, { - "left": "issue:31795", - "right": "issue:33453", + "left": "issue:46018", + "right": "issue:46172", "accept": false, - "reason": "Documentation confusion for forward args is unrelated to tokenizer loading regression." + "reason": "Both are DeepSeek-V4 related, but one reports gating semantics and the other a tensor transpose/casting issue; not clearly the same bug." }, { - "left": "issue:44315", - "right": "issue:44829", + "left": "issue:43519", + "right": "issue:44908", "accept": false, - "reason": "Liger Kernel application under `model_init` and flash_attention_3 training degeneration are different issues." + "reason": "A multimodal timestamp bug and a scheduler kwargs bug are unrelated." }, { - "left": "issue:43576", - "right": "issue:44273", + "left": "issue:41669", + "right": "issue:43704", "accept": false, - "reason": "Broken env command and lazy loading malfunction are not the same bug." + "reason": "Import-star cleanup/import time and a VRAM leak in dataloader threads are different problems." }, { - "left": "issue:44655", - "right": "issue:44861", + "left": "issue:31795", + "right": "issue:34689", "accept": false, - "reason": "Pipeline save_pretrained failure and tied-weight-key list crash are unrelated." + "reason": "Docs wording around `input_ids`/`past_key_values` is unrelated to a model-loading regression for Llama 3.2 Vision." } ] }, @@ -16344,77 +16652,64 @@ "label": "issue", "component_index": 1, "component_count": 19, - "review_unit_index": 52, + "review_unit_index": 53, "review_unit_count": 53, "status": "reviewed", "reason": null, "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", + "cluster_id": "cluster-10000-450", "nodes": [ "issue:30333", "issue:31515", "issue:31795", "issue:33453", - "issue:34689", "issue:36296", + "issue:37934", "issue:38617", - "issue:41669", "issue:42371", "issue:42913", "issue:42936", "issue:43066", + "issue:43240", "issue:43298", "issue:43352", "issue:43502", "issue:43519", "issue:43525", "issue:43572", - "issue:43576", "issue:43606", "issue:43618", "issue:43668", "issue:43704", "issue:43717", + "issue:43813", "issue:44031", "issue:44230", "issue:44261", "issue:44263", - "issue:44273", "issue:44351", - "issue:44360", "issue:44485", "issue:44492", - "issue:44556", "issue:44623", "issue:44861", "issue:44908", "issue:44998", - "issue:45230", "issue:45290", "issue:45310", + "issue:45341", "issue:45468", "issue:45850", "issue:46018", "issue:46095", "issue:46144", - "issue:46172", "issue:46218" ], "soft_pairs": [ - "issue:45230|issue:45310", - "issue:44360|issue:45468", - "issue:44273|issue:44908", - "issue:43576|issue:46218", - "issue:46018|issue:46172", - "issue:43519|issue:44485", - "issue:43519|issue:44908", - "issue:41669|issue:43704", - "issue:31795|issue:34689", "issue:31515|issue:31795", - "issue:44485|issue:45468", "issue:43704|issue:44485", + "issue:44485|issue:45468", "issue:43298|issue:46095", "issue:45850|issue:46218", "issue:43502|issue:43519", @@ -16429,10 +16724,10 @@ "issue:42913|issue:43066", "issue:44908|issue:44998", "issue:44492|issue:44998", - "issue:43618|issue:44998", "issue:44998|issue:45850", + "issue:43618|issue:44998", "issue:44861|issue:45290", - "issue:44556|issue:44998", + "issue:44998|issue:45341", "issue:43525|issue:44861", "issue:43572|issue:44861", "issue:31515|issue:38617", @@ -16444,400 +16739,297 @@ "issue:43298|issue:44230", "issue:43717|issue:46144", "issue:42936|issue:46018", - "issue:43668|issue:44031" + "issue:43668|issue:44031", + "issue:42936|issue:43240", + "issue:43240|issue:46144", + "issue:42936|issue:44031", + "issue:42936|issue:46144", + "issue:42936|issue:43813", + "issue:43240|issue:43668", + "issue:43668|issue:43813", + "issue:37934|issue:43668", + "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "6bcd7392b6ab2a1f94f209e15c79d8c8a918a94b731d86000e9bc5d4eb668e53", + "prepared_review_unit_hash": "62348282138d5ec4e8049bcb03c84c756f1e0ebd35842297ba63c46edc6c2764", "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 44, + "item_count": 44, "soft_pair_count": 42, - "serialized_chars": 39344, - "estimated_input_tokens": 9836, - "estimated_eval_tokens": 19928 + "serialized_chars": 37123, + "estimated_input_tokens": 9281, + "estimated_eval_tokens": 18818 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: mostly unrelated bug reports spanning model loading, tokenizers, processors, training, and tooling. None of the soft pairs look like the same concrete underlying issue, so I would not merge any of them as duplicates.", - "confidence": 0.86, - "canonical_issue_reason": "No clear canonical duplicate target exists because the items are mostly unrelated and describe different failure modes across different subsystems.", + "summary": "Loose cluster of mostly unrelated Transformers bug reports, with a weak theme around model/tokenizer loading regressions. None of the soft pairs are true duplicates under a strict triage standard.", + "confidence": 0.41, + "canonical_issue_reason": "issue:33453 is the clearest concrete regression report in the tokenizer-loading family and is more representative than the broader or more vague items.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45310 is the most concrete and actionable standalone bug report here: it names a specific version regression and a specific model loading failure path.", + "best_issue_reason": "issue:33453 is the strongest standalone issue to anchor this cluster: it is specific, user-facing, and closest to the small tokenizer-regression subtheme here.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45230", - "right": "issue:45310", - "accept": false, - "reason": "Both are bug reports, but one is generic and the other is a specific Qwen3.5 MoE from_pretrained regression; not the same bug." - }, - { - "left": "issue:44360", - "right": "issue:45468", - "accept": false, - "reason": "Different model areas and different symptoms; no shared code-path evidence." - }, - { - "left": "issue:44273", - "right": "issue:44908", - "accept": false, - "reason": "Lazy loading vs scheduler kwargs are unrelated subsystems." - }, - { - "left": "issue:43576", - "right": "issue:46218", - "accept": false, - "reason": "CLI/env command breakage and a model import failure are different issues." - }, - { - "left": "issue:46018", - "right": "issue:46172", + "left": "issue:31515", + "right": "issue:31795", "accept": false, - "reason": "DeepSeek gating logic and a tensor transpose/dtype bug are not the same change." + "reason": "Different bugs: slow checkpoint loading vs documentation confusion about input_ids/past_key_values." }, { - "left": "issue:43519", + "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "Qwen3VL timestamp math and GLM-5 RoPE implementation are unrelated." - }, - { - "left": "issue:43519", - "right": "issue:44908", - "accept": false, - "reason": "Processor timestamp calculation and scheduler kwargs handling are different code paths." - }, - { - "left": "issue:41669", - "right": "issue:43704", - "accept": false, - "reason": "Import-star cleanup and a VRAM leak in dataloader threads are unrelated." - }, - { - "left": "issue:31795", - "right": "issue:34689", - "accept": false, - "reason": "Documentation confusion and a tokenizer/model loading regression are different problems." - }, - { - "left": "issue:31515", - "right": "issue:31795", - "accept": false, - "reason": "Checkpoint load performance and forward-doc confusion are not the same issue." + "reason": "Different code paths: VRAM leak in dataloader-thread usage vs GLM-5 RoPE implementation question." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Distinct models and distinct implementation bugs." - }, - { - "left": "issue:43704", - "right": "issue:44485", - "accept": false, - "reason": "VRAM leak in multi-threaded use and RoPE implementation are unrelated." + "reason": "Both concern model internals, but the specific bugs are unrelated: RoPE math vs Gemma-4 audio positional encoding." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Audio continuous batching support and DeepSeek-V4 weight conversion are different requests." + "reason": "Feature/support request mismatch: continuous batching for audio models vs DeepSeek-V4 weight conversion tooling." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "Broad v5 breakage report and a specific Qwen3.5 import error are not the same concrete bug." + "reason": "Too broad vs specific import failure; not the same underlying bug." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "Local-files-only network requests and timestamp calculation are unrelated." + "reason": "Local-files-only network leakage vs incorrect timestamp calculation in a processor." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 settings guidance and a VRAM leak have no shared code-path." + "reason": "Unrelated areas: TF32 API behavior vs VRAM leak in dataloader-threaded inference." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Processor timestamp math and missing save_pretrained files are different processor bugs." + "reason": "Different processor bugs: timestamp calculation vs missing files from save_pretrained." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Timestamp computation and memory leakage are unrelated." + "reason": "Different failures in different code paths: processor timestamp math vs multi-thread VRAM leak." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch and audio positional encoding are different bugs." + "reason": "CPU-offload device mismatch is not the same bug as Gemma-4 audio positional encoding." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split indexing behavior and audio positional encoding do not match." + "reason": "Different model logic: torch.split return handling vs Gemma-4 audio positional encoding." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "MLA layernorm epsilon precision and audio positional encoding are unrelated." + "reason": "RMS norm epsilon precision issue is unrelated to Gemma-4 audio positional encoding." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "Both are import errors, but for different symbols/modules and different root causes." + "reason": "Importing HybridCache and importing Qwen3_5ForCausalLM are separate import-path failures." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "FA2 unsupported for NemotronH and a missing Qwen3_5 module are different failures." + "reason": "Flash-Attention support gating for Nemotron is unrelated to a Qwen3_5 module import error." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both concern tokenizer v5 behavior, but one is a general behavior difference and the other is a decoder type issue; not enough to call them the same bug." + "reason": "Related tokenizer-v5 theme, but one is a behavior regression and the other is a wrong decoder type; not clearly the same bug." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs bug and an unrelated placeholder issue title are not duplicates." + "reason": "Inverse-sqrt scheduler kwargs bug has nothing to do with the unrelated unemployment issue." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Cache-strategy typo and unrelated placeholder issue are not the same." + "reason": "Typo in cache strategies is unrelated to the unemployment placeholder issue." }, { - "left": "issue:43618", - "right": "issue:44998", + "left": "issue:44998", + "right": "issue:45850", "accept": false, - "reason": "CLIPOutput attentions regression and unrelated placeholder issue are not duplicates." + "reason": "Unemployment placeholder vs broad model-breakage complaint; no shared code path." }, { - "left": "issue:44998", - "right": "issue:45850", + "left": "issue:43618", + "right": "issue:44998", "accept": false, - "reason": "Placeholder issue title and broad v5 breakage report are not the same concrete issue." + "reason": "CLIPOutput attentions regression is unrelated to the unemployment placeholder issue." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weight-key AttributeError and chat-template tool-call crash are different code paths." + "reason": "Tied-weights utility crash vs chat-template crash with tool calls; different functions and symptoms." }, { - "left": "issue:44556", - "right": "issue:44998", + "left": "issue:44998", + "right": "issue:45341", "accept": false, - "reason": "Checkpoint reload regression and unrelated placeholder issue are not duplicates." + "reason": "Unemployment placeholder vs a testing_utils bug; unrelated." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id and tied-weight-key list handling are unrelated model/config bugs." + "reason": "Missing pad_token_id in Llama4Config is unrelated to _get_tied_weight_keys crashing on list input." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_idx and tied-weight-key AttributeError are different problems." + "reason": "StableLmConfig pad_token_idx regression is unrelated to tied-weight key handling." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow checkpoint loading and an import error from configuration_utils are not the same bug." + "reason": "Checkpoint loading slowdown and missing layer_type_validation import are unrelated issues." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 inference support for Qwen3-VL/Qwen3.5 MoE and missing processor save files are unrelated." + "reason": "FP8 inference support request is unrelated to processor.save_pretrained missing files." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLFlow reporting on failed training jobs and slow from_pretrained checkpoint loading are different issues." + "reason": "MLFlow status reporting on failed training jobs is unrelated to checkpoint loading speed." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading speed and tensor-parallel training behavior are unrelated." + "reason": "Checkpoint loading slowdown vs tensor-parallel training bug; different failure modes." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "Both mention Qwen3.5/MoE, but one is FP8 inference support and the other is a from_pretrained regression; different concrete problems." + "reason": "FP8 support request is not the same as the Qwen3.5 MoE from_pretrained loading error." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Both mention loading, but one is a performance complaint about checkpoints and the other is a tokenizer-loading regression." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models and FP8 inference support are distinct feature requests." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba init_weights behavior and RoFormer attention interface usage are unrelated." + "reason": "init_weights behavior in Mamba models is unrelated to RoFormer attention-interface usage." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2Former ignore_value handling and DeepSeek shared-expert gating are different model bugs." + "reason": "Mask2former ignore_value handling and DeepSeek-V4 shared-expert gating are different model-specific bugs." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "Incorrect type hint and regex warning spam are unrelated." - } - ] - }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, - "error_kind": null, - "error_message": null - }, - { - "label": "issue", - "component_index": 1, - "component_count": 19, - "review_unit_index": 53, - "review_unit_count": 53, - "status": "reviewed", - "reason": null, - "source": "llm", - "cache_hit": false, - "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-447", - "nodes": [ - "issue:37934", - "issue:42936", - "issue:43240", - "issue:43668", - "issue:43813", - "issue:44031", - "issue:46144" - ], - "soft_pairs": [ - "issue:42936|issue:43240", - "issue:43240|issue:46144", - "issue:42936|issue:44031", - "issue:42936|issue:46144", - "issue:42936|issue:43813", - "issue:43240|issue:43668", - "issue:43668|issue:43813", - "issue:37934|issue:43668", - "issue:42936|issue:43668" - ], - "prepared_review_unit_hash": "a28ca2ee7048e446ccf98df4895af43e0459a4cd4418401a9b5c1e289977df85", - "component_budget": { - "node_count": 447, - "item_count": 447, - "soft_pair_count": 1652, - "serialized_chars": 855326, - "estimated_input_tokens": 213832, - "estimated_eval_tokens": 427920 - }, - "budget": { - "node_count": 7, - "item_count": 7, - "soft_pair_count": 9, - "serialized_chars": 6638, - "estimated_input_tokens": 1660, - "estimated_eval_tokens": 3576 - }, - "overflow_policy": "truncate_then_skip", - "trimmed": false, - "aggressively_trimmed": false, - "split": true, - "analyst_result": { - "summary": "This cluster looks like a false positive: the issues span unrelated models and bug types (norm layer question, missing ignore_value usage, kwargs plumbing, type hint, typo, tokenizer warning, and RoFormer attention interface). None of the soft pairs appear to describe the same underlying bug or fix.", - "confidence": 0.97, - "canonical_issue_reason": null, - "canonical_pr_reason": null, - "best_issue_reason": null, - "best_pr_reason": null, - "soft_edge_verdicts": [ + "reason": "Incorrect type hint for norm_eps is unrelated to tokenizer regex warning noise." + }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Different components and failure modes: Mask2Former ignore_value handling vs loss-function kwargs propagation." + "reason": "Mask2former ignore_value usage and kwargs not passed to loss calculation are separate issues." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "General kwargs plumbing bug is unrelated to RoFormer attention not using the attention interface." + "reason": "Loss-function kwargs plumbing is unrelated to RoFormer attention-interface implementation." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Mask2Former loss argument usage is unrelated to tokenizer regex warnings." + "reason": "Different problem areas: segmentation ignore_value vs tokenizer regex warnings." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "Different models and code paths; no shared underlying bug." + "reason": "Different model bugs with no shared concrete code path." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2Former loss handling is unrelated to a typo in a quantization attribute name." + "reason": "Mask2former ignore_value bug is unrelated to a quantization-operation typo." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Passing kwargs to a loss function is unrelated to an incorrect type hint in ModernBERTConfig." + "reason": "Loss kwargs propagation and a config type-hint typo are unrelated." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "A type-hint issue and a typo issue are distinct, with no evidence they affect the same code path." + "reason": "Norm_eps type hint and quantization-operation typo are different, isolated fixes." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Both mention norm-related terms, but one is about Llama4TextL2Norm behavior and the other about a ModernBERTConfig type hint." + "reason": "Llama4TextL2Norm/RMS-norm question is unrelated to ModernBERTConfig type-hint correctness." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Different subsystems and bug classes; no plausible duplicate relationship." + "reason": "Ignore_value handling and norm_eps typing are unrelated model/config bugs." } ] }, @@ -16870,12 +17062,12 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "0cb08786b6f7da8efbfbe946b0f81df138208477e71b5fa455c398d8c1fccc66", + "prepared_review_unit_hash": "ff007f9e3a3f74e44d0abc898d94eb227657988274d7e73c30b9c000c749dd0e", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2878, + "serialized_chars": 2877, "estimated_input_tokens": 720, "estimated_eval_tokens": 1696 }, @@ -16883,7 +17075,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2878, + "serialized_chars": 2877, "estimated_input_tokens": 720, "estimated_eval_tokens": 1696 }, @@ -16892,30 +17084,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three issues are not duplicates; they cover unrelated topics: tokenizer truncation behavior, ReduceLROnPlateau-like scheduler functionality, and Wav2Vec2 word timestamp extraction.", - "confidence": 0.96, - "canonical_issue_reason": "#10519 is the most concrete and self-contained request in the set, centered on a specific truncation behavior change; the other two are unrelated topics.", + "summary": "This cluster does not appear to contain duplicates: the three issues are about unrelated topics (truncation behavior, ReduceLROnPlateau-like scheduling, and Wav2Vec2 word time offsets).", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: the items describe different feature requests in different parts of the library, so none is a true representative of the others.", "canonical_pr_reason": null, - "best_issue_reason": "#10519 is the best standalone representative because it describes a clear, actionable feature request with a specific code-path implication.", + "best_issue_reason": "If one must be selected, issue 11307 is the most discussion-heavy, but it is still not a good cluster representative because it is unrelated to the other two.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Only a superficial overlap on 'beginning/end'; one is about tokenizer truncation direction, the other about audio word time offsets in Wav2Vec2." + "reason": "Different requests: token truncation direction vs Wav2Vec2 word boundary offsets; no same underlying bug/change." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "Different functionality and subsystem entirely: truncation options versus ReduceLROnPlateau-like scheduling." + "reason": "Completely different features: truncation option vs ReduceLROnPlateau-like functionality." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "No shared bug or change; scheduler behavior is unrelated to Wav2Vec2 timestamp extraction." + "reason": "Unrelated topics: learning-rate scheduling behavior vs audio token/word offset extraction." } ] }, @@ -16947,12 +17139,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "2b09c12164fee42b55afda30126f7f0250d339e9ce26ce3fa927909599d2ba12", + "prepared_review_unit_hash": "a7f570ce78e4ef1966634cd978ab3fcd91a368937819f00523bd91b16cc1eec0", "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 }, @@ -16960,7 +17152,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2478, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -16969,24 +17161,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues are not duplicates. They cover three different problems: Tapas tokenization mismatch, private-model download/auth error messaging, and a model-loading failure for bert-base-uncased. The shared wording is only superficial.", - "confidence": 0.96, - "canonical_issue_reason": "None of the issues forms a clear duplicate hub for the others; if forced to pick one, #16618 is the broadest/generalized model-loading report, but it is not actually the same bug as the others.", + "summary": "The cluster is not a true duplicate set: one issue is about tokenization behavior, one about private-model auth/download errors, and one about a general model-loading failure. The closest canonical issue is the model-loading report, but the others are distinct bugs.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 16618 ('Can't load the model for 'bert-base-uncased'.') is the broadest and most central model-loading failure in the set, so it is the best anchor for the cluster even though the other issues are not true duplicates of it.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the most general user-facing failure report in the set, but it still describes a different underlying issue than the others.", + "best_issue_reason": "Issue 16618 is the best representative because it concerns a concrete loading failure and is the most likely duplicate target among these reports; the others are narrower and clearly about different code paths.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tapas tokenization differences versus TensorFlow code are unrelated to a bert-base-uncased loading failure; different code paths and bug types." + "reason": "Tokenization mismatch vs. model loading failure are different underlying bugs and different code paths." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Private-model download/auth error messaging is a different problem from failing to load bert-base-uncased; same broad area, but not the same bug." + "reason": "Private-model download/auth error is not the same as failing to load bert-base-uncased; they may both involve loading, but the failure mode and fix are different." } ] }, @@ -17016,12 +17208,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "18b96decdc1cf4e5a0a10ce0e7c2dd0cff1fff88ae5c6cdb730df64c45a8483d", + "prepared_review_unit_hash": "5a4d3ca3ecf13c79859842add75908b045e1024341ef657119dffc2fd50219f9", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -17029,7 +17221,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -17038,7 +17230,7 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate community feature-request issues about different capabilities (Flash Attention 2 vs GGUF). They share broad wording about supporting more architectures, but they do not describe the same underlying bug or change.", + "summary": "The two issues are both broad community-contribution requests, but they target different features: Flash Attention 2 support versus GGUF support. They do not describe the same underlying bug or change, so they should not be clustered as duplicates.", "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, @@ -17049,7 +17241,7 @@ "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Both are broad community contribution requests, but one is for Flash Attention 2 support and the other is for GGUF support; they concern different features and code paths, so they are not duplicates." + "reason": "Different feature requests in different code paths (Flash Attention 2 vs GGUF). The shared wording about adding support for more architectures is too generic to indicate the same underlying change." } ] }, @@ -17079,12 +17271,12 @@ "soft_pairs": [ "issue:26413|issue:31501" ], - "prepared_review_unit_hash": "b2e3fa47d5e8f351944d829177584b39ddb03cda5a5f0b199b0473f5735bbd50", + "prepared_review_unit_hash": "7f63e11b860fea25c0fe90c33eeafbc80aaeb7fd31a1d2c4f8c69a9f291e694a", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1593, + "serialized_chars": 1594, "estimated_input_tokens": 399, "estimated_eval_tokens": 1054 }, @@ -17092,7 +17284,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1593, + "serialized_chars": 1594, "estimated_input_tokens": 399, "estimated_eval_tokens": 1054 }, @@ -17101,18 +17293,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are only loosely similar. One reports a specific `resume_from_checkpoint` failure with a concrete exception, while the other is a broad \"training doesn't start\" symptom for FineWeb SLM training. There isn\u2019t enough evidence they describe the same underlying bug.", - "confidence": 0.82, - "canonical_issue_reason": "issue:26413 is the more canonical issue because it names a precise failure mode and error message, making it a better anchor for the underlying bug than the broader training-startup complaint in issue:31501.", + "summary": "These two issues both concern training flow failures, but they describe different symptoms and code paths: one is a checkpoint-resume iterator crash, the other is a training job that never starts for FineWeb SLM. I would not treat them as duplicates.", + "confidence": 0.48, + "canonical_issue_reason": "Issue 26413 is the more specific and actionable report, centered on a concrete `resume_from_checkpoint` failure and a clear error message.", "canonical_pr_reason": null, - "best_issue_reason": "issue:26413 is the best representative of the cluster because it is the most specific, actionable description of the problem; issue:31501 is too generic to serve as a reliable dedupe target.", + "best_issue_reason": "Issue 26413 is the best representative because it names a precise failure mode and likely narrower root cause than the broad 'training doesn't start' report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26413", "right": "issue:31501", "accept": false, - "reason": "Both are training-related, but one is a specific checkpoint-resume exception and the other is a vague startup failure. They do not clearly look like the same underlying bug." + "reason": "The reports do not describe the same underlying bug: one is a resume/checkpoint epoch-iterator crash, the other is a broader training-start failure on FineWeb. Shared training context alone is not enough to merge them." } ] }, @@ -17171,172 +17363,184 @@ "issue:44370|issue:46226", "issue:46222|issue:46226" ], - "prepared_review_unit_hash": "a76745bcf716dddc497d26d5262b2249cf7194675bacb171484da5e8e78d88c1", + "prepared_review_unit_hash": "5dda4b39e45831de65a7d22e4605e8c18380a20be289630bf13240754dd3bdb1", "component_budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 24, - "serialized_chars": 13527, - "estimated_input_tokens": 3382, - "estimated_eval_tokens": 7020 + "serialized_chars": 13524, + "estimated_input_tokens": 3381, + "estimated_eval_tokens": 7018 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 24, - "serialized_chars": 13527, - "estimated_input_tokens": 3382, - "estimated_eval_tokens": 7020 + "serialized_chars": 13524, + "estimated_input_tokens": 3381, + "estimated_eval_tokens": 7018 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of separate i18n translation requests for different languages. The only strong duplicate-looking subsets are the identical placeholder-title issues ([i18n-]...) and the identical Thai-titled pair; the language-specific issues should stay separate.", - "confidence": 0.74, - "canonical_issue_reason": "If a representative is needed, issue:43834 is the best canonical for the identical placeholder-title subgroup because it is the earliest of the near-exact matches; it does not subsume the language-specific translation requests.", + "summary": "This cluster is mostly a set of distinct i18n translation-request issues for different languages, so they should not be merged across languages. The only true duplicates are the near-identical generic placeholder issues ([i18n-]...) and the identical Thai pair.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue for the whole cluster: most items are separate language-specific translation requests. The generic placeholder issues form a smaller duplicate subgroup only.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43834 is the best overall issue representative because it matches the identical placeholder-title subgroup and is the clearest canonical anchor among the issue-like duplicates.", + "best_issue_reason": "43834 is the best representative of the generic placeholder duplicate subgroup because it is one of the earliest near-identical template issues, but it is not a global canonical for the full cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different translation request context; the similarity is only from the shared i18n template." + "reason": "Different target languages; same translation-request template, but not the same underlying issue." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Turkish vs generic placeholder template; not the same underlying request." + "reason": "Different target languages; not the same underlying translation request." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Turkish translation issue vs a generic placeholder issue; too broad to merge." + "reason": "Turkish request vs generic placeholder/template issue; not the same concrete issue." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Different languages (Turkish vs Romanian), so these are separate translation issues." + "reason": "Different target languages; these are separate translation requests." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different languages (Turkish vs Spanish); same theme, different issue." + "reason": "Turkish vs Spanish translation request; distinct issues." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish translation request vs generic placeholder template; not the same concrete issue." + "reason": "Spanish request vs generic placeholder/template issue; not the same underlying issue." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Spanish translation issue and a generic placeholder issue are not the same request." + "reason": "Different target languages; separate translation requests." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Spanish vs Romanian translation requests are distinct." + "reason": "Spanish vs Romanian translation request; distinct issues." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Spanish translation request vs generic placeholder template; no direct duplicate evidence." + "reason": "Different target languages; not duplicates." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian translation request vs generic placeholder issue; different underlying targets." + "reason": "Romanian request vs generic placeholder/template issue; not the same issue." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Romanian translation request vs generic placeholder issue; not a same-target duplicate." + "reason": "Romanian request vs generic placeholder/template issue; not duplicates." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian translation issue and a generic placeholder issue are not the same underlying request." + "reason": "Romanian request vs generic placeholder/template issue; distinct translation tasks." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Different issue content; the Romanian translation request is unrelated to the Thai-titled item." + "reason": "Different target languages/content; not the same underlying request." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Different issue content; the Romanian translation request is unrelated to the Thai-titled item." + "reason": "Different target languages/content; not duplicates." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "Same generic placeholder title/body pattern; these look like the same issue repeated." + "reason": "Near-identical generic placeholder/template issue; same underlying translation-request stub." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Same placeholder-title translation issue; no meaningful difference in the visible content." + "reason": "Near-identical generic placeholder/template issue; same underlying translation-request stub." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Generic placeholder/template issue vs different Thai-language issue; not the same underlying task." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Generic placeholder/template issue vs different Thai-language issue; not duplicates." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "Near-identical placeholder-title issues with the same underlying template text." + "reason": "Near-identical generic placeholder/template issue; same underlying translation-request stub." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Placeholder translation template vs a different Thai-titled issue; not clearly the same request." + "reason": "Generic placeholder/template issue vs different Thai-language issue; distinct issues." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Placeholder translation template vs a different Thai-titled issue; not clearly the same request." + "reason": "Generic placeholder/template issue vs different Thai-language issue; not duplicates." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Different visible issue text; similarity is generic, not evidence of the same request." + "reason": "Generic placeholder/template issue vs different Thai-language issue; not the same underlying request." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Different visible issue text; similarity is generic, not evidence of the same request." + "reason": "Generic placeholder/template issue vs different Thai-language issue; not duplicates." }, { "left": "issue:46222", "right": "issue:46226", "accept": true, - "reason": "These two are effectively identical in title and length, so they look like the same issue." + "reason": "Identical Thai text and same body shape; true duplicate issue." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays within the packet evidence: it distinguishes the exact-title placeholder subgroup and the exact Thai-titled pair, and it avoids overclaiming duplicates for the language-tagged issues. The soft-edge rejections are conservative given the available titles/body sizes and only soft-similarity evidence." + "feedback": "The summary is conservative and grounded in the packet. It correctly treats the language-specific translation requests as distinct issues, and it only groups the clearly near-identical generic placeholder issues plus the identical Thai pair as duplicates. The canonical-issue conclusion is appropriately cautious, and the soft-edge false/true verdicts do not overstate evidence." }, "evaluator_used": true, "retried": false, @@ -17412,7 +17616,7 @@ "issue:43837|issue:45177", "issue:45322|issue:46133" ], - "prepared_review_unit_hash": "8b3842ffed9ab7b4f0fec520791a061805a90878142cbb1875329640ee2584c0", + "prepared_review_unit_hash": "cab8c2e546cd855918ab27f1f2a13c38f758e61ca5a26d9ce99d569fcc5bad97", "component_budget": { "node_count": 16, "item_count": 16, @@ -17434,222 +17638,228 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are broadly similar model-request issues, but they are not duplicates: each asks to add a different model or modality (text, audio, ASR/TTS, OCR, vision-language, detector, sparse transformer).", - "confidence": 0.96, - "canonical_issue_reason": "No canonical issue: the cluster spans distinct model additions and not one underlying bug/change.", + "summary": "This cluster is mostly a set of distinct model-request issues with similar wording, not true duplicates. The overlap is broad template similarity around \"add/support model X\" rather than the same underlying bug or change.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue: the items request different models, architectures, and modalities, so they are not duplicates of one underlying ticket.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a strong global representative for duplicate triage because the requests target different models and code paths.", + "best_issue_reason": "Issue 43837 is a clear, moderately discussed support request and a reasonable representative of the cluster, but it is still not a duplicate target for the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "Different model families (EuroBERT vs DEIMv2); only the generic 'add model' intent overlaps." + "reason": "Different models (EuroBert vs DEIMv2); only generic request-template similarity." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBERT text-model support request vs Qwen3-TTS audio support; unrelated code paths." + "reason": "Different model families and modalities; not the same change." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBERT config addition and RF-DETR model addition are separate model integrations." + "reason": "EuroBert vs RF-DETR are unrelated model requests." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR detector vs DEIMv2; same broad subsystem, but not the same model or change." + "reason": "Different models and likely different integration paths." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "Object detection model request vs LFM2.5 Audio; different modality and implementation." + "reason": "RF-DETR and LFM2.5 Audio are unrelated requests." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "RF-DETR integration is not the same as adding a sparse transformer architecture." + "reason": "Object detection model vs sparse transformer request; not same underlying task." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "Detector model request vs Qwen3-ASR support; different target models and paths." + "reason": "Different model requests (RF-DETR vs Qwen3-ASR)." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR and Qwen3-TTS are unrelated additions despite similar request phrasing." + "reason": "RF-DETR and Qwen3-TTS are different modalities and code paths." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT text-model addition and Qwen3-ASR audio support are different changes." + "reason": "NeoBERT and Qwen3-ASR are unrelated model additions." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT integration is not the same concrete change as sparse-transformer support." + "reason": "Different model families; only broad 'add model' similarity." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT and LFM2.5 Audio are unrelated model requests." + "reason": "NeoBERT vs LFM2.5 Audio are separate requests." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are separate model additions with different implementations." + "reason": "RF-DETR and NeoBERT are distinct models with no shared concrete fix." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are different model families; similar naming is not enough." + "reason": "NeoBERT and ModernVBERT are different model integrations." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture request vs Qwen3-ASR support; different model and modality." + "reason": "MiniCPM3 vs Qwen3-ASR are different model requests." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 and circuit-sparsity are different additions, not one fixable PR." + "reason": "MiniCPM3 and weight-sparse transformer are not the same underlying change." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 vs ModernVBERT are separate model requests." + "reason": "MiniCPM3 and ModernVBERT are separate model-add tickets." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 support and Voxtral TTS support target different code paths and modalities." + "reason": "MiniCPM3 and Voxtral TTS are different model targets." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are unrelated model integrations." + "reason": "DEIMv2 and Qwen3-TTS are unrelated additions." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT and Qwen3-ASR are different models and features." + "reason": "ModernVBERT and Qwen3-ASR are different model families and tasks." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT and Qwen3-TTS are not the same underlying change." + "reason": "ModernVBERT and Qwen3-TTS are not the same code-path change." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT and generic BERT-like RoPE support overlap conceptually, but not as one concrete duplicate change." + "reason": "ModernVBERT vs BERT-like models with RoPE is a broad thematic overlap, not a duplicate." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like RoPE support is a text-architecture change, not Voxtral TTS support." + "reason": "RoPE support request vs Voxtral TTS support request are different changes." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "TTS vs ASR are related only by family name; they are distinct model/support requests." + "reason": "Both are Qwen3 speech-related, but TTS and ASR are different concrete model integrations." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Qwen3-TTS and Voxtral TTS are different models and likely different implementations." + "reason": "Qwen3-TTS and Voxtral TTS are different models and implementations." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model additions." + "reason": "Different models and domains; no shared duplicate bug." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Audio TTS support and sparse-transformer support are different code-path changes." + "reason": "TTS support request vs sparse transformer request are unrelated." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio are both audio-related but not the same model or PR." + "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model requests." }, { "left": "issue:43837", - "right": "issue:44205", + "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and SAM3-LiteText are different model requests." + "reason": "ASR support and LFM2.5 Audio are different model additions." }, { - "left": "issue:44121", + "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Sparse-transformer support and EUPE encoder support are different architecture additions." + "reason": "Qwen3-ASR and EUPE target different model families and tasks." }, { "left": "issue:43837", - "right": "issue:45322", + "right": "issue:44121", "accept": false, - "reason": "ASR support and EUPE vision-language encoder support are unrelated changes." + "reason": "ASR support request vs weight-sparse transformer request are unrelated." }, { - "left": "issue:43837", - "right": "issue:44121", + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio and SAM3-LiteText are different model requests." + }, + { + "left": "issue:44121", + "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR and sparse-transformer support are distinct implementation tasks." + "reason": "Sparse transformer and EUPE are different concrete model changes." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS are different audio-model integrations." + "reason": "Qwen3-ASR and Voxtral TTS are distinct speech-model integrations." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral TTS are not the same model or code path." + "reason": "NeoBERT and Voxtral TTS are unrelated requests." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 support are different modality/model requests." + "reason": "ASR support and DeepSeek-OCR2 are different model additions." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE and TIPSv2 are both vision-language related, but they are different model additions and not one mergeable PR." + "reason": "EUPE and TIPSv2 are both vision-language related, but they are different models and changes, not the same fix." } ] }, @@ -17679,12 +17889,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "66dd281a37cedeb26342cf59e43071496e3582a1ed1ee153377658c2e91d6d31", + "prepared_review_unit_hash": "3fb562e696d96d146870bfd902d591beefcb11bc0f24d0d1a7d033da25125203", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, + "serialized_chars": 1479, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -17692,7 +17902,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, + "serialized_chars": 1479, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -17701,18 +17911,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are unrelated: one is a generic block decoding request, the other is a request for DeepSeek v3.2 support. They do not describe the same bug or feature implementation.", + "summary": "These two issues are not duplicates: one requests block decoding behavior, while the other asks for DeepSeek v3.2 model support. They describe different features/problems with no shared targets or code path.", "confidence": 0.96, - "canonical_issue_reason": "Issue #43037 is more specific and actionable, with a clear model/version target ('DeepSeek v3.2 support').", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue #43037 is the better global representative because it names a concrete compatibility request rather than a broad feature phrase.", + "best_issue_reason": "No canonical issue stands out because the issues are about unrelated functionality rather than the same underlying bug or feature request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "The titles and available metadata point to different requests: decoding behavior vs. model support. No shared code path or explicit linkage suggests a duplicate." + "reason": "Different requests: block decoding vs DeepSeek v3.2 support. No evidence they refer to the same bug, change, or code path." } ] }, @@ -17742,12 +17952,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "c3bd680aff46b7eef87a5672e003e27312b37e4e3963b1922aae71170528db8c", + "prepared_review_unit_hash": "5e286194a90ffcbc2b4a7f2c542f0e1916bd06993b08db088c7b714030ebb4f4", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1583, + "serialized_chars": 1584, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -17755,7 +17965,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1583, + "serialized_chars": 1584, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -17764,18 +17974,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are both about Trainer behavior with custom trainers, but they describe different problems: one is about `remove_unused_column` customizability, the other about estimating FLOPs for custom trainers. They are not duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue: the topics overlap in Trainer extensibility, but the underlying problems are different and should be tracked separately.", + "summary": "The two issues are related to Trainer extensibility, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": "Issue #43598 is the more concrete and narrowly scoped request, centered on a specific Trainer behavior (`remove_unused_column`) and its customizability.", "canonical_pr_reason": null, - "best_issue_reason": "If one issue must represent the cluster, 43598 is slightly more concrete and self-contained around a specific Trainer behavior, but it still does not subsume 43600.", + "best_issue_reason": "Issue #43598 is the best issue to keep as the representative item because it has a clear, actionable change request tied to one Trainer code path.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Different underlying bugs/features: `remove_unused_column` customization versus FLOPs estimation for custom trainers. Shared Trainer context is too broad to treat as the same issue." + "reason": "They concern different Trainer behaviors: column removal customizability vs FLOPS estimation for custom trainers. Shared theme is extensibility, but not the same underlying bug or change." } ] }, @@ -17805,7 +18015,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "d3f270cb26b9943d98d905d9923325e9cbff164237f5bc7566fc1d6039091a7c", + "prepared_review_unit_hash": "e9206a51b1e18b30d13ed162778a8992d3e4c12ce37a4939cd43171183491bb1", "component_budget": { "node_count": 2, "item_count": 2, @@ -17827,18 +18037,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are separate feature requests: one asks for Qwen3-Omni registration on AutoModel/AutoModelForConditionalGeneration, while the other asks for AutoModelForSequenceClassification support for Qwen3.5. They target different model families and different auto classes, so they are not duplicates.", - "confidence": 0.94, - "canonical_issue_reason": "issue:43684 is the broader and more central request about adding model registration for Qwen3-Omni, whereas issue:44405 is a narrower, distinct request for a different task head and model config.", + "summary": "These are related Qwen feature requests, but they are not duplicates: one asks for registering Qwen3-Omni with AutoModel/AutoModelForConditionalGeneration, while the other requests AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config). They target different model families and different auto-class APIs.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical duplicate here; the requests are about different Qwen variants and different auto-model classes.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43684 is the best representative issue for the cluster because it covers the more general registration gap and is closer to the apparent shared theme, even though the two issues are not duplicates.", + "best_issue_reason": "Issue #43684 is the broader, more foundational registration request and is the better representative of this small cluster\u2019s theme.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different Qwen variants and different auto classes/tasks; they do not describe the same underlying bug or change." + "reason": "Different underlying requests: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration vs. Qwen3.5 sequence classification support. Different model/config targets and different code paths." } ] }, @@ -17875,79 +18085,79 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "a55946b6308d30e02323eaa4c3c7abcde202d28dc0cc9c5664d6ba6ed72b6766", + "prepared_review_unit_hash": "2adaf9613d744269dfa53ce4f61ddd96931983a92b51391988eb966ca4e92992", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4263, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4266, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4263, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4266, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The three \"huggingface\" issues likely form one duplicate cluster, with issue 43845 as the best representative. Issue 43835 looks unrelated. I would only accept the 43845\u219443846 soft edge; the 43859 pairs are too weakly supported given the time gap and lack of additional evidence.", - "confidence": 0.8, - "canonical_issue_reason": "Issue 43845 is the earliest of the near-identical \"huggingface\" reports and has no stronger distinguishing evidence than the others.", + "summary": "Three near-identical closed issues titled \"huggingface\" look like duplicates of the same report; the \"Fett\" issue appears unrelated. I\u2019d pick the most active \"huggingface\" issue as the canonical representative.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43859 is the best representative among the duplicate-looking \"huggingface\" issues because it shares the same title and has the most discussion activity.", "canonical_pr_reason": null, - "best_issue_reason": "43845 is the best issue candidate because it matches the repeated title cluster and predates the other similar reports.", + "best_issue_reason": "43859 is the strongest issue candidate: same apparent topic as the others and slightly more activity, making it the best cluster representative.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different title and no shared targets/files or other content evidence; not enough to treat it as the same issue." + "reason": "Different titles and no evidence they describe the same bug or request." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different title and no shared targets/files; does not look like the same underlying report." + "reason": "Different titles and no concrete shared problem shown." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different title and no shared targets/files; no clear indication of the same bug." + "reason": "Different title; no sign this is the same underlying issue." }, { "left": "issue:43845", "right": "issue:43846", "accept": true, - "reason": "Exact same title and adjacent creation times make these look like duplicate reports of the same issue." + "reason": "Same title, created back-to-back, and no distinguishing evidence suggests they are separate reports." }, { "left": "issue:43845", "right": "issue:43859", - "accept": false, - "reason": "Same title, but the time gap is much larger and there is no content/link/file evidence tying them together." + "accept": true, + "reason": "Identical title and very similar metadata strongly suggest duplicate reports of the same issue." }, { "left": "issue:43846", "right": "issue:43859", - "accept": false, - "reason": "Same title, but the 13-hour gap and lack of supporting evidence make this too weak to accept as a duplicate pair." + "accept": true, + "reason": "Identical title and near-identical timing make them likely duplicates of the same report." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The canonical issue choice is grounded, but the soft-edge verdicts are not consistently conservative. In particular, the accepted pairs involving issue:43859 overstate the evidence by citing \u201cclose creation times\u201d even though 43859 is about 13 hours later than 43845/43846. With only matching titles and no content/link/file evidence, those edges should be treated more cautiously or left unaccepted." + "accept": true, + "feedback": "The summary is grounded in the packet: 43845/43846/43859 share the same title and close creation times, while 43835 differs in title and has no evidence tying it to the others. The canonical choice of 43859 is conservative and supported by its slightly higher discussion activity. No PR-related claims were made." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -17982,88 +18192,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "a0665adce8eff19c56a854915ddf47421755faa42e64a206056c68b4ea14a429", + "prepared_review_unit_hash": "b9bd213bee093e8f6842b80ef5768a39bc5aeee011b5b22cfea493178d62ee43", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All items look like unrelated low-signal/spam issues rather than duplicates of one concrete bug or request. No soft pair should be merged.", - "confidence": 0.86, - "canonical_issue_reason": "No clear canonical issue exists; #44418 is only the most descriptive title, but the set appears to be unrelated spam/placeholder reports.", + "summary": "This cluster looks like unrelated low-information/spam issues rather than a duplicate set. One item is a documentation suggestion, one mentions \"Google LLC\", and the rest are generic spam placeholders.", + "confidence": 0.91, + "canonical_issue_reason": "No clear canonical issue: the items do not describe the same underlying bug or change, so there is no valid representative duplicate to merge around.", "canonical_pr_reason": null, - "best_issue_reason": "#44418 is the best placeholder canonical only because its title is more specific than the others; it still does not clearly represent the same underlying problem as the rest.", + "best_issue_reason": "Issue 44418 is the only semantically readable title, but it still does not match the other spam-like items and is not a good duplicate representative.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are generic spam placeholders with no shared concrete bug or request." + "reason": "Both are spam-labeled placeholders, but there is no evidence they report the same underlying issue or change." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Same as above: no evidence they describe the same underlying issue." + "reason": "Identical placeholder-style titles are not enough to treat them as duplicates; no shared bug/change is described." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "Titles are unrelated spam/documentation placeholders, not the same issue." + "reason": "A spam placeholder and a documentation enhancement suggestion are clearly different kinds of issues." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "One is a spam placeholder and the other is a different quoted-string title; no shared bug or change." + "reason": "No sign these describe the same problem; one is a spam placeholder and the other is a distinct titled item." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "Documentation-suggestion spam and generic spam placeholder are not the same underlying report." + "reason": "Documentation enhancement and spam placeholder are unrelated; no shared underlying change is evident." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "No concrete overlap beyond being low-information issues." + "reason": "Documentation enhancement and spam placeholder are unrelated; not the same issue." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Identical-looking spam titles, but no body or explicit details tying them to the same underlying bug/change." + "reason": "Same placeholder title alone is insufficient; there is no content tying them to the same underlying bug or request." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Generic spam placeholder versus a different quoted title; insufficient to treat as duplicates." + "reason": "No concrete overlap is shown between a spam placeholder and the \"Google LLC\" issue." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "No shared concrete problem; only superficial similarity from low-quality titles." + "reason": "No concrete overlap is shown between a spam placeholder and the \"Google LLC\" issue." } ] }, @@ -18093,12 +18303,12 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "e85cd0dede3ea2e939603810c805f9cc1ff04e5a7cbe62670e59d0fb5cb1bc0e", + "prepared_review_unit_hash": "f12390b74d023fc2eb88cd190687a802b139dcb41ffd1658646cd52a2c218826", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, + "serialized_chars": 1646, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -18106,7 +18316,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, + "serialized_chars": 1646, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -18115,24 +18325,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues have identical titles and very similar bodies, so a duplicate relationship is plausible, but the evidence is limited to soft similarity with no explicit cross-links. #45145 is the better canonical issue because it was opened earlier and has one inbound reference.", - "confidence": 0.74, - "canonical_issue_reason": "#45145 is earlier and has 1 inbound reference versus 0 for #45151, so it is the better canonical issue if these are duplicates.", + "summary": "The two closed issues have the same title and very similar body size, so they are likely duplicate or near-duplicate reports. #45145 is the better canonical issue because it was created first and already has an inbound reference.", + "confidence": 0.84, + "canonical_issue_reason": "#45145 is earlier and has an inbound reference, so it is the more established duplicate target.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the strongest issue representative because it was created earlier and has the only inbound reference in the pair.", + "best_issue_reason": "#45145 is the best single issue to keep because it appears to be the earliest and most representative report of the pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": false, - "reason": "The match is suggestive, with identical titles and similar bodies, but there are no explicit links or shared targets/files. That is not enough to confidently merge them as the same issue." + "reason": "They look likely related or duplicate, but the packet only provides soft similarity plus basic metadata; that is not enough to confidently treat them as the same underlying issue." } ] }, "evaluator_result": { "accept": false, - "feedback": "Reject: the duplicate-issue conclusion is plausible, but the writeup is not fully conservative. In particular, `best_issue_reason` says #45145 has 'slightly more discussion context,' yet the packet shows 0 discussion_activity for #45145 and 1 for #45151. Also, the soft-edge acceptance is stronger than the evidence warrants because the pair is supported only by soft similarity with no explicit links. The canonical-issue choice (#45145 for earlier creation and inbound references) is grounded, but the overall confidence and duplicate verdict should be toned down." + "feedback": "The packet supports a likely duplicate/near-duplicate relationship: the titles match exactly, the bodies are similar in size, and #45145 is earlier with one inbound reference. However, the soft-edge verdict is too strong for the evidence available. The packet only provides soft_similarity plus metadata, so the conclusion should stay conservative (e.g., 'likely duplicate' or 'probably same report'), not 'same underlying issue rather than merely related reports.'" }, "evaluator_used": true, "retried": true, @@ -18213,7 +18423,7 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "52db19df4d65de3f54ea26ac28c3c9b754315a3451af4f47c48fbf8d4f4b4a9a", + "prepared_review_unit_hash": "2b455ed2231a85a0e27d0db2c41d37847a11bb50729e400c6936a236e129b06e", "component_budget": { "node_count": 2, "item_count": 2, @@ -18235,18 +18445,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate 4-bit inference bugs in different code paths: one is a Gemma4/audio-module uint8/torch.finfo TypeError, the other is a BitsAndBytesConfig regression where llm_int8_skip_modules drops the default lm_head exclusion and triggers an AssertionError.", + "summary": "Both items are about 4-bit quantized inference, but they describe different regressions in different code paths: one is a Gemma4/audio-module uint8 vs. torch.finfo TypeError, the other is a BitsAndBytesConfig skip-modules regression that removes the default lm_head exclusion and triggers an AssertionError.", "confidence": 0.95, - "canonical_issue_reason": "No true canonical duplicate here: 45672 is a Gemma4-specific dtype/torch.finfo failure in audio modules, while 45674 is a BitsAndBytesConfig skip-module regression affecting lm_head exclusion. Different root causes and fixes.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "If one issue must represent the cluster, 45674 is the slightly better standalone representative because it describes a broader config regression, but it is not a safe duplicate of 45672.", + "best_issue_reason": "Issue 45674 is the cleaner standalone report: it names the configuration knob, the exact regression, and the resulting AssertionError, making it the better representative issue in this small cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Same broad 4-bit inference area, but the failure modes differ: Gemma4 audio uint8/torch.finfo vs. llm_int8_skip_modules clearing lm_head exclusion. Not the same underlying bug." + "reason": "Not the same underlying bug: 45672 concerns Gemma4/audio-module uint8 weights hitting torch.finfo during NF4 inference, while 45674 concerns llm_int8_skip_modules clearing the default lm_head exclusion and causing an AssertionError. They share only the broad 4-bit inference context." } ] }, @@ -18276,40 +18486,40 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "5b19e1c83bf6b4dfa790895b3cee889d32416b29d04cb33efe949e62ec3890de", + "prepared_review_unit_hash": "558f54fd9b8d868ea55224a139067871178efe8bb04ebe6b9a4dc069ea63d630", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1476, - "estimated_input_tokens": 369, - "estimated_eval_tokens": 994 + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1476, - "estimated_input_tokens": 369, - "estimated_eval_tokens": 994 + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two empty closed issues with different personal-name titles and no substantive content, so they do not look like duplicates of a real bug report.", + "summary": "The two issues are unrelated closed items with only superficial title similarity (personal names + dates). There is no evidence they describe the same bug or request, so they should not be clustered as duplicates.", "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a good canonical representative; both are empty personal-name posts with no bug context or shared technical problem.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Different names and dates, no body or technical context, and no evidence they describe the same underlying issue." + "reason": "Different named subjects and different dates; no shared problem statement, target, or code context. The similarity is incidental, not duplicate evidence." } ] }, @@ -18339,7 +18549,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "36d6218967995c8ea595c68c0e38cf763100e8eb3d1644f9e26accd348eae84b", + "prepared_review_unit_hash": "c9f858fcd367c816893b2f046e70e47f7abc68c12a559cfb4c641720cd8aa929", "component_budget": { "node_count": 2, "item_count": 2, @@ -18361,18 +18571,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated name/date entries and do not appear to describe the same underlying bug or request.", + "summary": "These two closed issues are just separate personal-name/date entries with no shared bug context, so they should not be clustered as duplicates.", "confidence": 0.98, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue 45764 is the earliest-created item and, given the lack of any substantive body or context in either issue, is the most reasonable placeholder canonical issue.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue 45764 is the least arbitrary representative for the same reason: earliest created and equally sparse as the other issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "Shared surname/date formatting is not enough; these are distinct person-name issues with no code, symptoms, or tracking links indicating the same underlying problem." + "reason": "The titles refer to different people and dates, with no shared problem statement, code path, or explicit linkage; they do not appear to describe the same underlying issue." } ] }, @@ -18402,7 +18612,7 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "34041f06ff79617c0f9a5b9b46b19307f2b6ae7415b257617f166d7eb284a2aa", + "prepared_review_unit_hash": "014f8f99a0ecbf77a1aff23c8ef51296065ea325282fed5ddeaea0682ab9172e", "component_budget": { "node_count": 2, "item_count": 2, @@ -18424,18 +18634,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are unrelated feature requests: one asks for a new Qwen3_5 token-classification/value-model class, while the other asks for FP8 training support. The soft similarity is too weak to treat them as duplicates.", + "summary": "The two issues are not duplicates: one requests a new Qwen3_5 token-classification/value-model class, while the other asks for FP8 training support. They concern different features and code paths.", "confidence": 0.96, - "canonical_issue_reason": null, + "canonical_issue_reason": "Issue #45810 is the more concrete, implementation-specific request; it names a missing model class and a clear target use case.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Issue #45810 is the best representative issue because it is narrowly scoped and directly tied to an actionable model-class addition.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Different underlying requests and code paths: model-class support for Qwen3_5 token classification vs. FP8 training support. They do not describe the same bug or change." + "reason": "These are unrelated feature requests: one is for adding Qwen3_5ForTokenClassification, the other is for FP8 training support. They do not describe the same underlying bug or change." } ] }, @@ -18475,88 +18685,88 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "c946245fed8e0f477e06b786dbcd52a85e3c2c1f46ac71bd15396f7ea7d2b0f8", + "prepared_review_unit_hash": "12b022c514266ee11c7a182d270a830930e482c70066c872d3b2e9c473dd12d2", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5726, - "estimated_input_tokens": 1432, - "estimated_eval_tokens": 3120 + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5726, - "estimated_input_tokens": 1432, - "estimated_eval_tokens": 3120 + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "45869 is separate from the rest. 45870/45871 and 45872/45873 look very similar, but the packet only shows metadata and body length, so I would not call either pair a confirmed duplicate. The cross-links are thematic rather than evidence of one shared underlying request.", - "confidence": 0.72, - "canonical_issue_reason": "If a single issue must represent the cluster, 45870 is the best candidate because it has the clearest overlap with 45871 and slightly more activity, but this is only a tentative representative choice, not a confirmed duplicate canonical.", + "summary": "The cluster appears to split into two highly similar issue pairs\u201445870/45871 and 45872/45873\u2014plus one unrelated dataset promotion issue (45869). Based on the available metadata, I would not call any pair proven exact duplicates; the evidence is suggestive but not conclusive.", + "confidence": 0.84, + "canonical_issue_reason": "No single issue cleanly represents the whole cluster because it contains at least two distinct topic groups and an unrelated dataset post. If a representative is needed, 45870 is the best anchor for the first DataForge example pair because it is the earlier of the two nearly identical reports.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the strongest overall representative issue in this packet: it has the most visible activity among the similar issues and the closest match to another same-title issue, while 45869 is clearly separate.", + "best_issue_reason": "45870 is the strongest overall issue representative only as the earliest, cleanest member of the most tightly matched pair. It is not a perfect global representative because 45872/45873 cover a different DataForge topic and 45869 is unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Different request types: a financial dataset bundle versus a DataForge example. Related domain only, not the same underlying issue." + "reason": "Dataset promotion vs DataForge prediction example; different topics." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Different request types: a financial dataset bundle versus a DataForge example. Similar theme only, not a duplicate." + "reason": "Dataset promotion vs DataForge prediction example; different topics." }, { "left": "issue:45870", "right": "issue:45871", "accept": false, - "reason": "Same title and body size strongly suggest a near-duplicate, but the full bodies are not shown here, so I cannot safely confirm them as the same underlying issue." + "reason": "Same title and similar metadata, but the packet does not prove an exact duplicate or same underlying issue." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Both are financial/time-series themed, but one is an example and the other is a scraping-service request. Not the same concrete issue." + "reason": "Both mention finance/DataForge, but one is a prediction example and the other is a scraping-service topic; distinct requests." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Both are financial/time-series themed, but one is an example and the other is a scraping-service request. Not the same concrete issue." + "reason": "Both mention finance/DataForge, but one is a prediction example and the other is a scraping-service topic; distinct requests." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Different requests: an example issue versus a scraping-service request. Similar topic, not a confirmed duplicate." + "reason": "Prediction example vs scraping-service topic; not the same underlying issue." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Different requests: an example issue versus a scraping-service request. Similar topic, not a confirmed duplicate." + "reason": "Prediction example vs scraping-service topic; not the same underlying issue." }, { "left": "issue:45872", "right": "issue:45873", "accept": false, - "reason": "Same title and body size suggest a near-duplicate, but without the full bodies I would not mark this as a confirmed duplicate pair." + "reason": "Same title and similar metadata, but the packet does not prove an exact duplicate or same underlying issue." } ] }, "evaluator_result": { "accept": false, - "feedback": "The summary is a bit too strong for the evidence in the packet. We can conservatively say 45869 is separate and that 45870/45871 and 45872/45873 look very similar, but the packet does not show the full bodies, so calling them \"exact duplicate\" pairs is overstated. Also, all pair entries are marked deterministic_accept=false, so the two true accepts should be softened or left unaccepted." + "feedback": "The summary overstates the evidence. The packet only provides titles, body lengths, and soft-similarity scores; it does not prove any pair is an exact duplicate. In particular, calling 45870/45871 and 45872/45873 \"clear\" or \"exact\" duplicates is stronger than the available metadata supports, and every pair in the packet has deterministic_accept=false. A more conservative summary would say the cluster appears to split into two highly similar topic pairs plus an unrelated dataset post, without asserting exact duplication." }, "evaluator_used": true, "retried": true, @@ -18567,7 +18777,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18618,166 +18828,172 @@ "pull_request:44827|pull_request:46048", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "5d259541d321071d43eebb812035177fe40b6adba4190b8e6bd909d98d2edd5a", + "prepared_review_unit_hash": "cba00d6424d124a7562e958998ca7cc20b0987b39b7d0ac50401fffcbeed37d2", "component_budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 22, - "serialized_chars": 43090, - "estimated_input_tokens": 10773, - "estimated_eval_tokens": 21802 + "serialized_chars": 43088, + "estimated_input_tokens": 10772, + "estimated_eval_tokens": 21800 }, "budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 22, - "serialized_chars": 43090, - "estimated_input_tokens": 10773, - "estimated_eval_tokens": 21802 + "serialized_chars": 43088, + "estimated_input_tokens": 10772, + "estimated_eval_tokens": 21800 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a loose set of unrelated PRs spanning model additions, docs, tests, and separate bug fixes. The only clear duplicate pair is the two `fixed_cross_entropy` PRs, which patch the same function for the same issue.", + "summary": "Only one true duplicate pair appears here: the two `fixed_cross_entropy` PRs both address issue #43240 by adding supported kwargs to the same loss call. The rest are distinct model additions, refactors, or unrelated fixes that merely share broad subsystem similarity.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "`pull_request:43251` is the better canonical PR because it explicitly targets issue 43240, has the fuller fix context, and is the clearer representative of the duplicate `fixed_cross_entropy` change.", + "canonical_pr_reason": "PR #43251 is the strongest canonical representative: it has the explicit issue target (#43240) and the fuller implementation of the same `fixed_cross_entropy` fix.", "best_issue_reason": null, - "best_pr_reason": "`pull_request:43251` is the strongest global representative among the provided items because it is the only clearly duplicated bug fix, is directly issue-linked, and is more complete than the near-identical alternative.", + "best_pr_reason": "PR #43251 is the best overall representative because it directly fixes the tracked issue and cleanly captures the underlying change in one place.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Different areas: VideoPrism model addition vs. ExecuTorch dynamic-shape export test. Shared theme is only broad framework support, not the same change." + "reason": "Videoprism model addition vs executorch dynamic-shape export test; unrelated changes." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Tokenizer method addition vs. export test; unrelated code paths and no same underlying bug." + "reason": "Mistral tokenizer method additions vs executorch export tests; different code paths and goals." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Mistral tokenizer helpers and a new MobileLLM model are separate feature additions." + "reason": "Tokenizer API additions vs adding MobileLLM model support; not the same underlying change." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor robustness updates across models vs. adding MobileLLM; not the same fix." - }, - { - "left": "pull_request:43251", - "right": "pull_request:43254", - "accept": true, - "reason": "Both modify `fixed_cross_entropy` in the same file to pass supported kwargs to `nn.functional.cross_entropy`; same underlying bug/fix." + "reason": "Processor auto-loading robustness work vs new MobileLLM model files; broad shared area only." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss-function kwargs fix vs. new MobileLLM model; unrelated." + "reason": "Loss-function kwargs fix vs new model implementation; unrelated." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Different changes: cross-entropy kwargs vs. Mistral4 test/model adjustments. Not mergeable as one bug fix." + "reason": "Cross-entropy kwargs fix vs Mistral4 test/model tweaks; no shared bug or code path." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "ExecuTorch export test and Mistral4 test/model work are distinct test-focused changes." + "reason": "Executorch dynamic-shape test vs Mistral4 test fixes; both are tests, but not the same issue." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "Dynamic-shape export test vs. model addition; no shared concrete bug." + "reason": "Executorch export test vs MobileLLM model addition; unrelated." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "Model addition vs. class config resolution ordering fix; unrelated." + "reason": "Model addition vs config-resolution ordering fix; distinct changes." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "New model implementation vs. Falcon output-collection refactor; separate code paths." + "reason": "MobileLLM implementation vs Falcon refactor for standardized output capture; different models and purposes." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon model refactor and a model-creation skill are unrelated." + "reason": "Falcon model refactor vs model-creation skill/docs; not the same code-path problem." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon refactor vs. Mistral4 test fixes; different models and issues." + "reason": "Falcon output-interface refactor vs Mistral4 test fixes; unrelated." }, { "left": "pull_request:44827", - "right": "pull_request:45170", + "right": "pull_request:46344", "accept": false, - "reason": "Mistral4 test fixes vs. `layrnorm`/`layernorm` rename across CLIP-like models; different bug classes." + "reason": "Mistral4 test fixes vs VitPose graph-break fix; different model subsystems and bugs." }, { "left": "pull_request:44827", - "right": "pull_request:45213", + "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 fixes vs. model-creation skill; not the same change." + "reason": "Mistral4 test fixes vs layrnorm/layernorm rename sweep; unrelated." }, { "left": "pull_request:44827", - "right": "pull_request:45655", + "right": "pull_request:46267", "accept": false, - "reason": "Model-specific test fixes vs. generic subclass config-resolution fix; unrelated." + "reason": "Mistral4 test fixes vs chat-response parsing overhaul; different feature areas." }, { "left": "pull_request:44827", - "right": "pull_request:46048", + "right": "pull_request:46367", "accept": false, - "reason": "Mistral4 test/model work vs. adding a TDT loss kernel; different subsystems and purposes." + "reason": "Mistral4 test fixes vs deepseek/xpu kernel integration; different model and kernel work." }, { "left": "pull_request:44827", - "right": "pull_request:46267", + "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 fixes vs. chat response parsing updates; not the same underlying bug." + "reason": "Mistral4 test fixes vs model-creation skill content; completely unrelated." }, { - "left": "pull_request:44827", - "right": "pull_request:46344", + "left": "pull_request:43823", + "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 test fixes and VitPose graph-break fix address different models and failures." + "reason": "New MobileLLM model support vs repository skill/docs change; not the same artifact." }, { "left": "pull_request:44827", - "right": "pull_request:46367", + "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 test fixes vs. finegrained FP8 kernel name changes; unrelated." + "reason": "Mistral4 test fixes vs config-resolution order fix; different problems." }, { "left": "pull_request:46048", "right": "pull_request:46367", "accept": false, - "reason": "TDT loss kernel integration vs. finegrained FP8 kernel integration; same general kernel area but clearly different features and code paths." + "reason": "TDT loss kernel integration vs finegrained-fp8/xpu kernel integration; same general integration area, but different kernels and code paths." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 test fixes vs TDT loss kernel addition; unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs patch the same `fixed_cross_entropy` function to pass through supported `cross_entropy` kwargs for issue #43240, and could plausibly be merged into one change." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the only accepted soft-edge pair (43251/43254) is a plausible single-bugfix duplicate based on the shared file, shared issue target, and near-identical cross-entropy kwargs change. The other rejections are conservative and do not overstate evidence." + "feedback": "The summary is grounded in the packet. The only clearly mergeable duplicate pair is the two `fixed_cross_entropy` PRs (#43251 and #43254), which both target issue #43240 and patch the same loss helper in compatible ways. The other pair verdicts are conservative and consistent with the listed filenames/diffs, and no unsupported broad claims stand out." }, "evaluator_used": true, "retried": false, @@ -18788,7 +19004,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18806,46 +19022,46 @@ "pull_request:40546|pull_request:42781", "pull_request:42781|pull_request:46331" ], - "prepared_review_unit_hash": "4a16c01e94ca48d35bbd3db57e18c61f36cc42b1a0e81037fbff66f85e85d3fe", + "prepared_review_unit_hash": "ad88523de97426c411f1255c18c2f8fadd252b7f4ff61517a621acdd53719b03", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10381, - "estimated_input_tokens": 2596, - "estimated_eval_tokens": 5448 + "serialized_chars": 10379, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10381, - "estimated_input_tokens": 2596, - "estimated_eval_tokens": 5448 + "serialized_chars": 10379, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three distinct pull requests: a base VibeVoice implementation, a VibeVoice Realtime extension, and an unrelated Parakeet-RNNT model addition. The two VibeVoice PRs share files because the latter builds on the same family, but they are not the same concrete change; Parakeet-RNNT is a separate model altogether.", - "confidence": 0.95, + "summary": "This cluster contains three distinct open PRs for speech-model additions: a VibeVoice implementation, a VibeVoice Realtime extension, and a Parakeet-RNNT addition. They overlap in auto-mapping/doc plumbing, but they are not the same underlying change.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 40546 is the best representative of the cluster: it is the original, non-draft core VibeVoice implementation with substantial review activity. PR 42781 is a draft follow-up for VibeVoice Realtime, and PR 46331 targets a different model family (Parakeet-RNNT).", + "canonical_pr_reason": "PR #40546 is the cleanest canonical representative: it is the original VibeVoice implementation, has substantial review activity, and is more central than the later draft realtime extension.", "best_issue_reason": null, - "best_pr_reason": "PR 40546 is the strongest single PR to anchor the cluster because it is the foundational VibeVoice change and has the broadest completed implementation footprint. The other PRs are either an incremental variant or unrelated.", + "best_pr_reason": "PR #40546 is the best single PR for the cluster because it is the core VibeVoice add-on; #42781 is a different realtime feature set, and #46331 is for a different model family entirely.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Related model family, but not the same change: 40546 adds core VibeVoice support while 42781 adds VibeVoice Realtime with additional tokenizer/feature-extractor/generation work. They could not be safely merged as one duplicate PR." + "reason": "Both are VibeVoice-related and share doc/auto-registration files, but #42781 adds VibeVoice Realtime with extra acoustic-tokenizer and generation changes, so it is a separate feature rather than the same concrete fix." }, { "left": "pull_request:42781", "right": "pull_request:46331", "accept": false, - "reason": "Different underlying models and code paths. Shared auto-registration files are incidental; VibeVoice Realtime and Parakeet-RNNT are separate additions, not duplicates." + "reason": "These target different model families (VibeVoice Realtime vs Parakeet-RNNT). Shared auto-mapping files are broad plumbing overlap, not evidence of the same underlying change." } ] }, @@ -18859,7 +19075,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18875,12 +19091,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "04a498d021a3395be73d4fd2c0fa982d3472d0efba58e6cb735073f6b3a86b75", + "prepared_review_unit_hash": "12ca58dec53934a4d49a2c62113833ca2dc03c6bc3a753c2f0576a43c38991ce", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -18888,7 +19104,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -18897,8 +19113,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: one adds covariate-based forecasting support to TimesFM, while the other adds DINOv3 image-classification support. They touch different models, code paths, and user-facing features.", - "confidence": 0.98, + "summary": "These two pull requests are unrelated: one adds TimesFM covariate forecasting support, while the other adds DINOv3 image classification support. They touch different models, code paths, and docs, so they should not be clustered as duplicates.", + "confidence": 0.99, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -18908,7 +19124,7 @@ "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different underlying changes and model families: TimesFM covariate forecasting vs DINOv3 image classification. They do not fix the same bug or implement the same feature, so they should not be merged as duplicates." + "reason": "Different underlying changes for different model families: TimesFM forecasting with covariates vs DINOv3 image classification support. No shared bug, feature, or mergeable code-path overlap beyond generic docs/tests/model plumbing." } ] }, @@ -18922,7 +19138,7 @@ { "label": "pull_request", "component_index": 4, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -18985,7 +19201,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19003,12 +19219,12 @@ "pull_request:41797|pull_request:45186", "pull_request:45186|pull_request:45630" ], - "prepared_review_unit_hash": "0da2a0d8b4c41db733ec56200d4b34d8db3b68f70c18725fb2deb4d15589f054", + "prepared_review_unit_hash": "45fb9d56dd0cee1594598486fd6dd977b9667ade073c9aabed567d7e5691ee23", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10436, + "serialized_chars": 10435, "estimated_input_tokens": 2609, "estimated_eval_tokens": 5474 }, @@ -19016,7 +19232,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10436, + "serialized_chars": 10435, "estimated_input_tokens": 2609, "estimated_eval_tokens": 5474 }, @@ -19025,24 +19241,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs all add different new models and documentation, but they do not fix the same underlying bug or implement the same change. They share generic model-registration scaffolding, yet each targets a distinct model family and code path.", + "summary": "Three open PRs add unrelated new models (DeepSeek OCR, Isaac, Kimi2-6). They overlap on onboarding boilerplate, but not on the same underlying change or code path.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: DeepSeek OCR, Isaac, and Kimi2-6 are separate model-addition pull requests with different model-specific files, APIs, and tests, not the same change.", + "canonical_pr_reason": "No canonical PR: these are distinct model additions for different model families and implementation paths, not duplicates.", "best_issue_reason": null, - "best_pr_reason": "No best PR as a duplicate representative; if forced, 45630 is the broadest model-addition PR, but it is still unrelated to the others beyond shared boilerplate.", + "best_pr_reason": "No single best PR for deduping this cluster; the shared files are mostly registration/docs boilerplate, while the substantive model code differs across all three PRs.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different model additions (DeepSeek OCR vs Isaac) with distinct model directories and implementations; shared auto/registration files are boilerplate, not evidence of the same code-path change." + "reason": "Different model additions (DeepSeek OCR vs Isaac). Shared auto-mapping/docs boilerplate is not enough to indicate the same underlying change." }, { "left": "pull_request:45186", "right": "pull_request:45630", "accept": false, - "reason": "Different model additions (Isaac vs Kimi2-6) touching separate model families and files; they are not plausibly mergeable into one PR for the same underlying fix or feature." + "reason": "Different model families and code paths (Isaac vs Kimi2-6). They share onboarding scaffolding, but not a concrete fix or mergeable duplicate PR." } ] }, @@ -19056,7 +19272,7 @@ { "label": "pull_request", "component_index": 6, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19072,7 +19288,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "85d660f4857e0281237db082dfd4a633f98aa26eaa697e022061f5f8c9586692", + "prepared_review_unit_hash": "1f2ddb817b71bde1ce4fd72d2f6dfd96701b01e18f923bf17d8218858bb6677a", "component_budget": { "node_count": 2, "item_count": 2, @@ -19094,18 +19310,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs both update GitHub Actions workflow dependencies, but they target different underlying changes: one is a Node 24 compatibility sweep for core actions, the other upgrades build/publish-related actions (notably docker/build-push-action) and other latest versions. They overlap in workflows but are not the same fix/change set.", - "confidence": 0.93, + "summary": "These are two separate GitHub Actions maintenance PRs: one upgrades general workflow actions for Node 24 compatibility, the other bumps docker/build-push-action to v6 in docker-related workflows. They overlap only at the workflow-infrastructure level, not the same concrete change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the better canonical representative because it has the clearest project-wide theme and broader workflow-action compatibility focus, making it the more central/recognizable upgrade PR in this cluster.", + "canonical_pr_reason": "No canonical PR: 42975 and 42976 touch different action families and solve different upgrade tasks, so they should not be deduplicated.", "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the best single representative because its title and diff describe the broader GitHub Actions compatibility upgrade, while PR 42976 is a narrower action-version bump centered on docker/build-push-action and related workflow updates.", + "best_pr_reason": "PR 42975 is the broader workflow-actions upgrade and is the more representative umbrella change, but it is still not a duplicate of 42976.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Both are GitHub Actions upgrade PRs, but they change different action families and different concrete code paths; they would not plausibly be merged as one PR without conflating separate upgrade goals." + "reason": "Different upgrades and different action ecosystems: Node 24 compatibility bumps core workflow actions, while the other only updates docker/build-push-action to v6. They are not the same underlying PR." } ] }, @@ -19119,7 +19335,7 @@ { "label": "pull_request", "component_index": 7, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19147,7 +19363,7 @@ "pull_request:43451|pull_request:46411", "pull_request:45490|pull_request:46411" ], - "prepared_review_unit_hash": "a89dbd507d82fab88d5ef1be64afcc1462c44301781f187d4b7c3f91e2c08669", + "prepared_review_unit_hash": "332ef72d80aab67c606a602127a60e24a1da61e5466869ae28c5c1c41baf4388", "component_budget": { "node_count": 5, "item_count": 5, @@ -19169,72 +19385,72 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are all separate model-addition changes (ViT NEPA, Molmo, Molmo2, CTSM, MiniMax M2.7) with overlapping framework plumbing, but they do not fix the same underlying bug or implement the same feature.", - "confidence": 0.97, + "summary": "All five PRs are separate model-addition efforts for different model families (ViT NEPA, Molmo, Molmo2, CTSM, MiniMax M2.7). They share a common scaffold of docs/auto-registration changes, but they do not look like duplicates of the same underlying bug or feature.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the cluster is heterogeneous, and the PRs target different models and code paths rather than duplicating one another.", + "canonical_pr_reason": "No single canonical PR stands out for deduplication because each PR adds a different model/package and touches different model-specific code paths.", "best_issue_reason": null, - "best_pr_reason": "PR 43451 (Molmo2) is the most substantial and review-heavy of the set, but it is still not a duplicate of the others.", + "best_pr_reason": "No PR is a good global canonical duplicate target; the PRs are unrelated model integrations rather than alternative fixes for one change.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different models and feature sets: ViT NEPA vs Molmo. Shared auto/docs plumbing is incidental, not the same change." + "reason": "Different models and code paths: ViT NEPA vs Molmo. Shared docs/auto-registration patterns are not enough to call them the same change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "ViT NEPA and Molmo2 are unrelated model additions with different architectures and files; only generic framework registration overlaps." + "reason": "Different model families: ViT NEPA vs Molmo2. They add separate implementations, not the same underlying feature." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model families (ViT NEPA vs CTSM); overlapping docs/auto-registration does not indicate the same underlying change." + "reason": "ViT NEPA and CTSM are unrelated model additions, despite both updating docs and auto mappings." }, { "left": "pull_request:42978", "right": "pull_request:46411", "accept": false, - "reason": "ViT NEPA model addition is unrelated to MiniMax M2.7 code import; they do not share a concrete bug or fix path." + "reason": "ViT NEPA and MiniMax M2.7 are clearly different model integrations with no shared concrete bug/change." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are distinct models with separate configs, processors, and tests; not mergeable as one PR." + "reason": "Molmo and Molmo2 are related by name, but they are distinct model releases with different files and capabilities, so not the same change." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are different model implementations; shared auto/docs updates are broad subsystem overlap only." + "reason": "Molmo and CTSM are different model additions; overlap is only in generic Transformers plumbing." }, { "left": "pull_request:43448", "right": "pull_request:46411", "accept": false, - "reason": "Molmo and MiniMax M2.7 are unrelated model additions; no common concrete bug or code-path fix." + "reason": "Molmo and MiniMax M2.7 target different model families and code paths." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are separate model additions with different model code and tests; similarity is only in registration/docs scaffolding." + "reason": "Molmo2 and CTSM are separate model integrations, not one shared fix or feature." }, { "left": "pull_request:43451", "right": "pull_request:46411", "accept": false, - "reason": "Molmo2 and MiniMax M2.7 do not address the same model or bug; they are distinct changes." + "reason": "Molmo2 and MiniMax M2.7 are unrelated model additions." }, { "left": "pull_request:45490", "right": "pull_request:46411", "accept": false, - "reason": "CTSM and MiniMax M2.7 are unrelated model-introduction PRs with no shared concrete fix path." + "reason": "CTSM and MiniMax M2.7 are distinct model-specific PRs with no evidence of a common underlying issue." } ] }, @@ -19248,7 +19464,7 @@ { "label": "pull_request", "component_index": 8, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19271,7 +19487,7 @@ "pull_request:43665|pull_request:46380", "pull_request:44815|pull_request:46380" ], - "prepared_review_unit_hash": "3fa9bd718c9ce7c71132158665b23fd1a26e86b18a35eec77b76a270744b0515", + "prepared_review_unit_hash": "192b572cc904d08ee5acfb45771970515bf3889215407ad45f7c07a0f9333565", "component_budget": { "node_count": 4, "item_count": 4, @@ -19293,48 +19509,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All four items are pull requests, but they appear to address different problems. PR 46380 is the strongest substantive fix; 44815 is a separate FP8 dequant/loading fix; 43291 is a Whisper tokenizer test/fix; and 43665 looks like a noisy/bogus test-only change rather than a real bug fix.", - "confidence": 0.91, + "summary": "These four PRs are not duplicates of one another: they target unrelated code paths (Whisper decoding, test-only edits, FP8 dequantization, and generation config validation). No soft-edge pair looks like the same underlying fix or change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "46380 is the most clearly grounded bug fix: it patches generation config validation for dict-valued watermarking_config and includes broad regression coverage, indicating a real shared code-path issue.", + "canonical_pr_reason": "PR #46380 is the strongest representative of the cluster because it contains a concrete runtime bug fix in core generation config validation, with broader but coherent follow-up test coverage.", "best_issue_reason": null, - "best_pr_reason": "46380 is the best overall PR because it fixes a concrete runtime/config validation bug in a central path and has the strongest evidence of a legitimate, mergeable change set.", + "best_pr_reason": "PR #46380 is the best standalone PR here: it fixes a clear validation/type-conversion bug in an actual code path rather than only adjusting tests or touching an unrelated subsystem.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Different code paths and likely unrelated. 43291 changes Whisper tokenizer decoding; 43665 only injects failing assertions into CLIP/ViT tests, which does not indicate the same underlying bug or fix." + "reason": "Whisper tokenizer decode logic and test-only CLIP/ViT edits are unrelated; they do not look like the same bug or change." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Unrelated fixes: Whisper decoding vs FP8 dequant/loading logic. No shared concrete bug or mergeable single change is evident." + "reason": "Whisper decoding and FP8 dequantization are different subsystems with different failure modes and code paths." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Both touch Whisper-related code, but the changes target different issues: tokenizer timestamp decoding vs generation watermarking validation / broader generation behavior. Not the same underlying bug." + "reason": "Both touch a Whisper file, but one fixes timestamp decoding while the other fixes generation watermarking config validation; not the same underlying bug." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Completely different artifacts: bogus test assertions in CLIP/ViT vs FP8 quantization/loading code. No shared bug/change." + "reason": "A test-only PR with placeholder asserts is not the same change as an FP8 dequantization fix." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "43665 is test noise across CLIP/ViT, while 46380 is a real generation/config fix. They do not look like the same underlying issue." + "reason": "Test placeholder edits across CLIP/ViT are unrelated to generation config validation and watermarking handling." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "Both are infrastructure-level fixes, but they address distinct problems: FP8 dequantization/padding vs generation watermarking validation and related model tests. Too broad to merge as one PR." + "reason": "FP8 dequant/padding changes and generation config validation are different concrete problems; shared broad ML subsystem is not enough." } ] }, @@ -19348,7 +19564,7 @@ { "label": "pull_request", "component_index": 9, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19364,12 +19580,12 @@ "soft_pairs": [ "pull_request:43488|pull_request:43532" ], - "prepared_review_unit_hash": "79d10f7777ddf2a71ddbd7b0ccf843ef815362ae03cacc08dac6d003b56d7eca", + "prepared_review_unit_hash": "c01b27e2d94d5d4b2b724689706b57ff6c1179903409e61d8392f7bb75d6a5d7", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4563, + "serialized_chars": 4562, "estimated_input_tokens": 1141, "estimated_eval_tokens": 2538 }, @@ -19377,7 +19593,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4563, + "serialized_chars": 4562, "estimated_input_tokens": 1141, "estimated_eval_tokens": 2538 }, @@ -19386,18 +19602,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests are not duplicates: one is a repo-bot / packaging test PR centered on setup.py and Vit tests, while the other is a workflow/CI diff PR touching GitHub Actions and testing utilities. The only overlap is an incidental shared test file.", - "confidence": 0.96, + "summary": "These two PRs are not duplicates: one is a small repo-bot/setup/version sanity check, while the other rewires CI/comment workflows and testing utilities. They only overlap on a shared Vit test file, which is not enough to indicate the same underlying change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #43532 is the better representative of the cluster because it contains the broader, concrete code/workflow change set; PR #43488 is a narrow bot-check/style test PR and does not share the same change intent.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "PR #43532 is the most substantial and generalizable artifact here, touching CI and test infrastructure across multiple files. PR #43488 is too specialized and unrelated to serve as the cluster representative.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Different underlying changes and code paths: 43488 modifies setup/versioning and a Vit test, while 43532 rewires CI/workflow files and test infrastructure. Shared test filename alone is not enough to treat them as the same PR." + "reason": "Different goals and code paths: #43488 changes packaging/versioning for a bot/check, while #43532 updates CI/workflow and testing utility behavior. Shared Vit test file alone is not enough to merge them as one PR." } ] }, @@ -19411,7 +19627,7 @@ { "label": "pull_request", "component_index": 10, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19427,12 +19643,12 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "c1456f77295952bcbb531bdfe0d43a239075d1c1861224a882b7b9a98fe45382", + "prepared_review_unit_hash": "279d5dc56023b89d5195bc4accce091ce0af63f160fd24985be800c27bf972d9", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -19440,7 +19656,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6082, + "serialized_chars": 6083, "estimated_input_tokens": 1521, "estimated_eval_tokens": 3298 }, @@ -19449,18 +19665,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two distinct open PRs: one adds a Promptable Concept Segmentation pipeline, the other adds a Promptable Visual Segmentation pipeline. They share some documentation and pipeline scaffolding, but they target different code paths, models, and tests, so they should not be deduplicated.", + "summary": "Two separate PRs add different new segmentation pipelines with similar scaffolding, but they target different tasks/models and are not duplicates.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the two PRs implement different pipeline features rather than the same change.", + "canonical_pr_reason": "PR #43613 is the broader/more representative pipeline addition in this cluster: it adds Promptable Visual Segmentation and wires in support across multiple related model docs/tests, making it the stronger canonical example of the shared pipeline-creation pattern.", "best_issue_reason": null, - "best_pr_reason": "No single best PR for deduplication; 43613 is broader, but it is still a separate feature from 43612.", + "best_pr_reason": "PR #43613 is the best representative of the cluster because it is a complete, self-contained pipeline feature addition with substantial integration work; however, it should not subsume PR #43612 since that PR adds a different pipeline.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different pipelines and underlying model/test coverage (concept segmentation vs visual segmentation). Shared docs/scaffolding are not enough to treat them as the same fix or merge them into one PR." + "reason": "Both PRs follow the same template, but they implement different concrete features: Promptable Concept Segmentation vs Promptable Visual Segmentation. They touch different task pages, pipeline modules, and model/test sets, so they are not the same underlying change and should not be merged as duplicates." } ] }, @@ -19474,7 +19690,7 @@ { "label": "pull_request", "component_index": 11, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19490,40 +19706,40 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "f48c1e7514e33f01db1491ae53841b741d248bb0c5d6662c03c5f34ada6fac2b", + "prepared_review_unit_hash": "3965e6b52e223fce6934c3136e678633f3f28406edd910dde0b4dfc2a1c10641", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4292, - "estimated_input_tokens": 1073, - "estimated_eval_tokens": 2402 + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4292, - "estimated_input_tokens": 1073, - "estimated_eval_tokens": 2402 + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two pull requests both touch `src/transformers/trainer.py` and both add overridable trainer internals, but they fix different behaviors: one addresses gradient-accumulation loss scaling, the other column filtering from model signatures. They are related in theme but not the same underlying change.", - "confidence": 0.96, + "summary": "These two PRs both touch `trainer.py`, but they fix different trainer extension points: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers/PEFT. They are not duplicates and should not be merged together.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: these are distinct trainer enhancements affecting different code paths and should not be treated as duplicates.", + "canonical_pr_reason": "No canonical PR: the changes target different code paths and different customization hooks in `Trainer`, so they are not the same underlying fix.", "best_issue_reason": null, - "best_pr_reason": "PR #43651 is the more substantive and self-contained change, introducing `_loss_is_scaled_for_ga` to control training-step loss scaling for custom trainers.", + "best_pr_reason": "PR #43651 is the more self-contained behavior change and directly addresses a concrete training-step loss-scaling issue; PR #43663 is a separate refactor/extension point for column filtering.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Both are trainer-extension PRs, but they address different problems: gradient accumulation loss scaling vs. signature-based column filtering. Same file and theme are not enough to merge them as duplicates." + "reason": "Different problems and merge targets: #43651 changes loss normalization during training, while #43663 adds a signature-column helper for data filtering. They are not the same bug or change." } ] }, @@ -19537,7 +19753,7 @@ { "label": "pull_request", "component_index": 12, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19553,12 +19769,12 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "8682019cf3c156e074ebed5dfd82d430c5f51ea86c05fbadebc34df0785937b1", + "prepared_review_unit_hash": "524ef028715d04e10b5efcfa22239f8dbb320321e204f10a1ecb11c7e2bd7e60", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3664, + "serialized_chars": 3663, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -19566,7 +19782,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3664, + "serialized_chars": 3663, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -19575,35 +19791,35 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs touch the same file and related typing/import areas, but they do not look like the same underlying bug fix. PR 43656 is a broader annotations/runtime typing cleanup, while PR 43842 is a narrower pydantic-guarded `TypeAdapter` import/type-hint fix; they read more like adjacent follow-ups than one mergeable change.", - "confidence": 0.67, + "summary": "Both PRs target the same import-time NameError in `src/transformers/cli/serve.py` around `TypeAdapter`/type annotations; PR 43842 is the cleaner, more targeted fix, while PR 43656 is a broader annotation-runtime cleanup that appears to address the same bug.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "43842 is the more targeted PR: it directly addresses the `TypeAdapter` NameError path with a minimal, concrete change and an explicit issue link.", + "canonical_pr_reason": "PR 43842 is the best canonical PR because it directly fixes the reported `TypeAdapter` NameError with the smallest, most specific change (guarding the import under `TYPE_CHECKING` and quoting the annotation).", "best_issue_reason": null, - "best_pr_reason": "43842 is the best representative of the cluster because it is the narrower, more specific fix for the CLI `TypeAdapter` failure mode.", + "best_pr_reason": "PR 43842 is the strongest representative of the cluster since it cleanly addresses the concrete CLI runtime error without mixing in broader typing refactors.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", - "accept": false, - "reason": "Related, but not clearly the same bug: 43656 is a broader typing/annotations cleanup, while 43842 specifically guards `TypeAdapter` for missing pydantic. The shared file name alone is not enough to treat them as one mergeable PR." + "accept": true, + "reason": "Same underlying code-path problem in `cli/serve.py`: runtime evaluation of annotations causing a `TypeAdapter`-related NameError. The two PRs are plausible alternative fixes for the same bug and could have been merged into one change." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The pair is related, but the verdict is too strong for the evidence. The packet only shows shared file/name similarity and different explicit issue links, with a soft-similarity pairing and no direct indication they address the exact same bug. PR 43656 appears to be a broader annotations/runtime typing cleanup, while PR 43842 is a narrower pydantic-guarded TypeAdapter fix. It\u2019s safer to treat them as adjacent follow-ups rather than confidently mergeable as one bugfix PR." + "accept": true, + "feedback": "The summary stays within the packet: both PRs touch `src/transformers/cli/serve.py`, both are about annotation/runtime `TypeAdapter` NameError fixes, and PR 43842 is the more targeted change. The soft-edge acceptance is conservative enough because the two diffs look compatible and plausibly mergeable into one bugfix PR." }, "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": false, + "retried": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, { "label": "pull_request", "component_index": 13, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -19669,7 +19885,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -19747,12 +19963,12 @@ "pull_request:44017|pull_request:44071", "pull_request:44073|pull_request:44101", "pull_request:44059|pull_request:44071", - "pull_request:44073|pull_request:44161", "pull_request:44018|pull_request:44059", + "pull_request:44073|pull_request:44161", "pull_request:44056|pull_request:44101", "pull_request:44059|pull_request:44068", "pull_request:44101|pull_request:44161", - "pull_request:43997|pull_request:44017", + "pull_request:44017|pull_request:44018", "pull_request:43997|pull_request:44056", "pull_request:44071|pull_request:44076", "pull_request:44074|pull_request:44101", @@ -19781,7 +19997,7 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "63a601a68db8d2b49688b7193ca4e6c570f18e374c25d68058d31ba034cb0cb1", + "prepared_review_unit_hash": "b6ca91714e984d79dc5a72a2ecd33ecdb505f9e987d29483288b6738db6d1771", "component_budget": { "node_count": 28, "item_count": 28, @@ -19794,339 +20010,369 @@ "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75091, - "estimated_input_tokens": 18773, - "estimated_eval_tokens": 37802 + "serialized_chars": 75088, + "estimated_input_tokens": 18772, + "estimated_eval_tokens": 37800 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of model-specific refactors for the shared output-tracing migration (#43979), not true duplicates across different architectures. The only clear duplicate groups are the GPT-J/CodeGen subgroup and the GPT-Neo subgroup; the rest should stay separate.", - "confidence": 0.71, + "summary": "Mostly a model-by-model output-tracing refactor sweep, not a single duplicate bug. The only clear duplicate groups are the GPT-Neo pair and the GPT-J/CodeGen family; the rest are parallel refactors in different subsystems.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "44722 is the strongest representative of the GPT-J/CodeGen output-tracing subgroup: it directly implements the standardized decorator path, overlaps heavily with the GPT-J-only follow-up, and appears to be the most consolidated version of that concrete change.", + "canonical_pr_reason": "PR 44066 is the strongest representative: it explicitly targets the shared output-tracing refactor and covers the central GPT-J/CodeGen code path that later follow-up PRs overlap with.", "best_issue_reason": null, - "best_pr_reason": "44722 is the best single PR in this cluster because it cleanly captures one concrete code-path refactor with strong overlap to other near-duplicate GPT-J/CodeGen patches, while the rest of the cluster spans different model families and should not be merged together.", + "best_pr_reason": "PR 44066 is the best single PR to represent this cluster because it is the broadest concrete implementation in the repeated GPT-J/CodeGen tracing-refactor family and has the clearest issue linkage.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families (FNet/CVT vs GPT-J); same tracking issue and similar refactor pattern are not enough to make this the same concrete change." + "reason": "Different model/code path; same issue theme is not enough to make these duplicates." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Unrelated model implementations (FNet/CVT vs DeBERTa) that happen to use the same output-tracing migration pattern." + "reason": "Both are output-tracing refactors, but they touch unrelated model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same code path, with the same decorator/capture_outputs migration and overlapping file changes." + "reason": "Same GPT-J output-tracing refactor family, with overlapping gptj/modeling_gptj.py changes." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "ResNet/RegNet/RT-DETR refactor versus EfficientNet refactor; same broad migration theme, but not the same concrete bug/change." + "reason": "ResNet and EfficientNet are separate code paths; these are parallel refactors, not the same fix." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet and GPT-J/CodeGen are unrelated model-specific migrations." + "reason": "Different models and different tracing code paths; shared theme alone is insufficient." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen versus EfficientNet; only the shared issue target is common." + "reason": "GPT-J/CodeGen vs EfficientNet are unrelated subsystems." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are unrelated implementations; this is only a shared refactor style." + "reason": "MobileNetV2 and DeBERTa are unrelated implementations." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen output tracing versus MGP-STR output tracing; different model code paths." + "reason": "Different model families and different concrete code paths." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen and MPT are separate model implementations, so these are not the same underlying fix." + "reason": "GPT-J/CodeGen output tracing and MPT output tracing are separate model-specific changes." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both PRs refactor GPT-Neo output tracing on the same model/code path and are plausible duplicates/iterations of the same change." + "reason": "Both are GPT-Neo output-tracing refactors in the same code path." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo is a different model than GPT-J/CodeGen; shared output-tracing infrastructure alone is insufficient." + "reason": "GPT-Neo is a different model/code path from GPT-J/CodeGen." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen versus GPT-Neo; same infrastructure goal, different concrete model change." + "reason": "GPT-J/CodeGen and GPT-Neo are distinct implementations." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet refactor versus GPT-J/CodeGen refactor; unrelated code paths." + "reason": "ResNet and GPT-J/CodeGen are unrelated subsystems." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are different models with different forward paths." + "reason": "Different model code paths; both mention output tracing but do not fix the same bug." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet versus GPT-J; not the same bug/change." + "reason": "ResNet and GPT-J are different concrete implementations." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are unrelated model migrations." + "reason": "GPT-Neo and ResNet are unrelated; shared issue target does not imply duplication." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are distinct code paths." + "reason": "Different model families and no shared code path." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo versus Flaubert; same migration theme, different concrete model behavior." + "reason": "GPT-Neo and Flaubert/XLM are unrelated concrete changes." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are unrelated implementations." + "reason": "Different model families and different output paths." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo versus GPT-J/CodeGen; they share a migration approach but not the same underlying change." + "reason": "GPT-Neo and GPT-J/CodeGen are separate model implementations." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are different model-specific refactors." + "reason": "MPT and GPT-J/CodeGen are unrelated code paths." }, { - "left": "pull_request:44018", - "right": "pull_request:44085", + "left": "pull_request:44059", + "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus GPT-J; different model code paths." + "reason": "GPT2 and MGP-STR are different model families." }, { - "left": "pull_request:44018", + "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo versus MGP-STR; not the same concrete fix." + "reason": "GPT-J and MGP-STR are unrelated implementations." }, { - "left": "pull_request:44068", - "right": "pull_request:44086", + "left": "pull_request:44098", + "right": "pull_request:44101", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated model implementations." + "reason": "ViLT and Flaubert/XLM are different model-specific output changes." }, { "left": "pull_request:44059", - "right": "pull_request:44086", + "right": "pull_request:44722", "accept": false, - "reason": "GPT-2 versus MGP-STR; same refactor style, different code paths." + "reason": "GPT2 and GPT-J/CodeGen are distinct code paths." }, { - "left": "pull_request:44085", - "right": "pull_request:44086", + "left": "pull_request:44017", + "right": "pull_request:44722", "accept": false, - "reason": "GPT-J versus MGP-STR; different underlying behavior." + "reason": "SegFormer and GPT-J/CodeGen are unrelated subsystems." }, { - "left": "pull_request:44098", - "right": "pull_request:44101", + "left": "pull_request:44017", + "right": "pull_request:44161", "accept": false, - "reason": "ViLT and Flaubert/XLM are unrelated models." + "reason": "SegFormer and LongT5 are separate model implementations." }, { - "left": "pull_request:44019", - "right": "pull_request:44722", + "left": "pull_request:44017", + "right": "pull_request:44073", "accept": false, - "reason": "ResNet versus GPT-J/CodeGen; not the same concrete bug/change." + "reason": "SegFormer and VisualBert are unrelated code paths." }, { - "left": "pull_request:44059", + "left": "pull_request:44068", "right": "pull_request:44071", "accept": false, - "reason": "GPT-2 and MPT are different model code paths." + "reason": "GPT-Neo and MPT are different model families." }, { - "left": "pull_request:44073", - "right": "pull_request:44101", + "left": "pull_request:44071", + "right": "pull_request:44086", "accept": false, - "reason": "VisualBERT and Flaubert/XLM are different model implementations." + "reason": "MPT and MGP-STR are unrelated fixes." }, { - "left": "pull_request:44017", - "right": "pull_request:44722", + "left": "pull_request:43997", + "right": "pull_request:44073", "accept": false, - "reason": "SegFormer refactor is unrelated to GPT-J/CodeGen." + "reason": "RegNet and VisualBert are different model implementations." }, { - "left": "pull_request:44086", - "right": "pull_request:44722", + "left": "pull_request:44056", + "right": "pull_request:44101", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are not the same underlying change." + "reason": "MPNet and Flaubert/XLM are unrelated code paths." }, { - "left": "pull_request:44019", - "right": "pull_request:44085", + "left": "pull_request:44059", + "right": "pull_request:44068", "accept": false, - "reason": "ResNet and GPT-J are different model paths." + "reason": "GPT2 and GPT-Neo are different model-specific output tracing changes." }, { - "left": "pull_request:44017", + "left": "pull_request:44101", "right": "pull_request:44161", "accept": false, - "reason": "SegFormer and LongT5 are unrelated; only the migration pattern is shared." + "reason": "Flaubert/XLM and LongT5 are separate subsystems." }, { - "left": "pull_request:44073", - "right": "pull_request:44161", + "left": "pull_request:44017", + "right": "pull_request:44018", + "accept": false, + "reason": "SegFormer and GPT-Neo are unrelated models." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are separate model families." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", "accept": false, - "reason": "VisualBERT and LongT5 are different model families." + "reason": "SqueezeBert and MPT are unrelated concrete code paths." }, { "left": "pull_request:44074", "right": "pull_request:44101", "accept": false, - "reason": "TextNet and Flaubert are unrelated refactors." + "reason": "TextNet and Flaubert/XLM are different model implementations." }, { "left": "pull_request:43997", - "right": "pull_request:44073", + "right": "pull_request:44018", "accept": false, - "reason": "RegNet/ResNet-family output tracing versus VisualBERT; different code paths." + "reason": "RegNet and GPT-Neo are unrelated changes." }, { - "left": "pull_request:44076", - "right": "pull_request:44101", + "left": "pull_request:44056", + "right": "pull_request:44161", "accept": false, - "reason": "ImageGPT and Flaubert/XLM are unrelated model changes." + "reason": "MPNet and LongT5 are different model-specific refactors." }, { - "left": "pull_request:44101", + "left": "pull_request:43997", "right": "pull_request:44161", "accept": false, - "reason": "Flaubert/XLM and LongT5 are different model implementations." + "reason": "RegNet and LongT5 are unrelated code paths." }, { - "left": "pull_request:43997", - "right": "pull_request:44017", + "left": "pull_request:44074", + "right": "pull_request:44161", "accept": false, - "reason": "RegNet/ResNet-family refactor versus SegFormer; not the same concrete change." + "reason": "TextNet and LongT5 are separate models." }, { - "left": "pull_request:43997", - "right": "pull_request:44056", + "left": "pull_request:44018", + "right": "pull_request:44161", "accept": false, - "reason": "RegNet/ResNet-family refactor versus MPNet; same migration theme, different code paths." + "reason": "GPT-Neo and LongT5 are unrelated." }, { - "left": "pull_request:44010", + "left": "pull_request:43997", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBERT and MPT are unrelated model implementations." + "reason": "RegNet and MPT are different model families." }, { "left": "pull_request:44010", - "right": "pull_request:44018", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are unrelated implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are separate model-specific changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "FNet and LongT5 are unrelated code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", "accept": false, - "reason": "SqueezeBERT and GPT-Neo are different concrete code paths." + "reason": "FNet and SegFormer are different model implementations." }, { "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 changes." + "reason": "DeBERTa v2 and TextNet are separate fixes." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Same GPT-J/CodeGen output-tracing migration with strong file and implementation overlap; 44722 looks like the consolidated version." + "reason": "Same GPT-J/CodeGen output-tracing refactor, with overlapping model file changes." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both touch the GPT-J output-tracing code path and implement the same decorator-based refactor, making them plausible duplicates/iterations." + "reason": "Both target the same GPT-J output-tracing code path and would plausibly merge into one PR." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J versus SpeechT5; no shared concrete code-path problem." + "reason": "GPT-J and SpeechT5 are unrelated model-specific changes." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "FNet/CVT and ResNet/RegNet/RT-DETR are different model families." + "reason": "FNet and ResNet are different concrete code paths." }, { "left": "pull_request:43996", "right": "pull_request:44129", "accept": false, - "reason": "FNet/CVT and SpeechT5 are unrelated model-specific refactors." + "reason": "FNet and SpeechT5 are unrelated models." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and SpeechT5 are different code paths." + "reason": "DeBERTa v2 and SpeechT5 are separate model implementations." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and VITS are unrelated model implementations." + "reason": "DeBERTa v2 and ViTS are unrelated fixes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays grounded in the packet and remains conservative about cross-model pairs. The only accepted soft-edge groups are the ones that plausibly look like same-model iterations (GPT-J/CodeGen and GPT-Neo), and the rejected pairs are mostly justified by distinct model families or code paths. Canonical PR choice 44722 is also supported by the evidence as the most consolidated GPT-J/CodeGen representative." + "feedback": "The summary is grounded in the packet: the accepted soft edges stay within the same concrete GPT-Neo or GPT-J/CodeGen code paths, and the rejected edges correctly avoid conflating unrelated model implementations that only share the high-level output-tracing theme. The chosen representative PR (44066) is also defensible from the evidence because it is issue-linked and central to the GPT-J/CodeGen family." }, "evaluator_used": true, "retried": false, @@ -20137,7 +20383,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 32, + "component_count": 33, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -20164,7 +20410,7 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "5ec88645481037ed7f5715ceb8c61bc7f6d8d66550fd825a32e48308aae79900", + "prepared_review_unit_hash": "4e9450537d6878aec8f5053c6bc2749a287e08f8ae519e3db67976210ab49ed1", "component_budget": { "node_count": 28, "item_count": 28, @@ -20177,57 +20423,57 @@ "node_count": 8, "item_count": 8, "soft_pair_count": 6, - "serialized_chars": 17124, - "estimated_input_tokens": 4281, - "estimated_eval_tokens": 8818 + "serialized_chars": 17127, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs all follow the same general theme\u2014migrating model outputs to a standardized capture/output-tracing pattern\u2014but they are model-specific refactors, not duplicate fixes for the same underlying bug or change. I would not collapse any of the soft edges.", - "confidence": 0.95, + "summary": "This cluster is a set of model-specific refactors to standardized output tracing/capture_outputs across different architectures. They are related by theme, but they are not duplicates of the same underlying bug or change because each PR touches a different model code path.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR stands out as a duplicate representative: each PR targets a different model/code path, so they are separate changes rather than one shared fix.", + "canonical_pr_reason": "PR 44129 is the strongest representative: it has an explicit issue target, a clear refactor title, and a more substantive diff showing the standardized output-capture pattern applied to SpeechT5.", "best_issue_reason": null, - "best_pr_reason": "PR 44129 is the best standalone representative of the theme because it has an explicit issue target and a focused, model-specific implementation, but it still is not a duplicate of the others.", + "best_pr_reason": "PR 44129 is the best standalone PR in the set because it is the most clearly scoped and explicitly linked to an issue, with a concrete model-specific implementation change rather than only a generic decorator migration.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Both are output-tracing refactors, but for different models (FocalNet vs ImageGPT) and different code paths; not the same underlying change." + "reason": "Both are standardized-output refactors, but they affect different models (FocalNet vs ImageGPT) and different forward paths; this is a shared pattern, not the same change." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Same refactor pattern, but VisualBert and TextNet are separate model implementations with distinct forward paths." + "reason": "VisualBert and TextNet are separate model implementations with different output-tracing code paths, so these are not the same concrete fix." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "Both modernize output handling, but MPNet and VisualBert are different model-specific refactors, not one fix." + "reason": "MPNet and VisualBert are unrelated model-specific refactors; similar decorator usage does not make them duplicates." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "Different models and different implementation details; they share a theme but not the same concrete code-path change." + "reason": "These PRs touch different models and different tracing logic, so they are not mergeable as one duplicate change." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are separate model refactors; the shared output-tracing motif is too broad to treat as duplicate." + "reason": "RegNet and SqueezeBert are distinct architectures; the PRs share a migration theme but not the same underlying bug or code path." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 are different model code paths; SpeechT5 also has an explicit issue target, indicating a separate task rather than a duplicate." + "reason": "SqueezeBert and SpeechT5 are separate model refactors; the shared output-capture pattern is too broad to treat as one duplicate PR." } ] }, @@ -20241,7 +20487,7 @@ { "label": "pull_request", "component_index": 15, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20306,274 +20552,268 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "26e513d78dadfa31d9f522b1890a39c9863d1506911007fa15a0fb55a87457f5", + "prepared_review_unit_hash": "e22d35c90497aa6473076010e9c989e6d8a6e46eac7b99bc5152c4c477dbecc0", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 39, - "serialized_chars": 39780, - "estimated_input_tokens": 9945, - "estimated_eval_tokens": 20146 + "serialized_chars": 39774, + "estimated_input_tokens": 9944, + "estimated_eval_tokens": 20144 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 39, - "serialized_chars": 39780, - "estimated_input_tokens": 9945, - "estimated_eval_tokens": 20146 + "serialized_chars": 39774, + "estimated_input_tokens": 9944, + "estimated_eval_tokens": 20144 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a set of model-specific output-tracing refactors, not duplicates of one concrete bug or feature change. The PRs mostly share a migration pattern (`can_return_tuple`, output capturing, signature cleanup), but they target different model implementations and code paths.", - "confidence": 0.94, + "summary": "This cluster is a set of similarly named but distinct model-specific output-tracing refactors, not duplicate bugfixes or one mergeable PR.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44003 is the best representative of the cluster theme: it applies the output-tracing refactor within a closely related model family and introduces the recording/capture plumbing, making it more central than the smaller one-file variants.", + "canonical_pr_reason": "No single canonical duplicate: each PR edits a different model path. If one representative is needed, 44025 is the broadest example because it spans two related files, but it does not subsume the others.", "best_issue_reason": null, - "best_pr_reason": "PR 44003 is a strong standalone exemplar of this refactor pattern because it is self-contained, concrete, and spans two tightly related model files without being just a trivial signature tweak.", + "best_pr_reason": "44025 is the best anchor for the cluster because it is the most representative multi-file output-tracing refactor, but it is still model-specific rather than a duplicate of the rest.", "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 only, not the same underlying change." + "reason": "Different model code paths (`timm_backbone` vs `upernet`); same refactor theme only." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different models and outputs (`timm_backbone` vs `vision_text_dual_encoder`); too broad a similarity to treat as one PR." + "reason": "Different models (`timm_backbone` vs `vision_text_dual_encoder`); not the same underlying change." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "`timm_backbone` and `univnet` are unrelated implementations; they share only the generic output-tracing migration pattern." + "reason": "Separate model implementations (`timm_backbone` vs `univnet`); too broad to treat as duplicates." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Different families (`timm_backbone` vs `mamba/falcon_mamba`); not the same concrete bug or fix." - }, - { - "left": "pull_request:43998", - "right": "pull_request:44004", - "accept": false, - "reason": "`timm_backbone` and `codegen` touch unrelated forward paths; similar refactor intent is not enough." + "reason": "`timm_backbone` and `falcon_mamba/mamba` are unrelated forward paths; same pattern, different change." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "`timm_backbone` vs `rwkv` are separate model implementations with different output plumbing." + "reason": "Different architectures (`timm_backbone` vs `rwkv`); not one mergeable fix." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "`mobilenet_v1` and `univnet` are unrelated model paths; the overlap is only the output-tracing refactor pattern." + "reason": "`mobilenet_v1` and `univnet` are distinct model paths; same output-tracing refactor theme only." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different model subsystems (`mobilenet_v1` vs `upernet`); not one shared code-path bug." + "reason": "Different files and models (`mobilenet_v1` vs `upernet`); not the same concrete bug/change." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "`mobilenet_v1` and `depth_anything` are separate models; no evidence of a shared concrete defect." + "reason": "`mobilenet_v1` and `depth_anything` refactor different model forwards; too different to merge as duplicates." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different encoder-decoder and CNN model code paths; same migration theme only." + "reason": "`mobilenet_v1` vs `vision_encoder_decoder`; same subsystem theme, different code path." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "`mobilenet_v1` and `speech_encoder_decoder` are unrelated implementations; not mergeable as one fix." + "reason": "`mobilenet_v1` and `speech_encoder_decoder` are separate model wrappers; not one underlying change." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "`vision_text_dual_encoder` and `univnet` are different model families; similarity is limited to output tracing cleanup." + "reason": "Different model families (`vision_text_dual_encoder` vs `univnet`); same refactor pattern, not duplicate." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different model code paths (`vision_text_dual_encoder` vs `upernet`); no shared underlying bug." + "reason": "`vision_text_dual_encoder` and `upernet` touch different forward paths; no shared concrete fix." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "These touch unrelated forward implementations; the commonality is only the refactor pattern." + "reason": "`vision_text_dual_encoder` vs `falcon_mamba/mamba`; unrelated model implementations." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "`vision_text_dual_encoder` and `depth_anything` are separate model stacks; not the same concrete change." + "reason": "`vision_text_dual_encoder` and `depth_anything` are different model-specific refactors; not mergeable as one PR." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "`univnet` and `upernet` are different domains and code paths; no duplicate bug evidence." + "reason": "`univnet` and `upernet` are different code paths; same output-tracing theme only." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different model families (`univnet` vs `mamba/falcon_mamba`); too generic to merge." + "reason": "Different model implementations (`univnet` vs `falcon_mamba/mamba`); not the same bug/change." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "`univnet` and `depth_anything` are unrelated implementations; same cleanup motif only." + "reason": "`univnet` and `depth_anything` refactor distinct forwards; too broad to combine." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "`upernet` and `rwkv` are separate model code paths; no evidence they fix the same issue." + "reason": "`upernet` and `rwkv` are unrelated model paths; similar refactor, not duplicate." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different model families and APIs; they share only a broad refactor intent." + "reason": "`upernet` vs `falcon_mamba/mamba`; separate model-specific changes." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "`mamba/falcon_mamba` and `codegen` are unrelated model implementations; not the same fix." + "reason": "`falcon_mamba/mamba` and `codegen` target different code paths; not the same concrete fix." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Different code paths and model types (`timm_backbone` vs `codegen`); only the output-tracing migration is shared." + "reason": "`timm_backbone` vs `codegen`; same refactor motif but unrelated implementations." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "`univnet` and `codegen` are unrelated models; not a single underlying bug/change." + "reason": "`univnet` and `codegen` are unrelated forward paths; not one underlying change." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "`upernet` and `codegen` differ in both architecture and output handling; no mergeable duplicate here." + "reason": "`upernet` and `codegen` touch different model internals; not a duplicate pair." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "`vision_text_dual_encoder` and `codegen` are different code paths; same refactor theme is insufficient." + "reason": "`vision_text_dual_encoder` vs `codegen`; similar output-tracing refactor, different model change." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "`depth_anything` and `vision_encoder_decoder` are unrelated model families; not the same concrete bug or change." + "reason": "Different wrappers (`depth_anything` vs `vision_encoder_decoder`); not the same code-path problem." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Different target models (`depth_anything` vs `speech_encoder_decoder`); no shared code-path fix." + "reason": "`depth_anything` and `speech_encoder_decoder` are separate model-specific refactors." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "`depth_anything` and `superpoint` are separate implementations; the overlap is only generic output tracing." + "reason": "`depth_anything` vs `superpoint`; same theme, different concrete model path." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "`depth_anything` and `rwkv` are unrelated model families; not the same underlying issue." + "reason": "`depth_anything` and `rwkv` are unrelated implementations; not mergeable as one PR." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "`depth_anything` and `dpr` affect different model stacks; no evidence of one bug." + "reason": "`depth_anything` vs `dpr`; different model code paths and scope." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44027", "accept": false, - "reason": "`vision_encoder_decoder` and `rwkv` are not the same code path; this is only a shared refactor style." + "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are different wrappers; similar refactor only." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44028", "accept": false, - "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are different wrappers with different forward behavior; not a duplicate." + "reason": "`vision_encoder_decoder` vs `superpoint`; not the same underlying bug/change." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44029", "accept": false, - "reason": "`vision_encoder_decoder` and `superpoint` are unrelated models; no shared concrete fix." + "reason": "`vision_encoder_decoder` and `rwkv` are unrelated model paths." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different model families (`vision_encoder_decoder` vs `dpr`); same theme, not same bug." + "reason": "`vision_encoder_decoder` vs `dpr`; distinct model-specific changes." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "`speech_encoder_decoder` and `superpoint` are separate implementations with distinct output paths." + "reason": "`speech_encoder_decoder` and `superpoint` are separate implementations; not a duplicate fix." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "`speech_encoder_decoder` and `rwkv` are unrelated code paths; not mergeable as one fix." + "reason": "`speech_encoder_decoder` vs `rwkv`; same refactor pattern, different concrete change." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Different models (`speech_encoder_decoder` vs `dpr`); no evidence of the same underlying defect." + "reason": "`speech_encoder_decoder` and `dpr` are unrelated model paths." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "`superpoint` and `rwkv` are unrelated model implementations; similarity is only superficial." + "reason": "`superpoint` and `rwkv` are different code paths; too dissimilar for a duplicate." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "`superpoint` and `dpr` target different code paths and outputs; not the same concrete change." + "reason": "`superpoint` vs `dpr`; distinct model-specific refactors." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "`rwkv` and `dpr` are distinct models; no shared bug or jointly mergeable PR." + "reason": "`rwkv` and `dpr` are unrelated implementations; not one mergeable PR." } ] }, @@ -20587,7 +20827,7 @@ { "label": "pull_request", "component_index": 16, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20603,12 +20843,12 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "984a57d985df5ddca7592da89d5bbe7ce863180c52fe417ebe77962dad9c05f8", + "prepared_review_unit_hash": "2c7fcb183c37c13bdd48e87f62b3e17b1ca8c00b65f28318eed786226d496963", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6551, + "serialized_chars": 6552, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -20616,7 +20856,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6551, + "serialized_chars": 6552, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -20625,18 +20865,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests are model add-ons, but they implement different models with different code paths: PenguinVL vs Voxtral TTS. The overlap is mostly standard Transformers onboarding boilerplate (docs, auto registries), so they should not be clustered as duplicates.", - "confidence": 0.97, + "summary": "These PRs are both new model integrations, but for different models and different code paths: PenguinVL vs Voxtral TTS. The overlap is only in generic Transformer scaffolding (docs, auto registries, model init), so they are not duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45401 is the cleaner representative because it is explicitly tied to issue 45027 and is scoped to a single concrete model integration.", + "canonical_pr_reason": "No canonical PR: 44662 adds PenguinVL while 45401 adds Voxtral TTS, so they do not represent the same underlying change.", "best_issue_reason": null, - "best_pr_reason": "PR 45401 is the strongest standalone candidate: it has an explicit issue target and a self-contained Voxtral TTS implementation.", + "best_pr_reason": "No single best PR for duplication purposes; both are standalone model-addition PRs and neither subsumes the other.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different models and different functionality: PenguinVL image/vision support vs Voxtral TTS text-to-speech. Shared registry/doc changes are boilerplate and do not indicate the same underlying change." + "reason": "Different target models and implementations; shared files are broad registry/docs updates, not the same concrete bug or change." } ] }, @@ -20650,7 +20890,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20670,12 +20910,12 @@ "pull_request:45363|pull_request:46339", "pull_request:45666|pull_request:46339" ], - "prepared_review_unit_hash": "0eb83fd3424fbfd299ecbd60a0bf2db46eef355c91fb25f37fe077a80068f0a2", + "prepared_review_unit_hash": "b817b71c7caaaa5f0f85555ee5731d2d1e7279c2e02596b1b2b651a32ed6beb0", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 9636, + "serialized_chars": 9633, "estimated_input_tokens": 2409, "estimated_eval_tokens": 5074 }, @@ -20683,7 +20923,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 9636, + "serialized_chars": 9633, "estimated_input_tokens": 2409, "estimated_eval_tokens": 5074 }, @@ -20692,47 +20932,44 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three closely related PRs evolving the same kernel-fusion feature, plus one earlier foundational API PR. The latest PR is the best representative for the n-to-1 KernelConfig fusion line, while the standalone Module Fusion API PR is related but not the same change.", + "summary": "All items are related to the same kernel/module fusion area, but they look like successive iterations rather than duplicate PRs. PR 44979 introduces the base Module Fusion API, 45363 adds n-to-1 fusion via KernelConfig, 45666 extends that path, and 46339 further extends/simplifies the same feature set with additional loading/refactor changes.", "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46339 is the most complete and latest iteration of the n-to-1 KernelConfig fusion work, and it touches the same core files as the earlier fusion PRs while presenting the most finalized scope.", + "canonical_pr_reason": "PR 46339 is the broadest and most recent implementation in this line, touching the core loading and hub-kernel code paths with follow-up simplification/refactoring on top of the KernelConfig fusion work.", "best_issue_reason": null, - "best_pr_reason": "#46339 is the best single PR to represent this cluster because it is the newest, broadest refinement of the same kernel-fusion feature and subsumes the earlier KernelConfig-based fusion work.", + "best_pr_reason": "PR 46339 best represents the cluster because it is the latest, most complete kernel-fusion iteration and subsumes the earlier n-to-1 fusion direction more than the others.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Related feature area, but #44979 is a general module-fusion API while #45363 is a specific n-to-1 KernelConfig implementation; they are not the same underlying change and would not merge cleanly as one PR." + "reason": "Related feature area, but 44979 is the foundational Module Fusion API while 45363 is a separate n-to-1 KernelConfig implementation; they are not the same concrete change." }, { "left": "pull_request:45363", "right": "pull_request:46339", - "accept": true, - "reason": "Both PRs implement the same concrete n-to-1 KernelConfig kernel-fusion path and share the same core loading/integration files; #46339 looks like a later refinement of the same change set." + "accept": false, + "reason": "46339 appears to be a later extension/simplification of the same fusion machinery, not the same single PR-level change; the shared files alone are not enough to treat them as duplicates." }, { "left": "pull_request:45666", "right": "pull_request:46339", - "accept": true, - "reason": "These are clearly iterations of the same kernel-fusion feature, with overlapping filenames and nearly identical scope around n-to-1 fusion via KernelConfig; they could plausibly be consolidated into one PR." + "accept": false, + "reason": "Both work on n-to-1 KernelConfig fusion, but 46339 is a follow-up refactor/extension with additional loading changes, so they are consecutive iterations rather than one duplicate fix." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "Grounded overall. The summary stays within the packet evidence, and the soft-edge decisions are conservative enough: #44979 is reasonably kept separate from the KernelConfig fusion line, while #45363/#45666/#46339 share the same n-to-1 fusion feature area, overlapping files, and sequential titles that support treating them as iterations of one change set." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, { "label": "pull_request", "component_index": 18, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20748,7 +20985,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "b7147f8b13daa6087db61daacf5da094e7301dfbde30ba6aa19502086701ca6c", + "prepared_review_unit_hash": "e0939d904095dbc9451ca32b12fa6e427575b43a194b432b70c79a1b6dd3805c", "component_budget": { "node_count": 2, "item_count": 2, @@ -20770,18 +21007,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs fix the same Mistral regex patch AttributeError in `src/transformers/tokenization_utils_tokenizers.py`; the second PR is a superset with added tests and the same explicit issue target.", - "confidence": 0.97, + "summary": "Both pull requests fix the same AttributeError in `_patch_mistral_regex` by switching from `tokenizer.backend_tokenizer.pre_tokenizer` to `tokenizer.pre_tokenizer`; 45317 is the more complete version because it includes tests.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45317 is the better canonical PR: it addresses the same concrete `_patch_mistral_regex` AttributeError as #45086, but is more complete because it also adds test coverage and reflects the later, more polished fix.", + "canonical_pr_reason": "PR 45317 is the better canonical PR because it contains the same core fix as 45086 plus test coverage, making it the more complete representative of this change.", "best_issue_reason": null, - "best_pr_reason": "PR #45317 is the strongest representative of the cluster because it fixes the same bug as #45086 while including tests and the same code-path change, making it the most complete version of the fix.", + "best_pr_reason": "PR 45317 is the strongest overall PR candidate because it directly addresses the bug, includes validation in tests, and appears to subsume the earlier simpler patch.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying bug in `_patch_mistral_regex`/Mistral regex patching, same changed code path, and the second PR is essentially the same fix with tests added, so they plausibly could be merged into one PR." + "reason": "Same concrete bug fix in the same code path, with identical key edits in `tokenization_utils_tokenizers.py`; 45317 simply adds test coverage, so these are duplicate/overlapping implementations of the same change." } ] }, @@ -20795,7 +21032,7 @@ { "label": "pull_request", "component_index": 19, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20811,12 +21048,12 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "0e0bc715c84611ff7db5987a1a4cca186bb336686ea02e4fdb12c1abbd742ea7", + "prepared_review_unit_hash": "b457cb0bb457a684fea8123e943e10cf7cdd77edeabbeb9968695466a06e42e8", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6065, + "serialized_chars": 6066, "estimated_input_tokens": 1517, "estimated_eval_tokens": 3290 }, @@ -20824,7 +21061,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6065, + "serialized_chars": 6066, "estimated_input_tokens": 1517, "estimated_eval_tokens": 3290 }, @@ -20833,18 +21070,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs overlap on NemotronH core files and auto-model wiring, but they are not the same change: #45115 is a narrow refactor/conversion-mapping update, while #45569 is a broader architecture split introducing dedicated NemotronHDense/NemotronHSparse models, docs, and tests.", - "confidence": 0.87, + "summary": "These PRs are related NemotronH architecture work, but they target different changes: one is a conversion-mapping/refactor PR for inheriting GraniteMoeHybrid-compatible weights, while the other is a broader split into NemotronH dense/sparse variants with docs and auto-mapping updates. They are not the same fix and should not be merged as duplicates.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "#45569 is the better canonical PR for the cluster because it is the broader, more complete NemotronH-family change and includes the new dense/sparse model split plus supporting docs/tests; #45115 is a narrower precursor-style refactor.", + "canonical_pr_reason": "PR 45569 is the better representative of the cluster because it is the broader, more complete NemotronH rework and includes the model split plus supporting documentation and registration changes.", "best_issue_reason": null, - "best_pr_reason": "#45569 is the best representative PR because it covers the larger, final user-visible implementation and supersedes the narrower refactor scope in #45115.", + "best_pr_reason": "PR 45569 best captures the main final direction of the NemotronH changes; PR 45115 is a narrower conversion-mapping refactor that overlaps only partially on shared model files.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "Not the same concrete change: #45115 focuses on conversion mapping and inheritance/refactor for NemotronH, while #45569 introduces dedicated NemotronHDense/NemotronHSparse modules, docs, and tests. Shared files are only the common integration points, not evidence of a mergeable single fix." + "reason": "They share NemotronH files, but the concrete changes differ: 45115 focuses on checkpoint conversion and GraniteMoeHybrid-compatible renaming, while 45569 introduces dense/sparse model variants, docs, and auto/model registration updates. Same subsystem, not the same underlying change." } ] }, @@ -20858,7 +21095,7 @@ { "label": "pull_request", "component_index": 20, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20874,40 +21111,40 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "f7028965b892fdd92b632db9070a269811d51312a012379d6b3f5026f7bc685c", + "prepared_review_unit_hash": "67b4f5d135617a858ff7d516494b441bc7a55ca90a373be98f82dc878aef120d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5464, - "estimated_input_tokens": 1366, - "estimated_eval_tokens": 2988 + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5464, - "estimated_input_tokens": 1366, - "estimated_eval_tokens": 2988 + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related but distinct PRs: one adds per-layer heterogeneous config support, the other adds broader heterogeneous model/runtime support that builds on it. They overlap in shared config helpers, but they are not the same concrete change.", - "confidence": 0.84, + "summary": "Two open PRs in the same heterogeneity feature area: one is a broad implementation of heterogeneous model support, while the other is a narrower per-layer configuration PR. They overlap in config scaffolding but do not look like duplicate fixes/changes.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45332 is the broader and more complete change, covering heterogeneous model support across cache/modeling/masking code paths in addition to config plumbing.", + "canonical_pr_reason": "PR #45332 is the broader umbrella change: it adds heterogeneous model support across config, modeling, cache handling, and tests. It subsumes the per-layer config groundwork and is the better representative of the cluster.", "best_issue_reason": null, - "best_pr_reason": "PR #45332 is the best standalone candidate because it subsumes the config work and extends it into the actual model/runtime behavior, making it the more representative end-state change.", + "best_pr_reason": "PR #45332 is the most complete and central implementation in this cluster, so it is the best candidate to keep as the primary PR for the feature.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "#45333 is a narrower config-only prerequisite/subset, while #45332 is a broader heterogeneous model-support PR that happens to include similar config plumbing. They do not look like the same underlying change and should not be merged as duplicates." + "reason": "Related feature work, but not the same concrete change. #45333 is configuration-only groundwork, while #45332 extends into modeling and cache behavior; they could coexist as layered PRs rather than being merged as duplicates." } ] }, @@ -20921,7 +21158,7 @@ { "label": "pull_request", "component_index": 21, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -20937,7 +21174,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "a9a8d2b64d3c58da6af22cc8ede0fa69b396aa0709ad15dfe5cb3b02c4eddd5b", + "prepared_review_unit_hash": "c56d93992db94261afe194e1b37407cc1516061a53ec316522767714e9cbf8bf", "component_budget": { "node_count": 2, "item_count": 2, @@ -20959,18 +21196,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests are type-checking related, but they address different changes: one is a broad mypy/typing cleanup sweep, the other adds specific typing protocols and helpers for runtime code paths. They are not the same underlying fix and should not be merged as duplicates.", - "confidence": 0.84, + "summary": "These are both typing-related PRs, but they address different changes: one is a broad type-checking sweep across many files, while the other adds typing helper protocols and targeted fixes around `modeling_utils`/flash-attention integration.", + "confidence": 0.72, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45425 is the better canonical PR because it makes a focused, concrete typing-infrastructure change in central modules, whereas #45415 is a broad type-checking sweep across many files with mostly cleanup/configuration work.", + "canonical_pr_reason": "PR 45425 is the better representative because it is a focused typing-infrastructure change with a clearer, narrower scope. PR 45415 is a broad repo-wide typing pass that bundles many unrelated file edits.", "best_issue_reason": null, - "best_pr_reason": "PR #45425 is the best representative of the cluster because it targets a specific code-path typing problem and introduces reusable typing abstractions; #45415 is broader and more mechanical.", + "best_pr_reason": "PR 45425 is the most suitable single representative for the cluster because its change set is more concrete and self-contained, whereas PR 45415 is a wide typing sweep with many touched files and mixed cleanup work.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "They overlap on type-checking themes and some files, but #45415 is a wide typing cleanup across many modules while #45425 adds specific typing protocols and annotations for modeling_utils/related code paths. Different scope and not the same concrete change." + "reason": "Both are about typing, but they do not appear to fix the same concrete bug or implement the same change. 45415 is a broad type-checking enablement across many modules; 45425 adds specific typing protocols and related fixes in a smaller area. They would not plausibly be merged as one PR." } ] }, @@ -20984,7 +21221,7 @@ { "label": "pull_request", "component_index": 22, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21000,7 +21237,7 @@ "soft_pairs": [ "pull_request:45513|pull_request:46084" ], - "prepared_review_unit_hash": "723012905e0c02dbaf3acd405edf64db0c9755d417f87565adf4817b8c7b928d", + "prepared_review_unit_hash": "baf48fe75836276b9d91b2731ff9080f9950162981d3a1fee2fc960d2d071a62", "component_budget": { "node_count": 2, "item_count": 2, @@ -21022,18 +21259,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs, but they address different cached-forward bugs in different model families: one for Qwen3.5/Olmo hybrid linear attention, the other for Mamba2 use_cache with seq_len > 1. The soft similarity is only at the pattern level, not the underlying bug/change.", - "confidence": 0.94, + "summary": "These two PRs both touch cached multi-token forward paths, but they target different model families and different bugs. They are superficially similar yet not the same underlying change, so the soft edge should be rejected.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46084 is the clearest canonical representative: it targets one concrete Mamba2 cache-path defect, has an explicit tracking issue, and the diff is narrowly scoped to the seq_len > 1 cached-forward behavior.", + "canonical_pr_reason": "No canonical duplicate PR: #45513 is a broad Qwen3.5/Qwen3-next/OLMo-hybrid cache fix, while #46084 is a focused Mamba2 seq_len>1 cache-path fix for a different code path.", "best_issue_reason": null, - "best_pr_reason": "PR #46084 is the best overall fit because it is the most focused and self-contained fix for a single code path, whereas #45513 spans multiple model files and appears to cover a broader family of attention/cache adjustments.", + "best_pr_reason": "#46084 is the cleaner standalone PR: it is tightly scoped, explicitly tied to issue #46032, and fixes one concrete cache/seq_len bug in Mamba2.", "soft_edge_verdicts": [ { "left": "pull_request:45513", "right": "pull_request:46084", "accept": false, - "reason": "Both concern cached forward paths, but they fix different models and different mechanisms (Qwen3.5/Olmo hybrid GDN linear attention vs. Mamba2 use_cache). The shared theme is too broad to treat as the same underlying bug or mergeable PR." + "reason": "Both involve use_cache and multi-token cached forwarding, but they address different architectures (Qwen/OLMo hybrid vs Mamba2) and different implementation bugs. They do not appear to be the same fix and would not plausibly merge into one PR." } ] }, @@ -21047,7 +21284,7 @@ { "label": "pull_request", "component_index": 23, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21063,7 +21300,7 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "39924d2a17961bb40f4dbdb706cfdc9a11f30aeb19b5c582ac5c1b021db1d329", + "prepared_review_unit_hash": "5bf18bcf0dea43cc3a836c3a5fee2b5261e2e8d474c8f499127efd0318ee9799", "component_budget": { "node_count": 2, "item_count": 2, @@ -21085,18 +21322,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs both touch expert-parallel loading, but they fix different concrete bugs: one is about EP + DeepSpeed ZeRO-3 / accelerate launch serialization and config handling, the other is about EP + FSDP2 preserving sharded expert parameters during loading. They are related in subsystem, not duplicates.", - "confidence": 0.96, + "summary": "The two PRs both involve expert parallelism, but they address different concrete bugs: one fixes EP + ZeRO-3 loading under accelerate launch, the other fixes EP + FSDP2 expert parameters being overwritten by rank-0 broadcast. They share a broad subsystem and one file, but not the same underlying change.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45662 is the better canonical representative for this cluster because it contains the more direct loading/sharding fix in the actual code path where expert parameters were being overwritten, and its bug statement is concrete and self-contained.", + "canonical_pr_reason": "PR #45662 is the better canonical representative for the cluster because it fixes the more explicit EP runtime corruption bug and includes the broader code-path changes across loading and MoE/tensor-parallel integration.", "best_issue_reason": null, - "best_pr_reason": "PR #45662 is the strongest standalone candidate: it targets a specific failure mode (rank-0 broadcast overwriting expert weights) with code changes in the loading pipeline, whereas #45548 is a separate accelerate/ZeRO-3 configuration/loading adjustment.", + "best_pr_reason": "PR #45662 is the strongest standalone candidate: it targets a concrete failure mode (experts overwritten by rank-0 broadcast) with a coherent fix across the relevant loading and distributed parameter paths.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Reject: both concern EP loading, but #45548 fixes distributed_config serialization/has_ep for accelerate+ZeRO-3, while #45662 fixes expert parameters being overwritten during FSDP2 loading. Different code paths and failure modes." + "reason": "Different bugs and different fix paths: #45548 is about EP + DeepSpeed ZeRO-3 loading via accelerate launch, while #45662 is about EP + FSDP2 rank-0 broadcast overwriting expert parameters. Shared EP context and one shared file are not enough to treat them as the same PR-worthy change." } ] }, @@ -21110,7 +21347,7 @@ { "label": "pull_request", "component_index": 24, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21126,12 +21363,12 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "186260e80639210f5eacb5d4b234ca36f3fb22c9576fc17c8e68d2e8b7dc67cd", + "prepared_review_unit_hash": "c1b0045dc92905bd5ad14ac59dd0c3c91f7fa498441ec9f94954f5ee11da0ef5", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4589, + "serialized_chars": 4590, "estimated_input_tokens": 1148, "estimated_eval_tokens": 2552 }, @@ -21139,7 +21376,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4589, + "serialized_chars": 4590, "estimated_input_tokens": 1148, "estimated_eval_tokens": 2552 }, @@ -21148,18 +21385,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same captured_info/xdist collision problem under the same tracking issue; PR 45645 is the broader, more complete fix and subsumes the core testing_utils change in 45639.", - "confidence": 0.84, + "summary": "Both PRs address the same underlying xdist collision in captured_info debug artifacts; #45645 is the broader, more complete follow-up that also updates CI log collection, while #45639 is the smaller core fix in testing_utils.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the better canonical PR because it includes the core xdist-safe output-file fix in testing_utils plus the CI workflow adjustment needed to consume multiple captured_info files, so it represents the fuller end-to-end change.", + "canonical_pr_reason": "PR #45645 is the better canonical PR because it covers the core xdist-safe captured_info fix plus the downstream workflow/log-handling changes needed to preserve debug output in CI.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the strongest representative of the cluster: it fixes the underlying xdist collision and also preserves CI debug log collection, whereas 45639 is a narrower subset of the same fix.", + "best_pr_reason": "PR #45645 is the best overall representative since it fully addresses the artifact-collision problem and includes the CI-side adjustments, not just the library-side change.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Accepted: both target the same captured_info xdist collision issue and share the same core testing_utils/tests fix; 45645 simply extends it with workflow/log-handling changes, so they are the same underlying change family." + "reason": "Both PRs fix the same concrete bug: xdist workers colliding on captured_info debug output. #45645 appears to extend the same fix with CI artifact aggregation, so they are plausibly mergeable as one change." } ] }, @@ -21173,7 +21410,7 @@ { "label": "pull_request", "component_index": 25, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21191,46 +21428,46 @@ "pull_request:45670|pull_request:46320", "pull_request:46320|pull_request:46401" ], - "prepared_review_unit_hash": "3069d642e7ba0f9ed9fb5ffc4a9046e94cf0f3f363ebdefb33aedcb5bf9b63fb", + "prepared_review_unit_hash": "42790d45d71d8710b14213bb15d55f197d452de65d29dbf9f053fb8de94907fc", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 7029, - "estimated_input_tokens": 1758, - "estimated_eval_tokens": 3772 + "serialized_chars": 7028, + "estimated_input_tokens": 1757, + "estimated_eval_tokens": 3770 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 7029, - "estimated_input_tokens": 1758, - "estimated_eval_tokens": 3772 + "serialized_chars": 7028, + "estimated_input_tokens": 1757, + "estimated_eval_tokens": 3770 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not true duplicates. One is an AutoModel registration tweak for GLM ASR, while the other two add attn_implementation dispatch support for different model families. The RecurrentGemma and BLIP-2 PRs are thematically similar but touch different code paths and would not realistically be merged into one PR.", - "confidence": 0.93, + "summary": "These PRs form a loose topic cluster around model/auto-model support and attention dispatch, but they are not duplicates of the same underlying change. The only potentially similar pair is the RecurrentGemma and BLIP-2-family SDPA/attn dispatch work, which targets different model code paths and would not plausibly merge into one PR. The GLM ASR auto-mapping PR is unrelated.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46401 is the strongest representative of the cluster: it makes a concrete SDPA dispatch change across the BLIP-2/InstructBLIP family and includes the relevant model/test updates. The others are either unrelated (GLM ASR auto-mapping) or a separate model-family implementation of a similar pattern (RecurrentGemma).", + "canonical_pr_reason": "PR #46401 is the best canonical representative only in the sense that it is the clearest, most self-contained implementation PR in the cluster: it adds SDPA dispatch for the BLIP-2 family with code changes and tests. However, it is not a duplicate of the RecurrentGemma or GLM ASR PRs.", "best_issue_reason": null, - "best_pr_reason": "PR #46401 is the best standalone PR because it has the clearest, concrete code-path fix with accompanying tests across affected implementations. PR #46320 is similar in theme but for a different model family, and #45670 is unrelated to the attn dispatch work.", + "best_pr_reason": "PR #46401 is the strongest standalone PR here because it has the clearest concrete code-path fix, spans the main implementation plus tests/docs, and is more substantial than the small auto-model mapping tweak in #45670 or the narrower RecurrentGemma dispatch change in #46320.", "soft_edge_verdicts": [ { "left": "pull_request:45670", "right": "pull_request:46320", "accept": false, - "reason": "Different underlying changes: GLM ASR auto-class registration vs. RecurrentGemma attention-dispatch support. They do not fix the same code path and would not be merged as one PR." + "reason": "Different underlying changes: auto-model registry mapping for glmasr vs RecurrentGemma attention-dispatch support. Same broad Transformers area, but not the same bug or mergeable change." }, { "left": "pull_request:46320", "right": "pull_request:46401", "accept": false, - "reason": "Both are about attn_implementation dispatch, but they target different model families and different implementations. This is a pattern match, not the same concrete bug/change." + "reason": "Both mention attn_implementation dispatch, but they fix different model families and code paths (RecurrentGemma vs BLIP-2/InstructBLIP). They are related themes, not duplicates of one concrete change." } ] }, @@ -21244,7 +21481,7 @@ { "label": "pull_request", "component_index": 26, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21260,40 +21497,40 @@ "soft_pairs": [ "pull_request:45720|pull_request:46259" ], - "prepared_review_unit_hash": "22833d5708f97d3ea4e31633e01e793a056bc98774a93608fb555df4abbe5035", + "prepared_review_unit_hash": "4c1941fb403da8d2182f6de834754e58e4c48a81daee9362ab79ad8b4ab63d70", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7187, - "estimated_input_tokens": 1797, - "estimated_eval_tokens": 3850 + "serialized_chars": 7189, + "estimated_input_tokens": 1798, + "estimated_eval_tokens": 3852 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 7187, - "estimated_input_tokens": 1797, - "estimated_eval_tokens": 3850 + "serialized_chars": 7189, + "estimated_input_tokens": 1798, + "estimated_eval_tokens": 3852 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two unrelated open PRs: one implements mlinter rule TRF003 across model forward signatures, the other adds modeling rule TRF018 for image/video processor boolean flags. They overlap only in broad transformer model file coverage, not in the underlying change or bug.", + "summary": "Two unrelated linter/modeling-rule PRs: one implements TRF003 for `return_dict`/tuple-return handling in model forwards, the other adds TRF016 for preprocessing flag consumption. They share some touched model files but not the same underlying bug or change.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45720 is the more concrete code-change PR, while #46259 is primarily a rule/docs expansion with a different target area.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #45720 is the best representative of the cluster because it contains the actual behavioral implementation across model files; #46259 is a separate lint-rule addition.", + "best_pr_reason": "No single PR is a good canonical representative because these are distinct rule additions/fixes affecting different code paths and different linter behavior.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "Different lint rules and different code paths: TRF003 changes model forward signatures/return handling, while TRF018 checks processor boolean flag consumption. Shared subsystem/files are not enough to treat them as the same PR." + "reason": "Different changes: TRF003 modifies model forward signatures/tuple-return behavior, while TRF016 adds a rule about dead `do_*` preprocessing flags. Shared model filenames are incidental and do not indicate the same fix." } ] }, @@ -21307,7 +21544,7 @@ { "label": "pull_request", "component_index": 27, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21373,7 +21610,7 @@ { "label": "pull_request", "component_index": 28, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21389,7 +21626,7 @@ "soft_pairs": [ "pull_request:45932|pull_request:46100" ], - "prepared_review_unit_hash": "4f72dd278fe838412cc400c7d30806c57073a645e8c1405b41aecf3ab23f47a8", + "prepared_review_unit_hash": "26f25861a59b30af689700ec2c1e11a92e1884bdb364ae72c64eea37f6b79dc0", "component_budget": { "node_count": 2, "item_count": 2, @@ -21411,18 +21648,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated do-not-merge PRs: one is a broad testing/telemetry workflow and model-file change set, the other is a ViT test-migration PR that injects failing assertions. They should not be treated as duplicates.", - "confidence": 0.95, + "summary": "The two PRs are not duplicates: one is a broad code/config/workflow change across multiple model files, while the other is a test-only PR adding deliberately failing assertions in ViT tests. They do not address the same underlying bug or change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR 45932 is the substantive cluster member, touching actual model/config/workflow code across multiple files; PR 46100 is only a test-migration placeholder with failing assertions and is not the same change.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "PR 45932 is the best representative because it is the real implementation change; PR 46100 is clearly test-only and not a candidate for duplicate consolidation with it.", "soft_edge_verdicts": [ { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "They do not fix the same concrete problem or code path. PR 45932 touches workflow secrets and multiple model implementations, while PR 46100 only edits a ViT test file to add intentional failures. They would not plausibly be merged as one PR." + "reason": "Reject: 45932 changes model/config/workflow plumbing across several files, while 46100 only edits ViT tests with dummy failing assertions. They do not fix the same code-path problem and would not plausibly be merged as one PR." } ] }, @@ -21436,7 +21673,7 @@ { "label": "pull_request", "component_index": 29, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21455,7 +21692,7 @@ "pull_request:46073|pull_request:46196", "pull_request:46154|pull_request:46196" ], - "prepared_review_unit_hash": "d1b8d484d4ef4cc515377ac44df33eeccdc68eb3bc327c1897ce3290544373e0", + "prepared_review_unit_hash": "e254aee5bacdb3f280e0af5528a85fc8262fef2b4e50552dccdef2fa7eb1417e", "component_budget": { "node_count": 3, "item_count": 3, @@ -21477,30 +21714,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three PRs are related to SDPA attention dispatch support, but they target different model families and code paths, so they should not be deduplicated into one change.", - "confidence": 0.95, + "summary": "Three separate model-specific PRs add SDPA dispatch support in different code paths (CTRL, T5, and wav2vec2/SeamlessM4T variants). They share a theme but do not look mergeable into one single duplicate PR.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46196 is the broadest representative of this rollout, covering multiple model families and corresponding tests/docs, so it best captures the cluster theme.", + "canonical_pr_reason": "PR 46196 is the broadest representative of the cluster: it covers multiple model families and the same SDPA-dispatch theme, making it the best single PR to anchor the cluster.", "best_issue_reason": null, - "best_pr_reason": "PR #46196 is the most comprehensive and representative SDPA-dispatch implementation in the set.", + "best_pr_reason": "PR 46196 is the best representative PR because it spans the widest set of affected models and most clearly captures the shared feature direction across the cluster.", "soft_edge_verdicts": [ { "left": "pull_request:46073", "right": "pull_request:46154", "accept": false, - "reason": "Both are SDPA dispatch support PRs, but one is for CTRL and the other for T5; they touch different model implementations and are not the same concrete fix." + "reason": "Both add SDPA support, but for different architectures (CTRL vs T5) with different model code paths and tests; not the same underlying change." }, { "left": "pull_request:46073", "right": "pull_request:46196", "accept": false, - "reason": "These cover different model families and code paths (CTRL vs wav2vec2_conformer/bert and SeamlessM4T/v2), so they are not mergeable as one PR." + "reason": "These target different model families and distinct attention implementations; they are related feature work, not duplicate fixes or one mergeable PR." }, { "left": "pull_request:46154", "right": "pull_request:46196", "accept": false, - "reason": "T5 SDPA dispatch is a separate model-specific change from the multi-model wav2vec2/SeamlessM4T dispatch work." + "reason": "T5 SDPA dispatch and wav2vec2/SeamlessM4T SDPA dispatch are separate model-specific changes with different affected files and logic." } ] }, @@ -21514,7 +21751,7 @@ { "label": "pull_request", "component_index": 30, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21530,12 +21767,12 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "4c0d1603df0357fe93ee282237c4ee328004f9a3451a06f75427ef2239976fa8", + "prepared_review_unit_hash": "fdd46eda2eb3f503daa9baf11ace80366f82182947d87d64b848c884ab035226", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5572, + "serialized_chars": 5571, "estimated_input_tokens": 1393, "estimated_eval_tokens": 3042 }, @@ -21543,7 +21780,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5572, + "serialized_chars": 5571, "estimated_input_tokens": 1393, "estimated_eval_tokens": 3042 }, @@ -21552,18 +21789,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both large distributed-integration PRs, but they are not the same change: one is a focused EP/Trainer integration patch, while the other is a broad distributed branch base touching FSDP, tensor parallel, integrations, and multiple models. Shared filenames alone aren\u2019t enough to treat them as duplicates.", - "confidence": 0.79, + "summary": "Both items are PRs in the distributed stack, but they are not the same change. PR 46126 is a focused EP/Trainer integration patch, while PR 46269 is a broad distributed branch base that spans many subsystems.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46269 is the broader, more central distributed-branch base with wider architectural scope and many touched subsystems, making it the better canonical anchor for this cluster.", + "canonical_pr_reason": "PR 46126 is the more focused, concrete change set and better represents a single underlying fix/feature than the broad base branch in PR 46269.", "best_issue_reason": null, - "best_pr_reason": "PR #46126 is the more concrete and narrowly defined change set, centered on EP + Trainer integration, so it is the cleaner standalone PR if a specific implementation example is needed.", + "best_pr_reason": "PR 46126 is the best representative PR for triage because it targets a specific EP routing/Trainer integration path and is narrowly scoped.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Different underlying work: #46126 adds RouterParallel/EP routing and Trainer integration, while #46269 is a broad distributed branch base with FSDP CI and many unrelated distributed/model updates. They do not look mergeable into one PR without collapsing distinct feature streams." + "reason": "Reject: the overlap is only superficial. 46126 implements EP/Trainer integration around tensor-parallel routing, while 46269 is a large distributed branch base touching many unrelated files; they do not appear to fix the same concrete code-path problem and would not plausibly merge into one PR." } ] }, @@ -21577,7 +21814,7 @@ { "label": "pull_request", "component_index": 31, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21593,7 +21830,7 @@ "soft_pairs": [ "pull_request:46152|pull_request:46207" ], - "prepared_review_unit_hash": "4a11209dd233e92ac470c6931e5f867f438ea44869ec38a8b6c51bf9bfb23d29", + "prepared_review_unit_hash": "30e14149d8fdf8fd0be3feddf93f0b290fe8c1dd1fd0b044d4c6d997713ff1b2", "component_budget": { "node_count": 2, "item_count": 2, @@ -21615,18 +21852,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs with a similar theme of avoiding direct file-open source inspection, but they target different modules and different code paths: one in import_utils for package self-source resolution, the other in modeling_utils for attention-implementation detection. They are not the same underlying change.", - "confidence": 0.92, + "summary": "Both PRs are about removing brittle source-file open() usage, but they fix different code paths in different modules: one in import_utils for resolving package traversables, the other in modeling_utils for attention-implementation detection. They are thematically related but not duplicates.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR stands out as a duplicate of the other; they fix different functions in different files, so neither can serve as a cluster-wide canonical for the pair.", + "canonical_pr_reason": "PR 46207 is the more concrete and self-contained fix: it updates the _can_set_attn_implementation path in modeling_utils to use inspect.getsource-style source inspection instead of file open(), with tests. It targets a single behavior change and is clearly scoped.", "best_issue_reason": null, - "best_pr_reason": "PR #46207 is the stronger standalone candidate because it targets a concrete behavior in _can_set_attn_implementation and includes tests, making the change easier to validate than the more internal helper refactor in PR #46152.", + "best_pr_reason": "PR 46207 is the best representative of the cluster because it addresses one specific bug-prone source-reading path directly. PR 46152 is broader and centered on a utility helper in import_utils, so it is less canonical for this cluster.", "soft_edge_verdicts": [ { "left": "pull_request:46152", "right": "pull_request:46207", "accept": false, - "reason": "Both avoid open() for source reading, but they affect different subsystems and distinct code paths with no shared concrete bug or mergeable unified patch." + "reason": "Same broad theme (avoid open() on source), but they modify different modules and different concrete mechanisms. 46152 adds a traversable resolver in import_utils; 46207 changes model capability detection in modeling_utils. Not the same underlying bug or mergeable into one PR." } ] }, @@ -21640,7 +21877,7 @@ { "label": "pull_request", "component_index": 32, - "component_count": 32, + "component_count": 33, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -21661,12 +21898,12 @@ "pull_request:46362|pull_request:46368", "pull_request:46362|pull_request:46366" ], - "prepared_review_unit_hash": "b504d55d0c83dfc1ab3fa49402a5b81a4271d917d34f388f9fccdd8cc3aa735a", + "prepared_review_unit_hash": "538fa3eed621a1be5726697d03b63ee2c8c18f113f0df68c3690d958ab26e4d7", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 8226, + "serialized_chars": 8228, "estimated_input_tokens": 2057, "estimated_eval_tokens": 4370 }, @@ -21674,7 +21911,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 8226, + "serialized_chars": 8228, "estimated_input_tokens": 2057, "estimated_eval_tokens": 4370 }, @@ -21683,36 +21920,99 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are mostly unrelated test-only fixes across different models; the two quant-cache skips look similar superficially but target different model-specific cache implementations, so they should not be merged as duplicates.", - "confidence": 0.92, + "summary": "The cluster is superficially similar because all four PRs touch model test files and three of them are test skips, but they address different models and different concrete behaviors. There is no true duplicate pair here.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "46362 is the clearest representative of the cluster\u2019s one recurring theme: a minimal, explicit skip for an incompatible quant-cache test in a specific model test file.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "46362 is the most self-contained and easy-to-review change: it directly documents and skips one incompatible test without introducing broader test-environment behavior changes.", + "best_pr_reason": "PR 46366 is the most substantive and self-contained change in the set: it adds backend detection and a clearer error path for the deepseek_v4 distributed test. Even so, it is not a duplicate of the other PRs.", "soft_edge_verdicts": [ { "left": "pull_request:46337", "right": "pull_request:46368", "accept": false, - "reason": "Different models and different fixes: cohere2_moe adjusts test configuration for accelerate splitting, while nemotron_h only skips a quant-cache test. Same broad test area, but not the same underlying bug." + "reason": "Both are skip-style test changes, but for different architectures and different model-specific failures: cohere2_moe test sizing/offload behavior vs nemotron_h quant-cache incompatibility." }, { "left": "pull_request:46337", "right": "pull_request:46362", "accept": false, - "reason": "These address unrelated test issues in different model families; one changes MoE test sizing, the other skips a DeepSeek V4 quant-cache test." + "reason": "cohere2_moe adjusts test configuration to fit accelerate splitting, while deepseek_v4 skips a quant-cache test; these are unrelated code paths and bugs." }, { "left": "pull_request:46362", "right": "pull_request:46368", "accept": false, - "reason": "Both skip a similarly named test, but the failures are model-specific and stem from different cache implementations. They are not the same concrete code-path problem and would not be merged into one PR." + "reason": "They share the same test name pattern, but the failures are model-specific quant-cache incompatibilities in different models, so they are not the same underlying change." }, { "left": "pull_request:46362", "right": "pull_request:46366", "accept": false, - "reason": "DeepSeek V4 quant-cache skipping and XPU/distributed-worker support are different code paths with different goals; they are not duplicate fixes." + "reason": "One PR skips a quant-cache test; the other adds CUDA/XPU backend selection for distributed execution. Different functionality, not one mergeable fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 33, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46376-2", + "nodes": [ + "pull_request:46376", + "pull_request:46417" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417" + ], + "prepared_review_unit_hash": "082c3754f1ae3cf547f9c1eb05f06f159c9f86af014f648bb0cc3ed501665595", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5960, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5960, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are unrelated: one adds multimodal support to continuous batching, the other adds a new HunYuan VL model integration and docs. No duplicate/merge candidate here.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical duplicate PR exists in this cluster; the PRs target different code paths and feature areas.", + "best_issue_reason": null, + "best_pr_reason": "PR 46417 is the clearer standalone representative because it is a complete model integration with tests/docs, but it is not a duplicate of 46376.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Different underlying changes: continuous batching multimodal/cache/scheduler work vs a HunYuan VL model add-on with docs, auto-mappings, and tests. No shared concrete bug or fix path, so they should not be merged." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index c0fb7d8edc65af923c46800a70e8bfa7d3476089..49e4a14f23cf5a9a2cec242274483810c7fca981 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260604t120021z", + "analysis_id": "hybrid-model-20260604t180024z", "archived_artifacts": { - "hybrid": "snapshots/20260604T120021Z/analysis-runs/hybrid-model-20260604t120021z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260604T120021Z/analysis-runs/hybrid-model-20260604t120021z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-06-04T12:10:14Z", + "published_at": "2026-06-04T18:11:32Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260604T120021Z", + "snapshot_id": "20260604T180024Z", "variant": "hybrid" } diff --git a/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.json b/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..8a1df5406fa12c14a650e02dab37e062795bef92 --- /dev/null +++ b/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.json @@ -0,0 +1,2081 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260604T180024Z", + "generated_at": "2026-06-04T18:11:31Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-41115-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41115, + "canonical_pr_number": 45613, + "issue_numbers": [ + 41115 + ], + "pr_numbers": [ + 41116, + 45613 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41116, + "right_pr_number": 45613, + "code_similarity": 0.797, + "size_similarity": 0.885, + "file_overlap": 0.75, + "area_overlap": 0.831, + "patch_similarity": 0.693, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "docs/source/en/model_doc/minicpm3.md", + "src/transformers/models/auto/modeling_auto.py", + "src/transformers/models/minicpm3/__init__.py", + "src/transformers/models/minicpm3/configuration_minicpm3.py", + "src/transformers/models/minicpm3/modeling_minicpm3.py", + "src/transformers/models/minicpm3/modular_minicpm3.py", + "tests/models/minicpm3/__init__.py", + "tests/models/minicpm3/test_modeling_minicpm3.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 713, + 720 + ] + ], + "right_ranges": [ + [ + 730, + 737 + ] + ] + }, + { + "filename": "docs/source/en/model_doc/minicpm3.md", + "left_ranges": [ + [ + 1, + 67 + ] + ], + "right_ranges": [ + [ + 1, + 45 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 285, + 291 + ], + [ + 700, + 706 + ], + [ + 1298, + 1304 + ] + ], + "right_ranges": [ + [ + 720, + 726 + ], + [ + 1331, + 1337 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/__init__.py", + "left_ranges": [ + [ + 1, + 29 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/configuration_minicpm3.py", + "left_ranges": [ + [ + 1, + 141 + ] + ], + "right_ranges": [ + [ + 1, + 126 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 544 + ] + ], + "right_ranges": [ + [ + 1, + 522 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modular_minicpm3.py", + "left_ranges": [ + [ + 1, + 444 + ] + ], + "right_ranges": [ + [ + 1, + 342 + ] + ] + }, + { + "filename": "tests/models/minicpm3/test_modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 109 + ] + ], + "right_ranges": [ + [ + 1, + 136 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45647-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45647.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45647, + "canonical_pr_number": 45738, + "issue_numbers": [ + 45647 + ], + "pr_numbers": [ + 45737, + 45738 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45737, + "right_pr_number": 45738, + "code_similarity": 0.526, + "size_similarity": 0.11, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.024, + "shared_filenames": [ + "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py", + "left_ranges": [ + [ + 491, + 497 + ] + ], + "right_ranges": [ + [ + 491, + 497 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.494, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.434, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3595, + 3602 + ], + [ + 3819, + 3853 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43656-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 43656, + "issue_numbers": [], + "pr_numbers": [ + 43656, + 43842 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43656, + "right_pr_number": 43842, + "code_similarity": 0.405, + "size_similarity": 0.5, + "file_overlap": 1.0, + "area_overlap": 0.0, + "patch_similarity": 0.036, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 30, + 36 + ], + [ + 315, + 323 + ], + [ + 665, + 671 + ], + [ + 931, + 937 + ], + [ + 1843, + 1849 + ], + [ + 1868, + 1874 + ] + ], + "right_ranges": [ + [ + 54, + 61 + ], + [ + 587, + 593 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, + "file_overlap": 1.0, + "area_overlap": 0.425, + "patch_similarity": 0.866, + "shared_filenames": [ + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "left_ranges": [ + [ + 26, + 46 + ], + [ + 138, + 143 + ], + [ + 182, + 187 + ], + [ + 283, + 295 + ], + [ + 327, + 341 + ], + [ + 348, + 354 + ], + [ + 360, + 369 + ], + [ + 399, + 406 + ], + [ + 411, + 419 + ], + [ + 428, + 433 + ], + [ + 472, + 492 + ], + [ + 507, + 513 + ], + [ + 519, + 528 + ], + [ + 541, + 559 + ], + [ + 562, + 567 + ], + [ + 595, + 601 + ], + [ + 607, + 614 + ], + [ + 627, + 643 + ], + [ + 685, + 690 + ], + [ + 708, + 714 + ], + [ + 720, + 727 + ], + [ + 740, + 757 + ], + [ + 761, + 766 + ], + [ + 780, + 786 + ], + [ + 791, + 798 + ], + [ + 807, + 822 + ], + [ + 840, + 845 + ] + ], + "right_ranges": [ + [ + 26, + 31 + ], + [ + 34, + 44 + ], + [ + 136, + 141 + ], + [ + 180, + 185 + ], + [ + 281, + 293 + ], + [ + 325, + 339 + ], + [ + 346, + 352 + ], + [ + 358, + 367 + ], + [ + 397, + 404 + ], + [ + 409, + 417 + ], + [ + 426, + 434 + ], + [ + 463, + 483 + ], + [ + 498, + 504 + ], + [ + 510, + 519 + ], + [ + 532, + 550 + ], + [ + 553, + 564 + ], + [ + 586, + 592 + ], + [ + 598, + 605 + ], + [ + 618, + 634 + ], + [ + 676, + 688 + ], + [ + 699, + 705 + ], + [ + 711, + 718 + ], + [ + 731, + 748 + ], + [ + 752, + 762 + ], + [ + 771, + 777 + ], + [ + 782, + 789 + ], + [ + 798, + 813 + ], + [ + 831, + 836 + ] + ] + }, + { + "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", + "left_ranges": [ + [ + 458, + 464 + ] + ], + "right_ranges": [ + [ + 458, + 464 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41115-3", + "canonical_pr_number": 45613, + "duplicate_pr_numbers": [ + 41116 + ], + "target_issue_number": 41115, + "reason": "PRs in cluster-41115-3 are treated as duplicates because they converge on issue #41115 with closing_reference evidence." + }, + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43656-2", + "canonical_pr_number": 43656, + "duplicate_pr_numbers": [ + 43842 + ], + "target_issue_number": null, + "reason": "PRs in cluster-43656-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-41211-3", + "issue_number": 41211, + "reason": "Issue #41211 is the strongest global issue candidate because it is open, belongs to a cluster with 3 artifacts, and carries 6 discussion comments plus 6 inbound references.", + "score": 148.0 + }, + "best_pr": { + "cluster_id": "cluster-43979-11", + "pr_number": 44007, + "reason": "PR #44007 is the strongest global PR candidate because it is open, belongs to a cluster with 11 artifacts, links to 1 issues, and carries 2 review/discussion events.", + "score": 238.886 + } +} diff --git a/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..c5efecb55bc693660dd99c23b4354d6e94001670 --- /dev/null +++ b/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,22027 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260604T180024Z", + "generated_at": "2026-06-04T18:11:31Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 1, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43296", + "issue:43425", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43827", + "issue:43828", + "issue:43866", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:46015", + "issue:46165", + "issue:46294", + "issue:46304", + "issue:46402" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44918|issue:46165", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:46294", + "issue:44805|issue:45161", + "issue:44560|issue:46165", + "issue:43596|issue:46165", + "issue:43572|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:45874|issue:46294", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:43828|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:45685", + "issue:44918|issue:45161", + "issue:44292|issue:44918", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43828|issue:45685", + "issue:44918|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:45237|issue:46015", + "issue:43039|issue:44805", + "issue:43866|issue:46402", + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:41762|issue:43596", + "issue:39290|issue:41720" + ], + "prepared_review_unit_hash": "53bd3fd6d0990d3a29343d0e99ae884846eb64591f640e9f1aa3fbb4f15259dc", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 64, + "serialized_chars": 46877, + "estimated_input_tokens": 11720, + "estimated_eval_tokens": 23696 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are unrelated. The only clear duplicate cluster is the DeepSpeed ZeRO-3 zero-sized tensor IndexError around Bert/Gemma model loading; the rest are distinct model-, backend-, or docs-specific issues.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43638 is the clearest representative of the ZeRO-3 zero-sized tensor IndexError cluster: it states the exact exception, the DeepSpeed ZeRO-3 context, and a concrete trigger (non-pretrained Bert).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43638 is the most general and explicit reproduction of the shared ZeRO-3 loading bug, so it is the best canonical issue for that cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different Qwen3.5 bugs: missing config attribute vs streaming chat/completions failure." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different hardware/backend failures: AMD GPU issue vs MPS histogram op limitation." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different models and failure modes; not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Different symptoms: autocast dtype mismatch vs AMD GPU failure." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different problems: AMD GPU runtime issue vs invalid method call in GraniteMoEHybrid." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Different models and error paths; not a duplicate." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "Distributed collective mismatch in training is unrelated to ONNX export failure." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "Export-to-ONNX bug and FineGrainedFP8 config failure are different code paths." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export issue vs auto device-mapping cuda assert; unrelated." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "Table QA crash and outdated RAG examples are unrelated." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 docs cleanups, but they target different removed docs content." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export bug and GenerationMixin import error are separate issues." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "TRL SFT embedding unpacking and training resume with remote code are different failures." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Generic mask-shape mismatch vs Gemma4 device-mapping mismatch; different bugs." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Different exception types and different model/runtime contexts." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Zero-3 init IndexError is unrelated to mask/tensor shape mismatch." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Mask-shape mismatch and Qwen3.5 video token mismatch are different problems." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Different failures: mask/indexing bug vs GPT-OSS TP issue." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video apply_chat_template failure is unrelated to resume/trust_remote_code training issue." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "DeepSpeed ZeRO-3 Bert loading bug is unrelated to training resume with remote code." + }, + { + "left": "issue:43572", + "right": "issue:46165", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice_preset bug is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both involve training/runtime issues, but the errors and code paths are different." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 index-out-of-bounds loading bug on Bert models; only the model wording differs." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross_entropy call and GPT-OSS TP failure are different code paths." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Incorrect detokenization and LoRA merge/reload mismatch are different bugs." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "NVFP4 runtime failure is unrelated to LoRA save/reload inconsistency." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "FSDP checkpointing size mismatch is unrelated to the mask indexing error." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "DeepSpeed weight-dimension error and collective mismatch are different failure modes." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Same model family, but CPU offload incompatibility and meta-to-empty uninitialized memory are different concrete bugs." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "Zero-3 Bert loading bug is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "Checkpointing/FSDP tensor size mismatch is unrelated to CustomTrainer compute_loss signature change." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "DeepSpeed ZeRO-3 Bert loading bug is unrelated to GPT-OSS TP failure." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Dtype mismatch in Mistral-3 preprocessing is unrelated to GraniteMoEHybrid invalid method calls." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Different Gemma4 input-embedding issue vs Qwen3.5 video chat-template mismatch." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3/vLLM attribute error and Mistral regex patch crash are unrelated." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Different Gemma4 bugs: device mismatch in forward vs expensive reverse embedding lookup." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch/version incompatibility is unrelated to Qwen-3 NVFP4 runtime failure." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP activation checkpointing mismatch and Qwen3 auto device-map assert are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Phi-tiny-MoE autocast dtype mismatch is unrelated to GPT-OSS TP failure." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral grouped_mm tracing error and Mistral-3 preprocessor dtype mismatch are different bugs." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Different models and different dtype/runtime failures." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Image-preprocessor dtype mismatch is unrelated to MPS histogram op failure." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL embedding unpacking issue is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Different models and code paths; not the same bug." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "GraniteMoEHybrid invalid method call and MPS histogram op error are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Mistral-3 dtype mismatch and Qwen-3 NVFP4 failure are different issues." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Liger Kernel cross_entropy call is unrelated to TRL SFT embedding unpacking." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": true, + "reason": "Same DeepSpeed ZeRO-3 index-out-of-bounds loading error; only the affected model differs." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs cleanup, but they refer to different removed features and should stay separate." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Mistral-3 dtype mismatch is unrelated to Qwen3.5 video template token mismatch." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Autocast dtype mismatch is unrelated to MPS integer histogram support." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "TRL trainer issue is unrelated to MPS backend limitation." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Liger Kernel cross_entropy usage and Phi-tiny-MoE autocast mismatch are different bugs." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch and compute_loss signature change are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "FSDP checkpointing mismatch is unrelated to GenerationMixin import failure." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "AMD GPU runtime issue is unrelated to outdated RAG docs." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Cross-entropy kernel usage and mask-shape mismatch are unrelated." + }, + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Corrupted checkpoint and Sam3Model weight-loading regression are different issues." + }, + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "Torch/transformers compatibility complaint and USE_DISTRIBUTED=0 regression are unrelated." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "vLLM/PaddleOCR-VL load failure and distributed build regression are different bugs." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Closely related DeepSpeed ZeRO-3 IndexError reports, but the model/context differs enough to keep them separate without more evidence." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3/vLLM attribute issue is unrelated to Qwen3 auto device-map cuda assert." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the soft-edge calls are generally conservative. The one clear duplicate cluster is the DeepSpeed ZeRO-3 zero-sized tensor IndexError around Bert/Gemma loading, while the rest are separated by clearly different model, backend, or docs contexts." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 2, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:36010", + "issue:39290", + "issue:41093", + "issue:41762", + "issue:43039", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43792", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43901", + "issue:43975", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44512", + "issue:44560", + "issue:44589", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45161", + "issue:45237", + "issue:45362", + "issue:45464", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304" + ], + "soft_pairs": [ + "issue:44589|issue:46304", + "issue:45823|issue:46294", + "issue:43901|issue:44512", + "issue:45237|issue:45901", + "issue:44841|issue:45823", + "issue:43039|issue:44560", + "issue:43975|issue:46015", + "issue:45812|issue:45920", + "issue:44863|issue:46015", + "issue:45020|issue:46304", + "issue:43866|issue:44863", + "issue:43872|issue:46304", + "issue:44292|issue:46015", + "issue:39290|issue:43531", + "issue:41093|issue:43638", + "issue:43854|issue:43866", + "issue:43792|issue:45901", + "issue:43425|issue:45070", + "issue:43596|issue:45800", + "issue:43975|issue:45901", + "issue:43792|issue:45070", + "issue:45362|issue:45464", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:43404|issue:43866", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:44841|issue:45084", + "issue:36010|issue:39290", + "issue:45800|issue:45823", + "issue:41093|issue:41762", + "issue:43296|issue:43366", + "issue:43572|issue:45070", + "issue:44805|issue:44918", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:43531|issue:45070" + ], + "prepared_review_unit_hash": "0d10c028f016d239a66ca3499e4dfe0212ddf46e73cfa1385e13e6eb88278dbc", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39796, + "estimated_input_tokens": 9949, + "estimated_eval_tokens": 20154 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are clearly unrelated, with the only strong duplicate signal in the visible v5 AutoTokenizer wrong-token-ID regression cluster. The rest are mostly superficially similar error strings or broad subsystem overlaps without enough evidence of the same underlying bug.", + "confidence": 0.87, + "canonical_issue_reason": "issue:45920 is the best canonical issue because it is the broadest visible report of the v5 AutoTokenizer wrong-token-ID regression and spans multiple model families; issue:45812 is the narrower Granite-specific instance of the same symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best global issue anchor in this packet because it describes the broadest visible tokenizer regression and names several affected model families, making it the most representative report here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Different incompatibility classes: missing Float8 storage support versus a USE_DISTRIBUTED=0 regression." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Different models and failure surfaces: Gemma4 device-map mismatch versus Qwen3.5 video/chat-template token-feature mismatch." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs-related, but they concern unrelated documentation problems." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "Unrelated failures in different areas: GPT-OSS AMD GPU support versus table-question-answering crashes." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Different models and bugs: Voxtral processor failure versus Gemma4 device mismatch." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Different code paths and symptoms: Liger cross_entropy dispatch versus Qwen3-vl video StopIteration." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Tokenizer detokenization bug versus outdated RAG examples; not the same issue." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same underlying v5 AutoTokenizer regression producing wrong token IDs across multiple model families." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "Model-loading/architecture support issue versus outdated examples; unrelated root causes." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Broad remote_code breakage is not the same as the distributed-build regression." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Checkpoint corruption versus implementation loading support; different causes." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes constructor mismatch is unrelated to the USE_DISTRIBUTED=0 regression." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Model-loading type error versus a documentation issue; no shared bug." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but they affect different models and likely different code paths." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Similar IndexError wording, but the shapes and contexts differ enough to treat as separate bugs." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Different load failures: unit-test model loading problem versus checkpoint corruption." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper model execution failure is unrelated to table-question-answering crashes." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Tokenizer detokenization bug versus table QA crash; unrelated." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper runtime failure and pydantic config regression are different bugs." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Same model family, but different entrypoints and symptoms: chat crash versus streaming chat/completions failure." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Different subsystems and failure modes: streaming inference versus MPS histogram support." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Weight-tying bug in Mistral3 is unrelated to a corrupted checkpoint report." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Different backend failures: torch dynamo/grouped_mm tracing versus MPS histogram op support." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Both involve MoE or mixed precision, but the concrete failures and models differ." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Both are runtime errors, but one is a mask-shape IndexError and the other is an embedding-unpacking failure." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Quantization OOM regression versus tied-weights issues; different code paths." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Different models and problems: Mixtral tracing error versus GPT-OSS tensor-parallel-only failure." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Dtype mismatch under autocast is unrelated to the Qwen3.5 streaming API failure." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument and a PretrainedConfig pydantic-field regression are not the same bug." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF gpt-oss support request is unrelated to the Qwen3.5 streaming API bug." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Sliding-window model issue versus pydantic PretrainedConfig regression; different root causes." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "GenerationMixin import failure and Gemma3 sliding_window_pattern are unrelated." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers compatibility issue versus Gemma4 forward device mismatch." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are Zero-3-related IndexErrors, but the reported shapes and model contexts do not indicate the same bug." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL load failure and GGUF gpt-oss support request are different model-specific issues." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig and a pydantic PretrainedConfig regression are different config bugs." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The pairwise judgments are mostly conservative and grounded in the titles, but the top-level issue reasoning overstates the packet: it refers to an \"accepted duplicate family,\" which is not established in the provided data. I\u2019d want the global-anchor/canonical-issue claims phrased more cautiously and only in terms visible in the packet (broadest v5 tokenizer regression, multiple model families, etc.)." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 3, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:36331", + "issue:38175", + "issue:39186", + "issue:39692", + "issue:40990", + "issue:41093", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43404", + "issue:43541", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43866", + "issue:43872", + "issue:43873", + "issue:43950", + "issue:43975", + "issue:43994", + "issue:44292", + "issue:44360", + "issue:44403", + "issue:44479", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44589", + "issue:44863", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45399", + "issue:45507", + "issue:45542", + "issue:45632", + "issue:45698", + "issue:45701", + "issue:45759", + "issue:45800", + "issue:45902", + "issue:46294" + ], + "soft_pairs": [ + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:43541|issue:44589", + "issue:44479|issue:46294", + "issue:41093|issue:43596", + "issue:43054|issue:43994", + "issue:45759|issue:45902", + "issue:45542|issue:45800", + "issue:39186|issue:44918", + "issue:43716|issue:44918", + "issue:40990|issue:44960", + "issue:45632|issue:45698", + "issue:43645|issue:45399", + "issue:43404|issue:45005", + "issue:42915|issue:45399", + "issue:44530|issue:45005", + "issue:43873|issue:45005", + "issue:44863|issue:45005", + "issue:44360|issue:44512", + "issue:44918|issue:45507", + "issue:44492|issue:44512", + "issue:43975|issue:44292", + "issue:44292|issue:45507", + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue:43828|issue:44928", + "issue:36331|issue:41762", + "issue:43872|issue:44589", + "issue:43950|issue:45902", + "issue:43054|issue:45701", + "issue:43866|issue:44403", + "issue:38175|issue:43054" + ], + "prepared_review_unit_hash": "db9132f1972d069a9d611394def93a6cf6aaaae2dcf439f7658a73da63139c23", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36897, + "estimated_input_tokens": 9225, + "estimated_eval_tokens": 18706 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous set of unrelated closed issues. The only clear duplicate thread is the Transformers 5.x non-persistent buffer corruption regression around from_pretrained / buffer materialization.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43950 is the clearest root report of the non-persistent buffer corruption regression, with the most direct repro and wording; issue:43644 is the earlier broader report and issue:44534 is a rephrased follow-up.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 is the best representative duplicate target because it describes the concrete failure mode precisely and anchors the confirmed duplicate chain in this cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Different models and different failure modes (dtype mismatch vs video chat template token/feature mismatch)." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "One is a remote_code regression, the other is a torch/transformers version compatibility problem; not the same bug." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated runtime errors in different code paths (grouped_mm tracing vs missing Float8 storage)." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": false, + "reason": "Both mention Qwen video, but one is a broad v5.3.0 regression and the other is a specific apply_chat_template mismatch; not clearly the same bug." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are IndexErrors under DeepSpeed, but the concrete shapes and trigger differ." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but one reports degraded embeddings and the other nonsensical pipeline output; too vague to be the same concrete bug." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Different issues: composite config unwrapping vs uninitialized rotary embedding memory after materialization." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "TensorBoard/TensorFlow backend handling is unrelated to the torch 2.4.1/transformers 5.8.0 incompatibility report." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP 2-D weight error and Qwen3.5 SFT embedding unpacking are different failures." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Different models and subsystems; image preprocessor dtype mismatch is unrelated to TRL SFT embedding unpacking." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "Completely different models and symptoms (perplexity vs GLM5-related report)." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both involve custom-code loading, but one is a leaf-directory cache collision and the other is wrong module loading after save_pretrained; not clearly the same bug." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Custom model init in Jupyter and flash-attn2 fallback gating are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but they concern different model families and different manifestations." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "FineGrainedFP8 failure and flash-attn2 fallback checks are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache group-type crash is unrelated to translation-model tied-weight issues." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior does not match the tied-weights regression." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint-loading failure is unrelated to translation-model weight tying." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer bug and a docs typo are unrelated." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Different model families and different error mechanisms." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A cache-strategy typo and a removed CLI docs mention are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Detokenization regression and Qwen NVFP4 runtime error are different bugs." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Different model-specific runtime failures." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Both report the same non-persistent-buffer corruption regression in Transformers 5.x; 44534 is a reworded follow-up to 43950." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "SigLIP2 docs example problems and SigLIP2 embedding quality regression are related only at a broad model level, not the same concrete bug." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Same regression: non-persistent buffers get corrupted/junk-filled during loading/materialization in Transformers 5.x." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE and Qwen3.5 RLHF NaN explosion are different issues." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Trainer API keyword mismatch and DeepSpeed ZeRO-3 load error are unrelated." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different quantization-related incompatibilities with different symptoms." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "Both mention buffer/materialization issues, but 45902 is a specific uninitialized rotary embedding case and not the same regression." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "Model embedding quality and tokenization changes across versions are different problems." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Checkpoint corruption and noisy loading are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both are SigLIP2 complaints, but zero probabilities and poor embeddings are not specific enough to treat as the same bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and the soft-edge verdicts are conservative overall. The accepted duplicate thread around 43644/43950/44534 is well supported by the issue titles, and the other pair rejections avoid over-claiming sameness. Minor caution: a couple of phrases like \"most direct repro\" go a bit beyond what the packet explicitly shows, but not enough to invalidate the cluster assessment." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 4, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "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:43828", + "issue:43856", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44509", + "issue:44661", + "issue:44910", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45305", + "issue:45356", + "issue:45399", + "issue:45520", + "issue:45663", + "issue:45685", + "issue:45715", + "issue:45823", + "issue:45859", + "issue:45902", + "issue:46315" + ], + "soft_pairs": [ + "issue:41553|issue:43883", + "issue:44910|issue:45859", + "issue:45663|issue:45902", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:41553|issue:43881", + "issue:44403|issue:44451", + "issue:43975|issue:45520", + "issue:44403|issue:44488", + "issue:43643|issue:45020", + "issue:42947|issue:43856", + "issue:41762|issue:42915", + "issue:45685|issue:45823", + "issue:43782|issue:44960", + "issue:43824|issue:43872", + "issue:44991|issue:45399", + "issue:43881|issue:43883", + "issue:45020|issue:45823", + "issue:42491|issue:44368", + "issue:44661|issue:45020", + "issue:43828|issue:45823", + "issue:45081|issue:45399", + "issue:45715|issue:46315", + "issue:45020|issue:45356", + "issue:43742|issue:45399", + "issue:43404|issue:43950", + "issue:43827|issue:44509", + "issue:43872|issue:44291", + "issue:42915|issue:43872", + "issue:43792|issue:44220", + "issue:43257|issue:45305" + ], + "prepared_review_unit_hash": "173a96ffbeddc66583a1e2fd4f903d1559bf57ceb1133086007f453a9906736e", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36798, + "estimated_input_tokens": 9200, + "estimated_eval_tokens": 18656 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a grab bag of distinct v5 regression reports around model loading, tokenizers, MoE, quantization, and docs. Only two soft pairs look like true duplicates: the stale v5 docs/pipeline references and the `_is_hf_initialized` init_empty_weights/Int8Params TypeError.", + "confidence": 0.57, + "canonical_issue_reason": "issue:45020 is the broadest umbrella report for recent Transformers v5 loading/remote-code regressions, making it the best representative of the cluster\u2019s shared theme.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the most central and general issue here; the rest are narrower, model-specific failures or unrelated regressions.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Different models and failure modes: Voxtral has a bad AutoTokenizer error message, Molmo is a missing `all_tied_weights_keys` attribute." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Both involve Qwen3.5/MoE, but one is a FlashAttention illegal memory access from `position_ids`, the other is a CPU offload compatibility bug in rotary embedding." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Distinct bugs: Gemma4 FSDP2 `shared_kv_states` reconstruction vs Qwen3_5Moe `inv_freq` uninitialized memory after `to_empty(cuda)`." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Different regressions: ineffective gradient checkpointing with LoRA versus extra reserved memory/OOM under int4 quantization." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Related model family, but not the same bug: one reports worse SigLIP2 text embeddings, the other a HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Separate loading problems for different models; no clear shared code-path or symptom beyond generic load failure." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "One is cosmetic loading noise, the other is an actual model load failure for ScandiBERT." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Unrelated subsystems: tokenizer detokenization bug versus `flash_attn` import-time `KeyError`." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Generic warning/noise issue is not the same as the specific `cjvt/sleng-bert` loading failure." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Both involve `trust_remote_code`, but one is missing fields in `AutoConfig`, while the other is a broader remote-code breakage report." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different problems: LoRA gradient checkpointing versus Qwen3 MoE training memory usage." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Different models and errors: Gemma3 ZeRO-3 index error versus Qwen3 MOE FineGrainedFP8 failure." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram int limitation in MoE is unrelated to Gemma4 PLE device mismatch in forward." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different model families and failure types; one is a Qwen3VL `weight_only` load error, the other is an unspecified GLM5 issue." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Import error for `Qwen2_5_VLForConditionalGeneration` is unrelated to the bitsandbytes `_is_hf_initialized` compatibility problem." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer load failure for a specific model versus FlashAttention2 fallback being blocked by unrelated checks." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Both are loading failures, but for different models and root causes; no evidence they are the same bug." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Remote-code loading regressions are not the same as Gemma4 PLE device placement bugs." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Same broad Qwen3.5 LoRA area, but one is a hf4.x\u2192hf5.x model incompatibility and the other is a tie-word-embeddings warning." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "`add-new-model-like` failing in `TOKENIZER_MAPPING_NAMES` is a different registry bug than remote-code loading breakage." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE and Gemma4 PLE device mismatch are different code paths." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral regex patch AttributeError is unrelated to FlashAttention2 fallback gating." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance regressions, but they affect different APIs and likely different hot paths (`convert_ids_to_tokens` vs `convert_tokens_to_ids`)." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "`trust_remote_code` loading breakage is not the same as the Kimi-K2.5 tokenizer codec regression." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "Different areas: MobileLLM loading key error versus FlashAttention2 fallback selection." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "`lm_head` tying bug and non-persistent buffer corruption are separate state-management regressions." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "These describe the same docs regression: stale v5 documentation still referencing removed pipeline tasks like summarization/translation/text2text-generation." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying compatibility bug: `init_empty_weights` / bitsandbytes passes `_is_hf_initialized` into `Int8Params.__new__`, causing the same TypeError." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Different issues: Qwen3 MOE FineGrainedFP8 failure versus bitsandbytes `_is_hf_initialized` TypeError." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Insufficient evidence that the Whisper load failure and `_torch_extract_fbank_features()` issue are the same bug; could be related but not clearly identical." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Different DeepSpeed-related problems: Qwen3 MOE weight conversion on load versus gradient averaging under GAS with `model_accepts_loss_kwargs=True`." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft edges. The two accepted pairs are well supported as same-bug/doc-duplicate cases, and the rejected pairs avoid overclaiming shared root cause." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 5, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:41093", + "issue:41553", + "issue:42617", + "issue:42915", + "issue:42947", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43632", + "issue:43638", + "issue:43645", + "issue:43646", + "issue:43824", + "issue:43828", + "issue:43854", + "issue:43940", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44805", + "issue:44863", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:44991", + "issue:45313", + "issue:45325", + "issue:45584", + "issue:45588", + "issue:45663", + "issue:45685", + "issue:45823", + "issue:45834", + "issue:45902", + "issue:45923", + "issue:46402" + ], + "soft_pairs": [ + "issue:36331|issue:41093", + "issue:42947|issue:45923", + "issue:43940|issue:45313", + "issue:41553|issue:44991", + "issue:43854|issue:44863", + "issue:44928|issue:45588", + "issue:41093|issue:44805", + "issue:44479|issue:44560", + "issue:44863|issue:46402", + "issue:41553|issue:44488", + "issue:24643|issue:30064", + "issue:44530|issue:44863", + "issue:45325|issue:45902", + "issue:43366|issue:43828", + "issue:45663|issue:45823", + "issue:43366|issue:43531", + "issue:44155|issue:45584", + "issue:43950|issue:45834", + "issue:44560|issue:45685", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:43366|issue:44589", + "issue:43550|issue:44910", + "issue:43645|issue:43646", + "issue:44560|issue:44918", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:44661|issue:45823", + "issue:44560|issue:44805", + "issue:43632|issue:45834", + "issue:38175|issue:43994", + "issue:43638|issue:44661" + ], + "prepared_review_unit_hash": "acaa79f4964227d1522ad4e4eb03c600d87eaaf1e19d2c290994db6c1ca86393", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37482, + "estimated_input_tokens": 9371, + "estimated_eval_tokens": 18998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers regressions across loading, tokenization, attention, and training. Two likely duplicate mini-clusters stand out: the Transformers 5.0.0 custom model initialization breakage, and the SigLIP2 zero-probability/nonsensical-output regression.", + "confidence": 0.87, + "canonical_issue_reason": "issue:43646 is the clearest broad duplicate target for the Transformers 5.0.0 custom model initialization regression; issue:43645 looks like a narrower notebook-specific repro of the same breakage.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43994 is the strongest standalone issue to keep for the SigLIP2 cluster: it has a concrete model-level repro and matches issue:38175\u2019s zero-probability / nonsensical-output symptoms.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Different bugs: a custom trainer API kwarg mismatch vs a tensor/mask shape mismatch." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective LoRA checkpointing behavior and the other is a missing capability flag." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 loading failures, but they affect different Qwen model families and report different concrete failures." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer loading failures, but for different models and different failure modes." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model-loading regressions on different architectures; not the same underlying bug." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "One is a Qwen3.5 RLHF gradient explosion from 3D position_ids/SDPA behavior; the other is a flash_attention.py None-handling crash." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": false, + "reason": "Both are mask/index shape errors, but the reported shapes and contexts differ enough to indicate separate bugs." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both involve video on Qwen variants, but one is a broad video-input regression and the other is a StopIteration in a different embedding path." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "Separate loading problems for different model families and different code paths." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are model/tokenizer loading issues, but they involve unrelated models and symptoms." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Training with DeepSpeed weight-shape error vs image processor failing on void segmentation maps; unrelated." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttentionCache invalid group-type crash vs NemotronH checkpoint loading failure; different code paths." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Different Qwen3.5-related bugs: one is temporal position_id scaling, the other is uninitialized memory after meta-to-empty materialization." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF gpt-oss support vs Phi MoE autocast dtype mismatch; unrelated." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but different bugs: per-layer shared_kv_states rebuild vs device_map device mismatch." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "GGUF architecture support issue vs Qwen3-MoE sliding_window behavior; unrelated." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "AudioFlamingo3 track leakage vs Whisper empty-transcription handling; different bugs." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer corruption regression vs Kosmos2.5 long-OCR index error; different code paths." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3-vl video StopIteration vs MoE MPS histogram kernel not implemented for Int." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Missing import symbol for a specific class vs meta-device model loading regression." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "GenerationMixin import failure vs meta-device loading breakage; unrelated failures." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF support request vs Float8 storage lookup error; not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure vs Qwen3.5 Flash Attention illegal-memory-access bug from 3D position_ids." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same Transformers 5.0.0 regression: custom model initialization breaks, with 43645 just being a notebook-specific repro of the broader issue." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Different failures: video StopIteration in Qwen3-vl-embedding vs TRL SFT unpacking of Qwen3.5 embeddings." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py execution failure vs GGUF model architecture support; unrelated." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "Missing GenerationMixin import vs FineGrainedFP8Config failure in Qwen3Moe; different problems." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer mapping-name bug vs Gemma4 PLE device mismatch; unrelated." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different symptoms and contexts: video StopIteration vs mask shape mismatch." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "_is_hf_initialized regression vs Kosmos2.5 OCR indexing bug; unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same underlying SigLIP2-base-patch16-224 inference regression: one report shows zero probabilities, the other nonsensical outputs from AutoModel/pipeline." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed ZeRO-3/Bert index error vs tokenizer mapping-name failure; unrelated." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is mostly conservative, but one key claim is overstated: issue:43994 does not have downstream references in the packet (inbound_references is 0), so that part of the best-issue rationale is unsupported. Keep the 43994/38175 pairing as a likely duplicate cluster, but ground it only in the shared SigLIP2 regression and repro symptoms. The rest of the soft-edge verdicts look appropriately cautious." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 6, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43404", + "issue:43454", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43720", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44075", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44387", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44877", + "issue:44910", + "issue:44960", + "issue:44977", + "issue:44991", + "issue:45030", + "issue:45276", + "issue:45292", + "issue:45313", + "issue:45335", + "issue:45357", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45588", + "issue:45812", + "issue:46050" + ], + "soft_pairs": [ + "issue:44877|issue:45030", + "issue:43596|issue:44661", + "issue:44661|issue:44805", + "issue:43317|issue:44756", + "issue:38175|issue:39692", + "issue:43644|issue:44534", + "issue:42491|issue:43720", + "issue:41720|issue:42915", + "issue:44960|issue:44977", + "issue:42915|issue:44756", + "issue:44451|issue:44488", + "issue:45313|issue:45357", + "issue:42491|issue:45276", + "issue:43708|issue:44075", + "issue:44488|issue:44991", + "issue:43644|issue:43645", + "issue:39692|issue:43994", + "issue:45488|issue:45812", + "issue:43856|issue:43873", + "issue:43854|issue:43931", + "issue:43257|issue:43854", + "issue:43856|issue:44387", + "issue:45362|issue:45406", + "issue:44960|issue:46050", + "issue:42491|issue:45292", + "issue:42491|issue:43931", + "issue:44910|issue:45588", + "issue:44164|issue:44291", + "issue:45276|issue:45335", + "issue:43404|issue:43454", + "issue:41720|issue:44155" + ], + "prepared_review_unit_hash": "d27bfa787da9b42891026160a0bd4e3a2a46cc7621621cd9b672437f3bd59da2", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36446, + "estimated_input_tokens": 9112, + "estimated_eval_tokens": 18480 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a collection of unrelated bug reports across different models and subsystems, with a few clear duplicate subclusters around config-validation load failures, non-persistent-buffer initialization, and resize_token_embeddings regressions. The BERT-loading pair issue:44451 vs issue:44488 is not strong enough to merge on the available evidence.", + "confidence": 0.79, + "canonical_issue_reason": "issue:45276 is the best canonical for the resize_token_embeddings subcluster: it states the core bug most generally, and it aligns with the related Gemma/T5Gemma reports about embeddings not updating after resizing.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45488 is the strongest standalone report overall: it describes a broad, silent tokenizer regression that breaks a major model family and has clear user impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Related loading failures, but the packet does not establish the same root cause; this is too speculative to merge as duplicates." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Different bugs: a DeepSpeed/BertModel zero-sized tensor error versus add-new-model-like failing on TOKENIZER_MAPPING_NAMES." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "No shared code path is evident; tokenizer-mapping registration failure and tensor mask shape mismatch are different failures." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Different problems: dequantized-model offload loading failure versus a Strix Halo mmap/OOM workaround." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but different concrete failures: zero probabilities at inference versus docs/example model-processor and quantization errors." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Both describe the same Transformers v5 non-persistent-buffer initialization regression where buffers get filled with junk." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Different failure modes: qwen3_moe LoRA incompatibility across versions versus BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both involve Qwen3, but one is auto device mapping cuda asserts and the other is FineGrainedFP8Config loading; not the same bug." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms: GLM5 versus Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Unrelated: FineGrainedFP8/Qwen3Moe issue versus mmap/OOM on Strix Halo." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading failures for BERT-like models, but the packet does not establish a shared root cause or the same failing code path." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Different code paths: DeepSpeed ZeRO-3 language_model loading versus incorrect visual-encoder keys in save_pretrained." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Different bugs: qwen3_moe LoRA/version incompatibility versus resize_token_embeddings not propagating to embeddings." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer resume/max_steps calculation bug is unrelated to SGD args not being used." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading failures, but one is a model load regression and the other is tokenizer loading for a specific repo; no clear same root cause." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Different regressions: junk-filled non-persistent buffers versus Jupyter custom-model initialization failure." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both mention SigLIP2, but one is docs/example breakage and the other is incorrect inference outputs; not the same concrete bug." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions: LlamaTokenizer pre-tokenizer override versus Granite token-id mis-generation." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Different areas: Qwen3 MoE training memory efficiency versus quantization offloading behavior." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different model-loading failures with different models and error types." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Different bugs: accelerate+deepspeed MOE weight conversion versus GLM-4.7-Flash loading in tests." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Training-memory inefficiency and int4 CUDA reserved-memory OOM are related only broadly, not the same bug." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different model/processor crashes: Qwen3.5 chat crash versus Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 issue versus quantized model broken after save/load are unrelated." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "Different bugs: qwen3_moe version incompatibility versus resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Different failures: qwen3_moe branch incompatibility versus Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both touch flash attention, but one is packed-sequence misinterpretation and the other is an s_aux=None AttributeError." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Different serialization/loading regressions: extra_state handling versus init_empty_weights unexpected argument." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Same resize_token_embeddings regression: both report target embeddings not being updated in related model variants." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both are lm_head tying problems, but they affect different model classes and are not clearly the same concrete fix." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Unrelated: cuda assert during auto device mapping versus AudioFlamingo3 batched inference token/embedding leakage." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "One soft-edge verdict is too strong for the evidence in the packet: issue:44451 vs issue:44488 are both model-loading failures, but the packet does not establish that they share the same root cause. The acceptance claim (\u201csame load regression affecting different BERT-like models\u201d) goes beyond the titles/metadata, so this is not conservative enough. The rest of the summary is mostly grounded, but this overstates one duplicate pair." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 7, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:42947", + "issue:43335", + "issue:43381", + "issue:43425", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43854", + "issue:43873", + "issue:43957", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44488", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44877", + "issue:44912", + "issue:44928", + "issue:44960", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:45859", + "issue:46315" + ], + "soft_pairs": [ + "issue:43425|issue:43957", + "issue:39692|issue:44960", + "issue:44402|issue:44488", + "issue:42947|issue:45704", + "issue:36010|issue:39404", + "issue:43873|issue:44387", + "issue:45663|issue:45759", + "issue:44291|issue:44756", + "issue:45663|issue:45859", + "issue:44155|issue:45823", + "issue:44292|issue:44912", + "issue:44387|issue:44756", + "issue:43335|issue:45663", + "issue:43957|issue:44756", + "issue:42947|issue:43381", + "issue:45276|issue:45292", + "issue:44164|issue:44756", + "issue:43854|issue:45216", + "issue:43381|issue:44928", + "issue:44479|issue:44877", + "issue:45684|issue:45698", + "issue:45588|issue:45663", + "issue:43646|issue:43708", + "issue:43645|issue:44164", + "issue:45292|issue:45335", + "issue:45812|issue:46315", + "issue:43828|issue:43957", + "issue:43425|issue:43828", + "issue:44402|issue:44451", + "issue:44589|issue:44756", + "issue:39404|issue:45701", + "issue:44683|issue:45823", + "issue:43611|issue:43645", + "issue:43425|issue:45823", + "issue:43425|issue:43716" + ], + "prepared_review_unit_hash": "27560eb294d5eed38b7f9336c7602e9072714cf7dce32354d4191b49c52ccb2d", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 35, + "serialized_chars": 37650, + "estimated_input_tokens": 9413, + "estimated_eval_tokens": 19082 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated Transformers issues grouped by loose textual similarity. The only clear duplicate family is the resize_token_embeddings regression affecting output embeddings / decoder embeddings in Gemma4/T5Gemma variants.", + "confidence": 0.86, + "canonical_issue_reason": "Issue 45292 is the best canonical issue because it states the core regression in the most general form: resize_token_embeddings does not update output embeddings. The neighboring reports are narrower model-specific manifestations of the same behavior.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45292 is the strongest global issue target because it is broad, concise, and describes the shared underlying bug rather than a single model variant.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Both are version/runtime breakages, but one is Torch compatibility and the other is meta-device model loading; they do not describe the same bug." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Completely different topics: SigLIP2 documentation errors vs. a GLM5 model-loading issue." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading complaints, but the model names and failure modes differ; not enough evidence of the same underlying bug." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing with LoRA and an apex RMSNorm memory leak are unrelated bugs." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "ImportError for GenerationMixin and Whisper pipeline return_language are different failures with no shared code-path indicated." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both mention quantization, but one is offloading behavior and the other is increased reserved memory/OOM under int4; too broad to treat as the same bug." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states and Qwen multimodal config unwrapping are different model-integration problems." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Unexpected _is_hf_initialized during init_empty_weights is unrelated to disabling mmap to avoid OOM." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Different models and failure modes: Gemma4 FSDP2 KeyError vs Qwen3_5Moe CPU offload incompatibility." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo3 batch leakage and Gemma4 device mismatch are unrelated issues." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Both involve quantization, but NVFP4 loading failure and MXFP4 fallback to bf16 are not the same bug." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Int4 quantization OOM and Strix Halo mmap OOM are separate memory issues." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and Gemma4 FSDP2 attention-state rebuilding are unrelated." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading breakage is not the same as Strix Halo mmap/OOM behavior." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective with PEFT LoRA and the other is incompatibility in eval mode; different bugs." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same underlying regression: resize_token_embeddings does not propagate updated sizes to model output embeddings; 45276 is a model-specific manifestation." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "save/from_pretrained extra_state handling and Strix Halo mmap OOM are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Loading GLM-4.7-Flash in tests and Qwen3.5 save_pretrained regression are different issues." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Gradient checkpointing in eval mode and RLHF NaN from 3D position_ids / SDPA fallback are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression in Qwen VL models and strict config blocking granite_speech are different problems." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Read-only permissions propagation during save_pretrained and loading the wrong custom module after save_pretrained are distinct save/load issues." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sinkless-model crash and Gemma4 FSDP2 shared_kv_states KeyError are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization in notebooks and Trainer max_steps calculation on checkpoint resume are unrelated." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook custom-model initialization and extra_state save/load handling are different code paths." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same resize_token_embeddings regression, with 45335 showing the decoder.embed_tokens variant of the same missing propagation." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Wrong token IDs from AutoTokenizer and slower convert_tokens_to_ids are different tokenizer regressions." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Autocast dtype mismatch for Phi-tiny-MoE and meta-device loading breakage are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility and autocast dtype mismatch are not the same bug." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Both are model-loading failures, but they concern different models and no common failure mode is shown." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup error and mmap OOM on Strix Halo are unrelated." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline return_language regression and tokenizer-version behavior change are different issues." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Base_model_prefix loading failure and notebook custom-model initialization are separate custom-model breakages." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch 2.10 incompatibility and Gemma4 PLE device mismatch do not share the same underlying bug." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch compatibility and Mistral-3 image-preprocessor dtype mismatch are different failures." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster summary is well grounded in the packet, and the soft-edge decisions are conservative. The only clearly supported duplicate family is the resize_token_embeddings/output_embeddings regression across 45276, 45292, and 45335; the other rejected pairs do not have enough evidence of a shared underlying bug." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 8, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:32090", + "issue:36032", + "issue:39404", + "issue:40990", + "issue:42915", + "issue:43299", + "issue:43421", + "issue:43425", + "issue:43452", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43716", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44756", + "issue:44805", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45759", + "issue:45859", + "issue:46050", + "issue:46285", + "issue:46326", + "issue:46402" + ], + "soft_pairs": [ + "issue:44530|issue:45084", + "issue:43906|issue:44530", + "issue:39404|issue:44912", + "issue:44265|issue:44898", + "issue:44479|issue:44805", + "issue:43906|issue:46326", + "issue:46050|issue:46326", + "issue:44589|issue:45084", + "issue:43901|issue:44509", + "issue:43425|issue:44918", + "issue:43299|issue:43931", + "issue:43994|issue:44079", + "issue:36032|issue:43452", + "issue:44164|issue:45759", + "issue:44671|issue:44991", + "issue:43596|issue:44683", + "issue:44991|issue:45081", + "issue:43856|issue:44756", + "issue:44164|issue:44479", + "issue:39404|issue:44671", + "issue:44737|issue:45859", + "issue:44488|issue:45081", + "issue:43582|issue:43632", + "issue:44661|issue:44683", + "issue:44488|issue:44987", + "issue:43611|issue:43644", + "issue:32090|issue:39404", + "issue:46285|issue:46402", + "issue:43716|issue:45084", + "issue:40990|issue:43421", + "issue:36032|issue:42915", + "issue:43643|issue:44530" + ], + "prepared_review_unit_hash": "23fca553bd9a56c734302a0a661de0566b842ed484b0b614b4ad4a5635c04fd1", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37406, + "estimated_input_tokens": 9352, + "estimated_eval_tokens": 18960 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a grab-bag of mostly unrelated Transformers issues, with several broad v5 regressions and many model-specific loading or runtime bugs. None of the soft pairs look like the same underlying defect closely enough to merge.", + "confidence": 0.93, + "canonical_issue_reason": "#43611 is the best canonical representative because it describes a broad, user-facing Transformers v5 loading regression that can subsume a family of similar breakages; the other issues are narrower, model-specific, or unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "#43611 is the strongest issue to anchor the cluster: it is broad, clearly regression-oriented, and about a core loading path rather than a niche edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Different failures: a Qwen3.5 PagedAttentionCache crash versus a tokenizer/template compilation TypeError." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "Unrelated code paths and symptoms; one is a reproduction of an older issue, the other is a cache/group-type crash on Qwen3.5." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline language-return regression and GPT-OSS MXFP4 quantization fallback are unrelated bugs." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export with torch_compilable_check is a different export-path bug than Perceiver image-size handling." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Different model families and failures: video-input regression versus a mask/index shape mismatch." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "One is a reproduction of an older issue; the other is a mistral_common tokenizer-specific continue_final_message bug." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized save/load corruption is not the same as the mistral_common final-message handling issue." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure and template-node compilation errors are different defects." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs-oriented, but they concern different pipeline deprecations and different documentation fixes." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch version incompatibility and Qwen3.5/TRL embedding unpacking are separate issues." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both involve Qwen3-VL, but one is a v5 loading regression for MoE models and the other is a specific weight-shape mismatch." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "SigLIP2 bad inference results and ModelOutput key assignment with None are unrelated." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "A T5 tokenizer method conflict is not the same bug as gguf_file breaking AutoTokenizer/AutoModel loading." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "save/from_pretrained extra_state handling and composite Qwen config unwrapping are different loading/saving paths." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT MLM prediction regression and EMBEDDIA tokenizer loading failure are different model-specific bugs." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "DeepSpeed ZeRO-3/BertModel initialization and compiled flex_attention on newer torch versions are unrelated." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer problems: EMBEDDIA load failure versus Mistral regex patch crashing on backend_tokenizer." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "MoE training memory inefficiency and Strix Halo mmap OOM are different memory issues with different causes." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "Extra_state save/load handling has nothing to do with the Qwen video-input regression." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline behavior and CamemBERT masked-LM predictions are separate model regressions." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "Both involve device/offload behavior, but the XLNet CPU arange bug and Qwen3_5MoE rotary embedding incompatibility are distinct." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Loading cjvt/sleng-bert and Mistral regex patching are different tokenizer-loading problems." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon warmup TypeError and the _is_hf_initialized v5 regression are unrelated." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer mapping-name failure and compiled flex_attention are on different subsystems entirely." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "They are both loading failures, but for different models and different failure modes." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 regressions, but one is about base_model_prefix loading and the other about non-persistent buffers being filled with junk." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer broadcast NoneType and Whisper return_language pipeline behavior are unrelated." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "CLIP text-model structure inconsistency and Sam3 text-encoder weights not loading are different model-specific issues." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Mistral-3 dtype mismatch during preprocessing/model use is unrelated to the template-node compilation error." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity quality on gpt-oss and tokenizer post-processor updates after runtime special-token changes are different bugs." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "T5 tokenizer method-name conflict and Qwen3Moe FineGrainedFP8Config failure are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "AutoConfig trust_remote_code field loss and Qwen3.5 PagedAttentionCache crashes are different code paths and symptoms." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 9, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:34567", + "issue:36032", + "issue:40444", + "issue:42175", + "issue:42915", + "issue:43089", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43452", + "issue:43482", + "issue:43540", + "issue:43541", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43746", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44206", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44530", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45507", + "issue:45584", + "issue:45987", + "issue:46178", + "issue:46315", + "issue:46326" + ], + "soft_pairs": [ + "issue:34567|issue:40444", + "issue:40444|issue:46326", + "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:44811|issue:45584", + "issue:42175|issue:46178", + "issue:42915|issue:44530", + "issue:43644|issue:43646", + "issue:44206|issue:44220", + "issue:43452|issue:43482", + "issue:43883|issue:44488", + "issue:36032|issue:43645", + "issue:43317|issue:43482", + "issue:44464|issue:45987", + "issue:45084|issue:45507", + "issue:44987|issue:44991", + "issue:44849|issue:45325", + "issue:43746|issue:46326", + "issue:43262|issue:43377", + "issue:43645|issue:44534", + "issue:43646|issue:44534", + "issue:43404|issue:44466", + "issue:44403|issue:45081", + "issue:42915|issue:45084", + "issue:43883|issue:44991", + "issue:43089|issue:45987", + "issue:43994|issue:44610", + "issue:43322|issue:43540", + "issue:43377|issue:43819", + "issue:44451|issue:45081", + "issue:44987|issue:45081", + "issue:44265|issue:44610", + "issue:44991|issue:46315", + "issue:42915|issue:44291", + "issue:44220|issue:44683" + ], + "prepared_review_unit_hash": "e57898660f82826c15f72c66d687e11995b0434f6c55b5fd3b38cbc45bbe74f4", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38323, + "estimated_input_tokens": 9581, + "estimated_eval_tokens": 19418 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is loosely centered on Transformers v5/v5.x regressions around model/tokenizer loading, init, and generation, but the listed issues are largely distinct bugs affecting different code paths and models. I found no soft-edge pair that clearly looks like the same underlying change/bug.", + "confidence": 0.34, + "canonical_issue_reason": "issue:43646 is the broadest representative: it describes a generic Transformers 5.0 custom-model initialization regression, which best matches the cluster\u2019s overall v5 load/init theme.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the most general and actionable issue in the set, with a clear regression framing and broader scope than the model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "Trainer token accounting vs multimodal iterable-dataset failure; unrelated bugs." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Different areas and failure modes: multimodal dataset handling vs mistral_common tokenizer messaging." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are model-load failures, but for different models and likely different tokenizer/model code paths." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch dynamo grouped_mm tracing error is unrelated to template-node compilation failure." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both involve Whisper, but one is batch_decode special-token handling and the other is empty-transcription generation after align_special_tokens." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Backend packaging omission is not the same as an unsupported backend-mapping error for tensorflow_text." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Different models and subsystems: FP8 config failure vs paged attention cache linear_attention crash." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Related to v5 initialization regressions, but one is junk non-persistent buffers and the other is custom model init failure." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Both are audio-feature related, but they point to different failures and code paths." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both mention GGUF loading, but one is a generic gguf_file/from_pretrained breakage and the other is a specific Qwen2.5-GGUF v5 failure." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Molmo tied-weights attribute error vs a separate model-loading failure; no shared underlying bug." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer method-name conflict and notebook custom-model init regression are unrelated." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "GPU/CPU offload dequantization failure is not the same as a Qwen2.5 GGUF loading regression." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked-generation inconsistency and StaticCache seq-length type mismatch are different bugs." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template compilation TypeError vs GraniteMoEHybrid invalid-method call; different concrete failures." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Different models and symptoms: one model fails to load, the other tokenizer loading fails for a different repo." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden-states bug vs Qwen2.5-VL RoPE position-id scaling bug; unrelated." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT adapter loading vs continue_final_message behavior in mistral_common tokenizer; not the same bug." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio chat-template sampling-rate default and MIMI padding-mask mismatch are different audio bugs." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but custom model init in notebooks is distinct from non-persistent buffer junk initialization." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Same reasoning: different init-related regressions with different observed failures." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Tied lm_head weights missing vs serialization differing by device are related broadly, but not the same concrete bug." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Logging/noise issue vs Mistral tokenizer regex patch crash; different scopes and failures." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "FP8-config model failure and init_empty_weights argument error are separate initialization/configuration bugs." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Molmo tied-weights attribute error is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation performance overhead is unrelated to StaticCache seq-length contract issues." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Nonsensical outputs from one model and processor/model input-size mismatch in another are distinct problems." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next segfault and Qwen3OmniMoe video-processing ValueError are different model-loading/processing failures." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "Both are audio reconstruction/comparison bugs, but they involve different models and different missing mechanisms (padding mask vs STE)." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Generic model load failure and Mistral regex-patch tokenizer crash do not share the same root cause." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different failure surfaces: loading a specific model repo vs tokenizer regex patch crash." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export failure with torch_compilable_check is unrelated to OmDet processor image-size mismatch." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer load failure and slow convert_tokens_to_ids performance regression are different bugs." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Different error classes and code paths despite both being v5-era regressions." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio fbank extraction issue and compiled flex_attention torch>=2.9 failure are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 10, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42915", + "issue:43296", + "issue:43381", + "issue:43404", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43742", + "issue:43792", + "issue:44062", + "issue:44079", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44479", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44829", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45020", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45479", + "issue:45538", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:46209", + "issue:46326" + ], + "soft_pairs": [ + "issue:45216|issue:45357", + "issue:44062|issue:44987", + "issue:36032|issue:43742", + "issue:46209|issue:46326", + "issue:43632|issue:44756", + "issue:42915|issue:44589", + "issue:44479|issue:45325", + "issue:36032|issue:41720", + "issue:43404|issue:45127", + "issue:43531|issue:44877", + "issue:40444|issue:45325", + "issue:43611|issue:44534", + "issue:43296|issue:43531", + "issue:43425|issue:44403", + "issue:40990|issue:44811", + "issue:43792|issue:44589", + "issue:43632|issue:44534", + "issue:44079|issue:45663", + "issue:44403|issue:44991", + "issue:33666|issue:40444", + "issue:43421|issue:44568", + "issue:44683|issue:45020", + "issue:43381|issue:45923", + "issue:44291|issue:44589", + "issue:35707|issue:40444", + "issue:44462|issue:45812", + "issue:43482|issue:45538", + "issue:44829|issue:45479", + "issue:43482|issue:43525" + ], + "prepared_review_unit_hash": "4d2d30ce9149dbf23e51fb608a3aa96dbff1191552b5d42fff0a9d2456de563f", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "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": "Mostly unrelated issues were grouped together. I only accept two true duplicate pairs: the Qwen3.5 save_pretrained regression and the AutoTokenizer/tokenizer.json Granite regression. The rest are adjacent by subsystem or model family, but not the same underlying bug.", + "confidence": 0.92, + "canonical_issue_reason": "Issue 45812 is the broadest, highest-signal representative of a real duplicate set: it describes the tokenizer regression at the root level, while 44462 reads like a narrower symptom. It also has more inbound references.", + "canonical_pr_reason": null, + "best_issue_reason": "45812 is the best single issue to keep for this cluster because it captures the clearest cross-cutting bug and is the most likely root report among the few plausible duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Same Qwen3.5 save_pretrained regression: both report incorrect checkpoint serialization, with 45357 reading as a later/continued symptom." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "Both are tokenizer-related, but the concrete failures differ: AddedToken argument collision vs loading a specific model tokenizer in transformers 5.1." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Different models and different errors; one is a T5 tokenizer method conflict, the other is a key error loading MobileLLM-125M." + }, + { + "left": "issue:46209", + "right": "issue:46326", + "accept": false, + "reason": "Completely different code paths: FP8 inference on B200 vs mistral_common continue_final_message behavior." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "One is an _is_hf_initialized regression, the other is an mmap/OOM workaround on Strix Halo; not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Both involve FP8, but one is a Qwen3Moe FineGrainedFP8Config failure and the other is a missing Float8 storage type; distinct failures." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Same model family, but different regressions: video input handling versus still-image temporal position_id scaling." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "No shared underlying bug; tokenizer method conflict vs auto device mapping cuda assert." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both mention weights/tie_word_embeddings, but one is a missing lm_head tie in a model class and the other is LoRA merge collapse; not the same fix." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Different issues in different components: Qwen3-MoE sliding_window behavior vs strict config loading for granite_speech." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Same Qwen2.5-VL area, but one is iterable-dataset multi-image finetuning failure and the other is a rope/position-id scaling regression." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Unrelated regressions: base_model_prefix loading broke in v5, while non-persistent buffers are being filled with junk." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "Different model families and different failure modes; PaddleOCR-VL loading vs Qwen3-MoE sliding_window." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch version incompatibility is not the same as noisy load logs; these are separate problems." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "No shared root cause: perplexity issue on gpt-oss-20b vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper model runtime failure and Float8 storage lookup failure are unrelated." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Different v5 regressions affecting different internal mechanisms; not one underlying bug." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment bug and Gemma-4 FSDP2 shared_kv_states KeyError are distinct issues." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "One concerns load-time noise, the other tokenizer loading for est-roberta; not the same defect." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both are Qwen-VL training issues, but one is multi-GPU training and the other is iterable-dataset multi-image finetuning; different bugs." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Both are special-token related, but one is runtime post-processor updates and the other is add_special_tokens not inserting BOS/EOS for a specific tokenizer." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Different layers entirely: compiled flex_attention on torch >=2.9 vs remote_code loading regressions." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is eval-mode misuse and the other is a missing capability flag on a remote-code model." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Different failures: init_empty_weights gets an unexpected argument, while this one cannot find a Float8 storage object." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Different generation/training scenarios and different failure modes; not the same underlying bug." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": true, + "reason": "Same tokenizer regression family: 44462 says AutoTokenizer ignores repository tokenizer.json, and 45812 reports the resulting wrong token IDs for Granite models." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated issues: GGUF loading with transformers v5 vs CLIPTokenizer model_max_length semantics." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both involve degenerate training, but the root causes differ: flash_attention_3 behavior vs a loss/label setup issue." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "No meaningful overlap beyond being load-time failures; Qwen2.5-GGUF loading and Llama4Config pad_token_id are separate." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays mostly conservative. The two accepted pairs are plausibly the same underlying regressions based on their titles, while the rejected pairs are generally framed as adjacent-but-distinct bugs rather than definite duplicates. No clear overstatement stands out from the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 11, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:40444", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43531", + "issue:43582", + "issue:43606", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44877", + "issue:44898", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45491", + "issue:45542", + "issue:45701", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:46402" + ], + "soft_pairs": [ + "issue:43317|issue:43856", + "issue:43818|issue:46402", + "issue:42175|issue:45542", + "issue:41553|issue:43582", + "issue:44079|issue:44534", + "issue:43818|issue:45357", + "issue:43531|issue:44291", + "issue:41720|issue:45081", + "issue:43377|issue:45491", + "issue:40444|issue:45381", + "issue:43644|issue:44079", + "issue:44265|issue:45072", + "issue:43749|issue:44079", + "issue:45325|issue:45381", + "issue:43611|issue:43646", + "issue:43632|issue:45072", + "issue:42175|issue:43296", + "issue:43792|issue:43957", + "issue:43994|issue:44898", + "issue:45325|issue:45834", + "issue:43454|issue:44466", + "issue:42915|issue:43482", + "issue:43701|issue:45325", + "issue:44479|issue:45823", + "issue:44291|issue:44877", + "issue:43873|issue:45701", + "issue:43606|issue:44265", + "issue:43296|issue:45800", + "issue:43262|issue:43746" + ], + "prepared_review_unit_hash": "2e44c74017f62985395dfa0eba00dc985c726cfadf891b538a2b1aabeac13413", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35884, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are related only at a subsystem level and should stay separate. The only clear duplicate-family matches are the Transformers v5 model-initialization regressions (43611/43646) and, more tentatively, the Qwen2.5-VL position-id regressions (45325/45381). No PRs are present.", + "confidence": 0.69, + "canonical_issue_reason": "issue:43646 is the broadest representative of the Transformers v5 initialization regression family; it captures the underlying custom-model init breakage better than the narrower base_model_prefix symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the most general and reusable issue report among the accepted duplicate candidates, making it the best single issue to anchor triage.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different bugs: one is load/offload failure with dequantized models, the other is Qwen3 MoE training memory usage." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Unrelated model families and failure modes: Video-LLaVA weight sharing vs Sam3 text-encoder loading." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both involve TensorFlow availability, but one is packaging extras and the other is an undefined backend error; not the same bug." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Completely different areas: tokenizer error message vs Apple Silicon warmup TypeError." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "One is ModelOutput key assignment, the other is non-persistent buffer serialization; different code paths." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Different product areas: Video-LLaVA architecture bug vs Qwen3.5 save_pretrained visual-key regression." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and init_empty_weights/_is_hf_initialized are separate regressions." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Qwen3 device-mapping crash and Mistral tokenizer patch crash are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both are batching/padding-related, but MIMI encoder padding-mask mismatch and Gemma3 NaNs are different bugs." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Both touch Qwen2.5-VL, but multiple-images-per-prompt finetuning failure is not the same as video position-id regression." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Tied-weight serialization inconsistency and ModelOutput key handling are different issues." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export/compiler check failure and bfloat16 dtype mismatch are unrelated." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "FSDP CPU RAM efficient loading is a loading-path bug, not a ModelOutput-key issue." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": true, + "reason": "Same Qwen2.5-VL visual position-id bug family; both report incorrect rope/vision position ids in the video/image path." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": true, + "reason": "Same Transformers v5 model-init/load regression; base_model_prefix failures and custom initialization failures are the same underlying breakage." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "_is_hf_initialized handling in v5 and dtype mismatch in inference are different regressions." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Package/backend installation issue vs vLLM model-load failure; not duplicates." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper runtime failure and meta-device loading breakage are distinct problems." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Different vision-model symptoms: SigLIP2 nonsense outputs vs Perceiver non-default resolution failure." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Both mention OCR/video-like inputs, but Qwen2.5-VL position ids and Kosmos2.5 long OCR index errors are not the same bug." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both involve lm_head/tied weights, but one is a broken tie in model config and the other is device-dependent serialization inconsistency." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Different model-loading failures: Qwen3Moe FP8 config vs Qwen2.5 GGUF with transformers v5." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL position-id computation." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Both are device/offload-related, but Qwen video regression and Gemma4 PLE forward mismatch are separate bugs." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument and strict granite_speech config loading are different issues." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantization offload behavior and tokenization changes across versions are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU offload device mismatch and torch.export/compiler-check failure are different code paths." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "PaddleOCR-VL/vLLM load failure and torch/transformers version incompatibility are not the same bug." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sampling-rate default and GraniteSpeech PEFT checkpoint loading are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the soft-edge calls are conservative overall. The two accepted duplicate-family pairs are reasonably supported by the issue titles, and the many rejections avoid overstating similarity. No PRs are present, so the null PR fields are appropriate." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 12, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43720", + "issue:43749", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44186", + "issue:44265", + "issue:44466", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44610", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45584", + "issue:45701", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46291", + "issue:46305", + "issue:46315" + ], + "soft_pairs": [ + "issue:40990|issue:45584", + "issue:34634|issue:35707", + "issue:35707|issue:41720", + "issue:44568|issue:45920", + "issue:44186|issue:44265", + "issue:41720|issue:45823", + "issue:43606|issue:43819", + "issue:43454|issue:43550", + "issue:39692|issue:40444", + "issue:43881|issue:44488", + "issue:46305|issue:46315", + "issue:43377|issue:44610", + "issue:43454|issue:45072", + "issue:43257|issue:45923", + "issue:43572|issue:45759", + "issue:43317|issue:43873", + "issue:43377|issue:44898", + "issue:40990|issue:42947", + "issue:43322|issue:46050", + "issue:43454|issue:43720", + "issue:41720|issue:43322", + "issue:45325|issue:46315", + "issue:43749|issue:44534", + "issue:44991|issue:46291", + "issue:42617|issue:42915", + "issue:41720|issue:44977", + "issue:44466|issue:45357", + "issue:43526|issue:43883", + "issue:40990|issue:45701", + "issue:43881|issue:44991" + ], + "prepared_review_unit_hash": "6e157776a4ad77e721deb1c5b1a02fa5219f3331e063296532503f0eb32a4833", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36233, + "estimated_input_tokens": 9059, + "estimated_eval_tokens": 18374 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a grab-bag of model-specific regression reports, but the closest common thread is loading/offloading/quantization behavior. The best umbrella issue is the generic offloading + quantization report.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43873 is the broadest and most reusable issue in the set: it captures the shared offloading/quantization failure mode without being tied to one model, and it has enough discussion to serve as an umbrella.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest global representative because it is open, broadly phrased, and closest to the recurring loading/offload regressions that show up in nearby issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Different symptoms and domains: perplexity on GPT-OSS vs Whisper empty-transcription generation." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Unrelated models and bugs: Bark voice preset vs progressive generation with inputs_embeds/past_key_values." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Different code paths: generation with cached embeddings vs Qwen3 auto device mapping crash." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both tokenizer-related, but one is special-token insertion for a specific tokenizer and the other is wrong token IDs across many models; not the same concrete bug." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Different subsystems: LayoutLMv2 tokenizer padding/NER crash vs torch.export failure on torch_compilable_check." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both mention device_map=auto, but they are different models and different failure modes (assert vs PLE forward mismatch)." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Both are device-related, but one is CPU offload mismatch for Bark and the other is DAC latent/STE mismatch." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "Unrelated bugs: weight tying in AyaVision vs torch.compile/SDPA on Bamba." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "Different problems: SigLIP2 docs/model-processor mismatch vs Qwen2.5-VL finetuning with multiple images." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Both are load failures, but for different models and likely different causes; not enough to be the same bug." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "Completely different areas: streaming window masking vs tokenizer conversion performance." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Different model components: MIMI encoder padding-mask correctness vs OmDet-Turbo processor image size mismatch." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Different issues: tied weights/garbage generation vs bfloat16 dtype mismatch." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Different bugs: Qwen3 MoE weight conversion with accelerate/deepspeed vs missing gradient-checkpointing flag." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Different configuration problems: missing pad_token_idx vs from_config text_config unwrapping for composite multimodal configs." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Related offloading/quantization theme, but the reports are still distinct: dequantized model loading on gpu+cpu offload vs generic offloading not working as expected with quantization." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "Different model families and failures: MIMI batching/padding vs Perceiver image resolution/interpolation." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "Unrelated: perplexity regression vs gradient checkpointing ineffectiveness with PEFT LoRA." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Different bug classes: Llava Next load-time segfault vs sinq quantization save/load corruption." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Different code paths: lm_head weight tying vs packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Both are loading/runtime failures, but for different models and root causes." + }, + { + "left": "issue:45325", + "right": "issue:46315", + "accept": false, + "reason": "Different areas entirely: Qwen2.5-VL rope indexing vs tokenizer conversion speed." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "Both involve loading/state handling, but one is FSDP CPU-efficient loading and the other is non-persistent buffer initialization; not the same bug." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Unrelated: tokenizer loading regression vs kernels dependency requiring version/revision." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Different model/runtime problems: 3d_parallel.py execution vs Qwen3Moe FineGrainedFP8Config failure." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Both touch Qwen3-family inference, but one is device-map crash and the other is flash-attention generation quality/behavior; not the same concrete bug." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both are save_pretrained serialization issues, but they affect different tensors and different models." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "Unrelated: BeitImageProcessorFast label reduction vs Molmo tied-weights attribute missing." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Different symptom space: perplexity issue vs tokenizer behavior changes across versions." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Different failures: model load crash vs tokenizer load regression for a different repository." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 13, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42175", + "issue:42617", + "issue:42915", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43317", + "issue:43404", + "issue:43425", + "issue:43450", + "issue:43525", + "issue:43531", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43873", + "issue:44008", + "issue:44186", + "issue:44387", + "issue:44423", + "issue:44610", + "issue:44671", + "issue:44743", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45161", + "issue:45357", + "issue:45406", + "issue:45632", + "issue:45684", + "issue:45701", + "issue:45704", + "issue:45812", + "issue:45823", + "issue:46291", + "issue:46402" + ], + "soft_pairs": [ + "issue:44387|issue:45704", + "issue:36331|issue:42617", + "issue:43065|issue:46402", + "issue:44186|issue:45072", + "issue:44008|issue:45823", + "issue:43577|issue:43708", + "issue:36331|issue:42915", + "issue:44877|issue:44991", + "issue:43450|issue:45701", + "issue:43873|issue:44756", + "issue:43550|issue:43720", + "issue:43404|issue:45357", + "issue:42175|issue:46291", + "issue:44423|issue:45406", + "issue:41720|issue:42617", + "issue:43550|issue:43606", + "issue:44610|issue:44898", + "issue:44186|issue:44610", + "issue:42915|issue:43317", + "issue:43295|issue:45701", + "issue:45632|issue:45684", + "issue:43425|issue:46291", + "issue:41720|issue:45161", + "issue:36010|issue:42617", + "issue:43525|issue:46291", + "issue:44671|issue:44743", + "issue:45081|issue:45812", + "issue:43531|issue:43572", + "issue:43316|issue:44913" + ], + "prepared_review_unit_hash": "e57a32d3b6ab5f0bd3309d26800b2d2c94fcb84652066ca0a15c9ba4594e3b82", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35944, + "estimated_input_tokens": 8986, + "estimated_eval_tokens": 18228 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the soft pairs are mostly same-subsystem but not the same bug/change, so I reject them all. The broadest and most representative issue here is the open quantization/offload report, but there is no real duplicate canonical across the set.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 43873 is the broadest, most active, and most referenced report in this set, so it is the best anchor if one must be chosen; however, the cluster does not contain a true duplicate group.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 has the strongest combination of scope (offloading + quantization), discussion, and external references, making it the most suitable representative issue among otherwise unrelated reports.", + "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 leak; different code paths and fixes." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Trainer compute_loss signature breakage is unrelated to not being able to run 3d_parallel.py." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Both are Sam3-related, but one is a dummy Conv2d in the pixel decoder and the other is missing text-encoder weights." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer padding/NER crash and bfloat16 dtype mismatch are unrelated." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Different models and different failures: variable-name collision in Gemma3n vs device mismatch in Gemma4 PLE." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "BLIP2 dtype loading bug and Trainer resume/max_steps regression are separate problems." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Unexpected compute_loss kwarg and Qwen3 MoE FP8 failure do not share the same underlying bug." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading issues, but one is strict config handling and the other is tokenizer loading for a specific model." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batching shape bug is unrelated to tokenization changing across versions." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both involve memory/offload, but one is quantization offloading and the other is mmap/OOM on Strix Halo." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "torch.compile/SDPA failure and accelerate packed-weight loading failure are different code paths." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Mistral lm_head tying and Qwen3.5 visual encoder key saving are unrelated serialization/model issues." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Packaging the torch extra and kernels version/revision requirements are unrelated dependency issues." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both affect serving, but one is continuous-batching with multimodal inputs and the other is a Gemma4Processor attribute error." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both are image-resolution mismatches, but they affect different models and processor/model contracts." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer crash on NER/batched padding is unrelated to OmDet-Turbo image size mismatch." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3 MoE FP8 failure and dequantized model offload failure are different bugs." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Regression around processor.tokenizer/images is not the same as a general tokenization change across versions." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collision and read-only permission propagation are separate save/load issues." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch 2.10 incompatibility and kernels import/version requirements are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "cudaErrorAssert on auto device mapping is not the same as TP failing for GPT-OSS MoE." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error and 3d_parallel.py runtime failure are unrelated." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is unrelated to kernels import/version handling." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM regression and Qwen3.5 recurrent-state reset are different model-specific bugs." + }, + { + "left": "issue:45081", + "right": "issue:45812", + "accept": false, + "reason": "Mistral regex patch crash and Granite token-ID regression involve different tokenizer paths and failures." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and StableLm pad_token_idx regression are unrelated config issues." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct reload regression are different config bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 14, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:36331", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43322", + "issue:43329", + "issue:43388", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:44186", + "issue:44361", + "issue:44451", + "issue:44492", + "issue:44509", + "issue:44514", + "issue:44530", + "issue:44545", + "issue:44589", + "issue:44625", + "issue:44991", + "issue:45059", + "issue:45357", + "issue:45406", + "issue:45464", + "issue:45479", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:46291", + "issue:46402" + ], + "soft_pairs": [ + "issue:43540|issue:43782", + "issue:42915|issue:43322", + "issue:44186|issue:44361", + "issue:43421|issue:45812", + "issue:36331|issue:41720", + "issue:43526|issue:43881", + "issue:29159|issue:36032", + "issue:33357|issue:35707", + "issue:43388|issue:45059", + "issue:30064|issue:34634", + "issue:43550|issue:43653", + "issue:43606|issue:43720", + "issue:45406|issue:45464", + "issue:43873|issue:45704", + "issue:41720|issue:43746", + "issue:44530|issue:46291", + "issue:43421|issue:45701", + "issue:45357|issue:46402", + "issue:44492|issue:44509", + "issue:43606|issue:43653", + "issue:43531|issue:45759", + "issue:34634|issue:41720", + "issue:42490|issue:43329", + "issue:34567|issue:35707", + "issue:44514|issue:44545", + "issue:44625|issue:45479", + "issue:44451|issue:44991", + "issue:43531|issue:44589" + ], + "prepared_review_unit_hash": "7011f91bff7b82282dc749d02e763fe07497cdd0c91f0d241a4d63124e68e88f", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36231, + "estimated_input_tokens": 9058, + "estimated_eval_tokens": 18372 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues. The only clear duplicate is the Qwen2.5-VL `apply_chat_template` crash on batched input with `padding=False` (44514/44545).", + "confidence": 0.64, + "canonical_issue_reason": "Issue 44514 is the earliest and most explicit report of the only confirmed duplicate bug in the set, with a concrete reproducer and clear failure condition.", + "canonical_pr_reason": null, + "best_issue_reason": "44514 is the strongest representative because it is a precise, reproducible bug report and anchors the only accepted duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Both involve Qwen multimodal paths, but one is a video-input ValueError and the other is a `from_pretrained`/`weight_only` load error in a different class and code path." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Different models and failure modes: FineGrainedFP8 loading for Qwen3-MoE vs a Llava Next segmentation fault." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Tokenizer bugs, but for different tokenizers, tasks, and error mechanisms; not the same underlying issue." + }, + { + "left": "issue:43421", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer-related, but one is runtime post-processor synchronization and the other is a Granite token-id regression." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Unrelated: trainer loss signature mismatch vs Qwen3 auto device-mapping CUDA assert." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "ImageProcessor label reduction bug and glm-4v model loading failure are separate problems." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Both mention tokenizers, but one is slow/fast behavioral inconsistency and the other is a T5Tokenizer load-time attribute conflict." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS CLIP bus error and progressive generation with `inputs_embeds`/`past_key_values` are different code paths." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Metrics label-truncation bug vs SAM3 text/bounding-box behavior; no shared concrete defect." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Void segmentation map processing and BarkProcessor `voice_preset` are unrelated image/audio processor issues." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "`torch.compile`/SDPA failure in Bamba is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both touch loading/offload behavior, but they are different models and different bugs (CPU offload mismatch vs unpacking packed weights)." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Serving crash from missing `_tokenizer` and streaming inference failure on Qwen3.5 are distinct runtime issues." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization/offloading behavior and a T5 RMSNorm memory leak are not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Auto device mapping CUDA assert vs PEFT local checkpoint loading; different failure surfaces." + }, + { + "left": "issue:44530", + "right": "issue:46291", + "accept": false, + "reason": "PagedAttention cache group-type crash and `evaluate` import/version requirement are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:45701", + "accept": false, + "reason": "Runtime special-token updates and version-dependent tokenization are too broad and not the same concrete defect." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Both involve Qwen/Sam3 multimodal weight handling, but one is incorrect save_pretrained keys and the other is missing text-encoder weights on load." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A typo in cache-strategy docs and obsolete pipeline-task documentation are unrelated documentation issues." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "Device-mismatch on CPU offload is not the same underlying bug as a tokenizer mask-token registration problem." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and composite config `text_config` unwrapping are separate model/config issues." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor `voice_preset` and Qwen3 auto device mapping are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Trailing-slash import warning and undefined video-branch variables in multimodal token counting are different bugs." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counters and progressive generation with embeddings/past KV are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same processor, same method, same batched `padding=False` crash; this is a direct duplicate." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Qwen3.5 config propagation and a general zero-loss classification issue are not the same defect." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Different models and tokenizer load failures; not a duplicate." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and missing `Float8_e4m3fnStorage` are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is conservative and grounded in the packet. The only accepted duplicate pair, issue:44514 / issue:44545, is strongly supported by the nearly identical titles and shared failure mode, and the rejected soft edges are explained cautiously without overclaiming. No clear overstatement stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 15, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:42915", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43645", + "issue:43653", + "issue:43746", + "issue:43782", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44062", + "issue:44077", + "issue:44186", + "issue:44220", + "issue:44368", + "issue:44451", + "issue:44610", + "issue:44625", + "issue:44671", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45588", + "issue:45663", + "issue:45704", + "issue:45759", + "issue:45859", + "issue:45907", + "issue:45923", + "issue:46050", + "issue:46294" + ], + "soft_pairs": [ + "issue:44898|issue:45072", + "issue:44077|issue:44625", + "issue:45759|issue:45859", + "issue:43421|issue:43927", + "issue:45588|issue:45759", + "issue:44062|issue:44451", + "issue:44368|issue:45357", + "issue:44987|issue:45042", + "issue:44918|issue:44936", + "issue:43782|issue:45381", + "issue:39692|issue:44610", + "issue:43746|issue:46050", + "issue:42886|issue:44843", + "issue:45704|issue:45923", + "issue:43381|issue:43844", + "issue:43421|issue:43653", + "issue:45663|issue:45923", + "issue:45704|issue:45859", + "issue:41720|issue:45704", + "issue:43531|issue:43645", + "issue:43994|issue:44671", + "issue:43421|issue:44186", + "issue:43746|issue:43873", + "issue:45081|issue:45356", + "issue:39692|issue:43873", + "issue:43540|issue:46294", + "issue:43596|issue:44220", + "issue:42915|issue:43746", + "issue:45356|issue:45907", + "issue:30064|issue:36010" + ], + "prepared_review_unit_hash": "a0e70ac03b9be15595bee2d41bc7c934deab096870a253bcc7bafafe56b011f6", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37086, + "estimated_input_tokens": 9272, + "estimated_eval_tokens": 18800 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated Transformers bugs and feature requests. The soft pairs are mostly thematic lookalikes (tokenizers, Qwen, quantization, training), but they do not appear to be the same underlying defect or change.", + "confidence": 0.89, + "canonical_issue_reason": "No clear canonical issue: the items span unrelated subsystems and failure modes, so none is a strong duplicate representative for the whole cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "No single best issue emerges because the cluster is heterogeneous and the soft-paired items are not true duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Both are vision-related, but one is Perceiver resolution handling and the other is a bfloat16 dtype mismatch in different models/code paths." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Unrelated problems: patchtsmixer post_init policy vs Qwen3.5 num_labels propagation." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Both involve Qwen configs, but one is composite config unwrapping and the other is CPU-offload incompatibility in a rotary embedding path." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Token ID / special-token handling is broadly related, but one is runtime post-processor updates and the other is save/load config persistence." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Different code paths: flash-attention sink handling vs Qwen multimodal config unwrapping." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken argument duplication is a tokenizer construction bug; loading ScandiBERT is a separate model-loading failure." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "One is a warning about tie_word_embeddings, the other is incorrect visual-encoder key serialization." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Loading a specific model versus PIL image processors wrongly requiring torchvision; different failures." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Trainer/SFT embedding unpacking and evaluate-after-train are distinct trainer lifecycle bugs." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen vision issues, but one is weight_only loading and the other is incorrect video position IDs." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Documentation example errors for SigLIP2 are unrelated to OmDet-Turbo processor resolution mismatch." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "PEFT local checkpoint loading and quantized save/load corruption are separate persistence paths." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both mention HF_HUB_OFFLINE, but 44843 is specifically _patch_mistral_regex/model_info, not the general cache-load failure in 42886." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "A T5 RMSNorm memory leak is unrelated to a missing supports_gradient_checkpointing flag." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Gradient checkpointing in eval mode and gradient blow-up under ZeRO-3 are different problems." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both touch special tokens, but one is runtime post-processing updates and the other is BigBirdTokenizer mask-token registration." + }, + { + "left": "issue:45663", + "right": "issue:45923", + "accept": false, + "reason": "Gemma4 shared_kv_states rebuild bug is unrelated to a trust_remote_code flag omission." + }, + { + "left": "issue:45704", + "right": "issue:45859", + "accept": false, + "reason": "Different failures: apex FusedRMSNorm leak vs Qwen3_5MoeTextRotaryEmbedding CPU-offload incompatibility." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and T5 RMSNorm memory leak are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is separate from Jupyter custom-model initialization in Transformers 5.0." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Both are bad outputs, but they affect different models and likely different preprocessing/model-selection bugs." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Runtime special-token updates and LayoutLMv2 NER/padding crashes are not the same bug." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT checkpoint loading and quantization/offloading behavior are related only at a high level, not the same code path." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve fix_mistral_regex, but one crashes on backend_tokenizer access while the other is a codec-handling regression in a different tokenizer." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 docs/example mistakes are unrelated to quantization offloading behavior." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both are video-input issues in Qwen models, but they fail at different stages with different root causes." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "DeepSpeed zero3/BertModel indexing and audio fbank extraction are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43746", + "accept": false, + "reason": "FineGrainedFP8Config failure in Qwen3Moe is unrelated to GraniteSpeech local checkpoint loading." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regression in Kimi-K2.5 is unrelated to list_repo_files missing revision forwarding." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Void segmentation map processing and a GenerationMixin import error are completely different issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 16, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:30064", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43295", + "issue:43322", + "issue:43335", + "issue:43388", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43618", + "issue:43643", + "issue:43653", + "issue:43697", + "issue:43716", + "issue:43720", + "issue:43756", + "issue:43818", + "issue:43819", + "issue:43950", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44220", + "issue:44448", + "issue:44488", + "issue:44661", + "issue:44671", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44991", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45183", + "issue:45381", + "issue:45479", + "issue:45685", + "issue:45834", + "issue:45987", + "issue:46032" + ], + "soft_pairs": [ + "issue:43720|issue:43819", + "issue:43653|issue:43720", + "issue:44220|issue:44661", + "issue:43550|issue:45072", + "issue:43540|issue:44008", + "issue:43295|issue:45834", + "issue:43618|issue:43697", + "issue:44062|issue:44488", + "issue:43550|issue:43819", + "issue:45183|issue:45479", + "issue:44186|issue:44898", + "issue:43322|issue:45685", + "issue:43716|issue:44008", + "issue:43653|issue:43819", + "issue:43531|issue:43643", + "issue:43335|issue:43756", + "issue:43388|issue:46032", + "issue:43950|issue:45987", + "issue:44843|issue:45081", + "issue:43643|issue:44877", + "issue:30064|issue:36331", + "issue:44448|issue:44671", + "issue:41720|issue:44991", + "issue:43540|issue:45381", + "issue:43421|issue:44811", + "issue:43577|issue:43818", + "issue:44821|issue:45084", + "issue:39692|issue:45834" + ], + "prepared_review_unit_hash": "0ae1db94a7e5ad43226dc81a019702be45f0502d1d5159c79c951129dfd3952f", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35954, + "estimated_input_tokens": 8989, + "estimated_eval_tokens": 18234 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is highly heterogeneous: most items are unrelated bugs spanning tokenizers, multimodal processors, model loading, caching, and model-specific forward paths. No soft edge looks like the same underlying duplicate, though a few are loosely related by theme.", + "confidence": 0.91, + "canonical_issue_reason": "issue:44843 is the closest representative of the only obvious subcluster: a concrete tokenizer-loading regression with a direct code-path failure and a nearby related Mistral regex issue. Even so, the overall cluster is too mixed to treat it as a strong canonical duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44843 is the best single issue to anchor the cluster because it is a specific, reproducible regression in core loading logic and is the only item that clearly sits near another issue in the set. It is still not a safe umbrella for the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Different failures: accelerate loading of packed BitNet weights vs DAC latent reconstruction without STE." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token registration bug and BitNet packed-weight loading bug are unrelated." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction issue vs add-new-model-like failing inside TOKENIZER_MAPPING_NAMES; different code paths." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different model/runtime problems: torch.compile+SDPA on Bamba vs bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Video-input processing failure in Qwen3OmniMoe vs Gemma3n variable-name collision; not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor.tokenizer regression and long OCR index error are separate failures in different models." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIP attentions serialization issue and RTDetrV2 output divergence are different model behaviors." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken special-argument duplication vs loading a specific tokenizer model are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Both are model bugs, but one is compile/SDPA and the other is DAC latent/forward mismatch; not mergeable as one fix." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Validation error messaging for input_ids/inputs_embeds is unrelated to single-label classification zero-loss behavior." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "LayoutLMv2 tokenizer padding/NER crash and Perceiver image-classification resolution issue are distinct." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next loading crash and MPS histogram int support error are completely different subsystems." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Both touch multimodal inputs, but one is image-preprocessor dtype mismatch and the other is audio mask attribute collision." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Special-token registration in BigBirdTokenizer is unrelated to DAC forward/from_latents mismatch." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE and missing fields with trust_remote_code are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation and Smollm3 RoPE-layer dropping are different configuration bugs." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Metrics label trimming bug and Mamba2 cache wrong-results bug are separate code paths." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "Non-persistent buffer corruption on load is unrelated to StaticCache get_seq_length shape/type contract." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Same helper name, but different failures: offline model_info call vs backend_tokenizer AttributeError. Not the same underlying bug." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Missing fields from trust_remote_code config and strict granite_speech config rejection are separate loading issues." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Void segmentation maps in image processor and CustomTrainer compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Pegasus v4/v5 output differences and CamemBERT masked-LM prediction regressions are model-specific and distinct." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "A800 cudaErrorAssert with auto device mapping and tokenizer loading failure for est-roberta are different problems." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both involve video inputs, but they affect different models and different failures (processing error vs vision_position_ids mismatch)." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime special-token/post-processor update bug and Whisper batch_decode skip_special_tokens bug are not the same code path." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype loading mismatch and Video-LLaVA missing temporal attention/shared weights are unrelated." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and template-node compilation failure are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 doc-example errors and Kosmos2.5 long-OCR index error do not share the same underlying bug." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "Tokenizer special-token registration and DAC latent/forward mismatch are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Duplicate of the earlier pair: sliding-window MoE behavior is not the same as trust_remote_code config loading." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 17, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:42175", + "issue:43262", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43644", + "issue:43720", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43906", + "issue:43994", + "issue:44186", + "issue:44291", + "issue:44442", + "issue:44479", + "issue:44485", + "issue:44509", + "issue:44561", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44857", + "issue:44936", + "issue:44977", + "issue:45003", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45372", + "issue:45406", + "issue:45584", + "issue:45701", + "issue:45834", + "issue:45987", + "issue:46192", + "issue:46209", + "issue:46285", + "issue:46326" + ], + "soft_pairs": [ + "issue:43782|issue:44977", + "issue:45584|issue:45834", + "issue:44186|issue:44442", + "issue:43644|issue:43749", + "issue:43606|issue:45072", + "issue:44479|issue:45834", + "issue:43825|issue:43994", + "issue:44805|issue:44936", + "issue:34634|issue:36010", + "issue:43577|issue:43582", + "issue:44485|issue:44509", + "issue:44561|issue:45020", + "issue:45987|issue:46192", + "issue:28218|issue:29159", + "issue:43720|issue:45072", + "issue:44610|issue:45072", + "issue:44671|issue:45701", + "issue:43782|issue:45325", + "issue:43825|issue:43827", + "issue:43906|issue:46209", + "issue:43582|issue:44857", + "issue:43749|issue:44291", + "issue:45081|issue:45406", + "issue:42175|issue:45081", + "issue:45372|issue:46326", + "issue:43262|issue:43819", + "issue:34634|issue:36331", + "issue:45003|issue:46285" + ], + "prepared_review_unit_hash": "724dd2db3cc1034403828bc4e5e4ad17b0fdb1eb0fa923e567ab491ccb5865e9", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35678, + "estimated_input_tokens": 8920, + "estimated_eval_tokens": 18096 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous, with only one plausible duplicate pair in the tokenizer behavior area. The strongest representative issue is the broad tokenization-regression report, while the rest are distinct bugs across unrelated subsystems.", + "confidence": 0.28, + "canonical_issue_reason": "issue:45701 is the best umbrella representative: it is broad, recent, open, and captures the recurring theme of version-driven tokenization changes better than the more specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the most cluster-like issue overall; it has the broadest scope and the highest likelihood of subsuming adjacent tokenizer-regression reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Same model family, but different failure modes: loading with weight_only vs generation under flash-attention." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty transcription failure and Kosmos2.5 long-OCR index error are unrelated code paths." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "Tokenizer crash on NER padding/truncation is not the same as AutoTokenizer failing to load a tokenizer class." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer corruption and FSDP RAM-efficient loading are different bugs in different loading paths." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU-offload device mismatch and bfloat16 dtype mismatch are distinct failure modes." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Video-input regression for Qwen VL models is unrelated to Kosmos2.5 long OCR indexing." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "One is an error-message wording issue; the other is a model-output correctness bug." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask-shape IndexError during indexing is not the same as evaluate() failing after train()." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset and GenerationMixin import failure are unrelated issues." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "dtype propagation for BLIP2 and Apple Silicon warmup TypeError are separate bugs." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE discussion and docs mentioning removed pipeline tasks are not the same underlying defect." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "A specific helper removal breaking trust_remote_code is not the same as the broader remote_code regression report." + }, + { + "left": "issue:45987", + "right": "issue:46192", + "accept": false, + "reason": "StaticCache seq-length contract breakage and return_tensors being ignored are unrelated API bugs." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": true, + "reason": "Both point to tokenizer behavior divergence around added tokens and slow/fast tokenizer inconsistency, plausibly the same underlying whitespace bug." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Packed-weight unpacking during accelerate loading and bfloat16 dtype mismatches are different problems." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor output size mismatch and dtype mismatch are unrelated failure modes." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression is not the same concrete bug as the generic tokenization-change report." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen VL loading/weight_only error and RoPE temporal scaling are separate bugs." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "They share the v5 pipeline-removal theme, but one is runtime messaging and the other is documentation; not the same concrete issue." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated repro of an older issue and an FP8 inference failure on B200 are unrelated." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon allocator warmup and CUDA AMP loss crash are different hardware-specific problems." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both involve loading, but FSDP CPU RAM-efficient loading and init_empty_weights argument handling are distinct code paths." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Mistral regex patch crash and Gemma4Processor missing _tokenizer are different tokenizer/processor bugs." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Backend installation extras and tokenizer regex patching are unrelated." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Missing ReasoningEffort import and continue_final_message behavior are different mistral_common-related defects." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Sampling-rate defaulting and DAC latent/forward mismatch do not share the same bug." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice_preset and compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "Unsafe sys.modules access and CLIPTextModel structural inconsistency are different regressions." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative: it identifies one plausible tokenizer-related duplicate pair and treats the rest as distinct, which matches the titles. The representative issue choice is also supported by the broad scope of issue:45701. No soft-edge verdict appears to overstate the evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 18, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43454", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43638", + "issue:43645", + "issue:43650", + "issue:43653", + "issue:43696", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43746", + "issue:43819", + "issue:43854", + "issue:43906", + "issue:43927", + "issue:44060", + "issue:44368", + "issue:44387", + "issue:44554", + "issue:44589", + "issue:44610", + "issue:44617", + "issue:44811", + "issue:44912", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45020", + "issue:45042", + "issue:45084", + "issue:45459", + "issue:45538", + "issue:45600", + "issue:45803", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:43526|issue:43697", + "issue:43653|issue:44912", + "issue:44060|issue:44368", + "issue:43742|issue:45020", + "issue:45084|issue:45538", + "issue:43526|issue:44554", + "issue:43650|issue:43906", + "issue:43329|issue:45600", + "issue:44589|issue:45042", + "issue:43452|issue:44991", + "issue:43746|issue:44811", + "issue:43638|issue:44936", + "issue:43550|issue:44912", + "issue:44387|issue:44617", + "issue:43819|issue:43927", + "issue:46421|issue:46424", + "issue:43257|issue:43329", + "issue:43854|issue:44977", + "issue:36010|issue:43645", + "issue:43720|issue:44610", + "issue:38175|issue:44610", + "issue:39692|issue:44912", + "issue:36010|issue:36331", + "issue:43531|issue:45042", + "issue:45459|issue:45803", + "issue:43454|issue:43577", + "issue:42490|issue:43257", + "issue:43262|issue:44811", + "issue:43257|issue:43696" + ], + "prepared_review_unit_hash": "782bb380ad668c4c6d035e480ae83294f8ee870085a3970ece9311e96d2c4f24", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36612, + "estimated_input_tokens": 9153, + "estimated_eval_tokens": 18562 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous single-bug reports, not one broad duplicate set. The only strong duplicate-like pair is the Qwen tied-weights warning reports (44060/44368); the rest are too far apart in subsystem and failure mode to merge.", + "confidence": 0.72, + "canonical_issue_reason": "No single canonical issue cleanly covers the whole cluster; if forced, 44060 is the best anchor because it states the most concrete Qwen tied-weights warning bug and appears to match 44368.", + "canonical_pr_reason": null, + "best_issue_reason": "44060 is the best representative issue because it is the clearest, most actionable description of the only likely duplicate set in the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "Different models and failure modes: BEiT label reduction vs RT-DETR output drift." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Tokenizer special-token decode bug vs MXFP4 quantization fallback; unrelated code paths." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": true, + "reason": "Both describe the same Qwen tied-weights warning regression; one is the core warning bug and the other is the same warning surfacing during Qwen3.5 LoRA fine-tuning." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "MobileLLM key error is too specific to be the same bug as a broad remote_code-loading regression." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Compilation error on template nodes is unrelated to CLIPTokenizer model_max_length." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "No shared underlying bug; image processor label reduction and MPS attention correctness are distinct." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "A vague request for data is not the same as an isolated reproduction of another issue." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Both touch multimodal/model-mapping areas, but they are different concrete bugs with different symptoms." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to PIL backend torchvision requirements." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Both involve loading, but one is gguf/from_pretrained breakage and the other is tokenizer loading for a specific model." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "Local PEFT checkpoint loading and Whisper batch_decode skip_special_tokens are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Zero3/non-pretrained BERT indexing and trainer.evaluate after train are different trainer lifecycle bugs." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile+SDPA on Bamba is not the same problem as MXFP4 quantization loading." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention OOM, but one is int4 reserved-memory regression and the other is a model-specific memory exhaustion report." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC from_latents/STE mismatch and DiaConfig token ID persistence are unrelated." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache-related, but one is recompilation from lazy init and the other is max_cache_length semantics; not the same bug." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE weight conversion and sliding_window behavior are different Qwen3 issues." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "GLM-4.7-Flash load failure and Qwen3.5 flash-attention generation problems are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Importing GenerationMixin and Jupyter custom-model initialization are different regressions." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Packed-weight loading in BitNet and OmDet-Turbo processor size mismatch are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 zero probabilities and OmDet-Turbo input-size mismatch are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Documentation example errors and MXFP4 quantization loading are not the same underlying issue." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "ImportError for GenerationMixin vs unexpected compute_loss keyword are unrelated training/loading bugs." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding_window and PIL backend torchvision requirements are different code paths." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both involve exception handling, but they affect different components and different failure modes." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision weight tying and Blip2 dtype handling are separate model-loading defects." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Trailing-slash path warning and Qwen3 MoE weight conversion are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio apply_chat_template sample-rate default and Whisper batch_decode behavior are not the same bug." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MoE conversion and GPT-oss OOM are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The clustering is grounded in the packet and remains conservative. The single accepted soft edge (44060/44368) is plausible from the titles alone, and the rest are appropriately rejected as cross-subsystem or cross-failure-mode pairs. The summary does not overclaim a broader duplicate set, and the canonical/best-issue choices are presented cautiously rather than as certainty." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 19, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43329", + "issue:43450", + "issue:43452", + "issue:43540", + "issue:43582", + "issue:43606", + "issue:43643", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43824", + "issue:43927", + "issue:43931", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44560", + "issue:44568", + "issue:44756", + "issue:44821", + "issue:44849", + "issue:44912", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45137", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45405", + "issue:45488", + "issue:45538", + "issue:45701", + "issue:45759", + "issue:45859", + "issue:45907" + ], + "soft_pairs": [ + "issue:43643|issue:45759", + "issue:43329|issue:44479", + "issue:43329|issue:43450", + "issue:43606|issue:45859", + "issue:34567|issue:40990", + "issue:43749|issue:44756", + "issue:43257|issue:43931", + "issue:44568|issue:45356", + "issue:29159|issue:34567", + "issue:42915|issue:43742", + "issue:44291|issue:45042", + "issue:42915|issue:44821", + "issue:44987|issue:45538", + "issue:43452|issue:43582", + "issue:43697|issue:45357", + "issue:43927|issue:44912", + "issue:42915|issue:45538", + "issue:44849|issue:45701", + "issue:44568|issue:45488", + "issue:41720|issue:43540", + "issue:33357|issue:34567", + "issue:44479|issue:45381", + "issue:42175|issue:43824", + "issue:43317|issue:45538", + "issue:43540|issue:44479", + "issue:44062|issue:44991", + "issue:43540|issue:44560", + "issue:28282|issue:34689", + "issue:45137|issue:45405", + "issue:24643|issue:29159", + "issue:43742|issue:44291", + "issue:44987|issue:45907" + ], + "prepared_review_unit_hash": "0a5399bd818fa84c0ad5cca7a31f1a8efaa98e203c491883f1ab81843f89335c", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37325, + "estimated_input_tokens": 9332, + "estimated_eval_tokens": 18920 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are unrelated standalone issues; there is one likely duplicate cluster around the Qwen v5.3.0 video-input regression. No PRs are present.", + "confidence": 0.68, + "canonical_issue_reason": "issue:44479 is the broadest and most representative report of the Qwen video-input regression, naming the release and multiple affected model families.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the best standalone issue to anchor the cluster because it captures the shared regression surface more generally than the model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Different bugs: missing fields with trust_remote_code vs composite Qwen text_config unwrapping." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "Same underlying video-input regression in the Qwen VL path; 43329 appears to pinpoint the root helper bug behind the broader 5.3.0 regression." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "One is an undefined-variable bug in token counting; the other is a batched shape issue in video processors." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both mention CPU offload, but they affect different models and code paths." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Trainer token-count tracking is unrelated to perplexity on gpt-oss-20b." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Both concern loading/memory, but one is FSDP RAM-efficient loading and the other is mmap OOM on Strix Halo." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen loading problems, but one is accelerate/deepspeed weight conversion and the other is a Qwen3-VL shape mismatch." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Distinct tokenizer regressions affecting different models and different behaviors." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState token counting." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "FP8 fine-grained config failure is unrelated to MobileLLM key errors." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Init-empty-weights argument handling is unrelated to PIL/torchvision backend requirements." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Different areas: Qwen3Moe FP8 vs AutoImageProcessor URL loading." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Model-loading failure for one repo is not the same as CLIPTokenizer max-length bug." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "gguf loading issue is unrelated to AppleSilicon allocator warmup TypeError." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetrV2 output drift and Qwen3.5 save_pretrained key regression are different bugs." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Config save/load token IDs issue is unrelated to MXFP4 quantization fallback." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Qwen3Moe FP8 and CLIPTokenizer model_max_length are unrelated." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-states behavior and generic tokenization changes are not the same bug." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Different tokenizer regressions: special-token insertion vs LlamaTokenizer pre-tokenizer override." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different symptoms and models: CUDA assert on A800 vs video ValueError in Qwen3OmniMoe." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS CLIP bus error is unrelated to TrainerState token accounting." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Same Qwen video-input regression family; 45381 is a model-specific report within the broader 44479 regression." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "TensorFlow backend packaging issue is unrelated to Qwen2.5-VL import failure." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Device-map dequantized-model offload failure is unrelated to CLIPTokenizer max-length behavior." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both are video-related, but Qwen3OmniMoe ValueError is not clearly the same bug as the multi-model 5.3.0 regression." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "AddedToken special-argument conflict is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Different video-input failures in different Qwen models; not enough evidence they share one bug." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "PyTorch-missing import error is unrelated to Llama 3.2 model loading breakage." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error is unrelated to the PEFT version bump issue." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "DeepSpeed training weight-shape error is unrelated to tokenizer slow/fast inconsistency." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Both are loading-related, but one is a key error for a model and the other is an init_empty_weights argument regression." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Repo loading failure and list_repo_files revision forwarding are different Hugging Face Hub issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The summary stays conservative, correctly centers issue:44479 as the broad Qwen video-input regression, and the accepted soft edges are plausible from the titles alone. No PRs are present in the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 20, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:27957", + "issue:28282", + "issue:29127", + "issue:34567", + "issue:36331", + "issue:38175", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43404", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43531", + "issue:43577", + "issue:43582", + "issue:43643", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43906", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44488", + "issue:44589", + "issue:44737", + "issue:44792", + "issue:44821", + "issue:44898", + "issue:44913", + "issue:44964", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45479", + "issue:45588", + "issue:45600", + "issue:45920", + "issue:45987", + "issue:46121", + "issue:46399" + ], + "soft_pairs": [ + "issue:43577|issue:45072", + "issue:28282|issue:29127", + "issue:42915|issue:43749", + "issue:38175|issue:44898", + "issue:42175|issue:44488", + "issue:43257|issue:43756", + "issue:43582|issue:43782", + "issue:44737|issue:45588", + "issue:43493|issue:43994", + "issue:34567|issue:36331", + "issue:44291|issue:44821", + "issue:43404|issue:45987", + "issue:43526|issue:45479", + "issue:43756|issue:46399", + "issue:44062|issue:45081", + "issue:43906|issue:45381", + "issue:27957|issue:28282", + "issue:43756|issue:45325", + "issue:44913|issue:46121", + "issue:44792|issue:44964", + "issue:45042|issue:45081", + "issue:43452|issue:44488", + "issue:43742|issue:44589", + "issue:45357|issue:45920", + "issue:43531|issue:44821", + "issue:45405|issue:45600", + "issue:43643|issue:45375" + ], + "prepared_review_unit_hash": "c09a82e0820529e2c6e04409f8c52e95ecb0b4edef0a0dc6cb8c839c1ec42536", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36185, + "estimated_input_tokens": 9047, + "estimated_eval_tokens": 18350 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous. The only strong duplicate candidate is the SigLIP2 pair, where one issue describes the underlying HF-vs-JAX implementation mismatch and the other reports the same bad outputs through AutoModel/pipeline. No PRs are present.", + "confidence": 0.68, + "canonical_issue_reason": "issue:43493 is the best canonical issue because it points to the underlying SigLIP2 implementation discrepancy; issue:43994 appears to be the same bug surfacing as nonsensical inference output.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43493 is the strongest representative: it is root-cause oriented, specific to the affected model, and can subsume the user-facing symptoms in issue:43994.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype mismatches, but they affect different model families and code paths (Blip2/QFormer vs SwitchTransformers/TimmWrapperModel)." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Unrelated problems: one is a missing-PyTorch import error, the other is a layoutlmv3 validation/error-message clarity issue." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Different failures in different loading/training mechanisms; both involve advanced memory/loading settings but not the same defect." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Same broad model area (vision), but the symptoms and mechanics differ: zero probabilities vs non-default resolution/interpolation failure." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "One is about backend dependency packaging, the other about loading a specific model; no shared concrete bug." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Both are Qwen/SmolLM-related, but one is a MoE weight conversion/loading issue and the other is a RoPE layer-count/configuration bug." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Different subsystems: Apple Silicon allocator warmup TypeError vs Qwen3VL weight_only loading failure." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "One is a CPU-device placement bug in XLNet; the other is a flash-attention sink/s_aux crash. Not the same code path." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both are about SigLIP2 producing incorrect outputs; 43493 frames the implementation discrepancy, while 43994 reports the same underlying bug through AutoModel/pipeline usage." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "Both concern trainer behavior, but one is a state counter not updating and the other is an unexpected argument in custom compute_loss." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "Different loading failures: unexpected `_is_hf_initialized` argument versus inability to load AutoImageProcessor from URL." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "A tied-weight bug in Mistral3 is unrelated to StaticCache returning a tensor instead of an int." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "Both are classification/label-related, but one is an image processor label-reduction bug and the other is a degenerate loss issue for single-label classification." + }, + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Different rotary-related areas: one is a model architecture/config mismatch, the other is a kernel registration crash." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer construction failure vs Mistral regex patching backend attribute crash; they are separate bugs." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "One is just a reproduction of an older issue; the other is a specific Qwen2.5-VL position-id regression. Insufficient evidence they are the same bug." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "XLMRoberta Flash Attention support and a missing-PyTorch ImportError are unrelated." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both touch positional/rope logic in multimodal models, but they target different models and failure modes." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Different config serialization issues: rotary_pct reload regression vs list-handling TypeError in rope validation." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "A failed janus test and a Phi-4 multimodal load failure are distinct symptoms on different models." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Dependency requirement regression in image processors vs a tokenizer regex patch crash; not the same defect." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "GGUF loading breaks in tokenizers/CausalLM paths, while cjvt/sleng-bert is a separate model-loading issue." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "KeyError loading MobileLLM-125M is unrelated to the Float8 storage-object TypeError." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "One is a save_pretrained visual-key regression for Qwen3.5; the other is a tokenizer ID regression across many models." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior is unrelated to AutoImageProcessor URL loading." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "One is a version pin bump for PEFT; the other is a broken auto_mappings reference to removed configs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet: it stays conservative about the cluster being mostly heterogeneous and only treats issue:43493/issue:43994 as the plausible duplicate pair. The soft-edge rejections are cautious and consistent with the titles/body signals, and no PR-pair mergeability concerns apply because there are no PRs." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 21, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29127", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:38175", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43577", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43749", + "issue:43819", + "issue:43906", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44466", + "issue:44484", + "issue:44493", + "issue:44514", + "issue:44568", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44811", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44936", + "issue:45005", + "issue:45216", + "issue:45290", + "issue:45356", + "issue:45372", + "issue:45381", + "issue:45542", + "issue:45823", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46202" + ], + "soft_pairs": [ + "issue:44625|issue:44877", + "issue:43720|issue:44898", + "issue:43697|issue:45216", + "issue:24643|issue:29127", + "issue:41762|issue:44936", + "issue:44514|issue:45381", + "issue:44493|issue:46050", + "issue:44466|issue:45005", + "issue:43452|issue:43577", + "issue:43906|issue:46202", + "issue:43452|issue:44843", + "issue:30064|issue:33666", + "issue:43688|issue:43749", + "issue:43819|issue:44811", + "issue:45372|issue:46170", + "issue:28282|issue:32090", + "issue:44936|issue:45823", + "issue:43295|issue:45356", + "issue:28282|issue:33357", + "issue:44361|issue:44442", + "issue:43653|issue:44361", + "issue:44484|issue:44704", + "issue:38175|issue:43295", + "issue:45542|issue:46178", + "issue:24643|issue:27957", + "issue:43653|issue:44568", + "issue:43577|issue:43697", + "issue:44514|issue:45290", + "issue:44464|issue:44743" + ], + "prepared_review_unit_hash": "255ddaee4f7c666691da35767c5354b42d633564bcf41a37e8adc4b6d66bb15a", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36684, + "estimated_input_tokens": 9171, + "estimated_eval_tokens": 18598 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a mixed cluster of unrelated Transformers issues across model loading, tokenizers, processors, training, and quantization. Only one soft pair looks like the same underlying regression: the v5 tied-weights/save_pretrained serialization bug.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 45005 is the better canonical because it states the broader tied-weights regression in v5, while 44466 reads like a specific manifestation of that same save/serialization problem.", + "canonical_pr_reason": null, + "best_issue_reason": "45005 is the most representative issue in the only likely duplicate pair: it captures the shared root cause (tied weights breaking in v5) rather than one device-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Different bugs: Qwen3.5 label propagation vs granite_speech strict-config loading." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "Unrelated code paths: BitNet packed-weight loading vs Perceiver image classification interpolation." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "Different regressions in different models: RTDetrV2 output changes vs Qwen3.5 checkpoint save_pretrained corruption." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "DeepSpeed weight-shape crash and LayoutLMv3 error-message clarity are not the same bug." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Both involve training/loading, but one is Gemma3 ZeRO-3 loading and the other is trainer state after train()\u2014distinct failures." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Same family (Qwen2.5-VL) but different problems: chat-template batching vs video vision_position_ids." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Position-id warnings and sinq save/from_pretrained breakage are different serialization/config issues." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": true, + "reason": "Both describe the same tied-weights/v5 serialization regression; 44466 is a concrete device-dependent symptom of the broader 45005 issue." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "Tokenizer/model-from-pretrained GGUF breakage is unrelated to BLIP2 dtype remaining float32." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "A reproduction of one issue and a Gemma pipeline roundtrip tokenization bug are not the same underlying defect." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Different tokenizer loading failures: GGUF-related breakage vs offline model_info() call in mistral regex patching." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Image-segmentation map handling and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Aux-loss normalization and FSDP CPU RAM-efficient loading affect different subsystems and code paths." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC latent reconstruction mismatch is unrelated to Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Gemma 4 processor import failure and Windows ROCm torch.distributed issues are different environment/import problems." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Missing PyTorch import error and Trainer _gpu_broadcast_one NoneType failure are unrelated." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Trainer evaluate-after-train failure is not the same as Gemma4 device_map=\"auto\" forward device mismatch." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but one is processor/tokenizer API breakage and the other is Kimi-K2.5 codec/fix_mistral_regex behavior." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer loading failure are different tokenizer implementations." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBird special-token registration/empty decode is unrelated to MLukeTokenizer task errors." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard-size behavior and AutoProcessor cached_file kwargs are different issues." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "SigLIP zero probabilities and processor.tokenizer regression are unrelated." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "TensorBoard-without-TensorFlow backend error and tensorflow_text backend mapping error are distinct backend-registration problems." + }, + { + "left": "issue:24643", + "right": "issue:27957", + "accept": false, + "reason": "DeepSpeed training weight-shape error and XLMRoberta Flash Attention 2 are different failures." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "BigBird mask-token special-token bug is unrelated to mDeBERTa BOS/EOS add_special_tokens regression." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "BLIP2 dtype propagation and RTDetrV2 output drift are unrelated model bugs." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both hit Qwen2.5-VL chat templating, but one is batched padding=False and the other is assistant tool-call content handling." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Chunked generation with compiled forward and Qwen3.5 recurrent-state reset are different generation/runtime behaviors." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft pair, issue:44466 and issue:45005, is a plausible same-bug match based on their titles around tied weights / save_pretrained regression in v5. The other verdicts do not overstate overlap and are reasonably rejected as distinct issues." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 22, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:29159", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:34567", + "issue:36010", + "issue:41720", + "issue:43262", + "issue:43295", + "issue:43452", + "issue:43504", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43927", + "issue:44361", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44554", + "issue:44561", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44843", + "issue:44977", + "issue:44991", + "issue:45440", + "issue:45491", + "issue:45588", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46002", + "issue:46050", + "issue:46209", + "issue:46402" + ], + "soft_pairs": [ + "issue:29159|issue:33357", + "issue:44843|issue:44991", + "issue:43646|issue:43749", + "issue:43531|issue:43742", + "issue:43756|issue:46402", + "issue:43688|issue:45440", + "issue:43688|issue:45704", + "issue:45491|issue:46002", + "issue:43653|issue:43927", + "issue:43550|issue:43927", + "issue:43688|issue:45701", + "issue:44464|issue:44977", + "issue:44479|issue:44625", + "issue:43643|issue:44561", + "issue:43550|issue:44361", + "issue:41720|issue:44464", + "issue:44977|issue:46209", + "issue:32090|issue:34567", + "issue:43577|issue:45588", + "issue:43577|issue:44554", + "issue:43504|issue:43720", + "issue:44466|issue:46050", + "issue:32090|issue:33666", + "issue:43504|issue:43550", + "issue:43504|issue:44610", + "issue:33666|issue:36010", + "issue:43742|issue:44821", + "issue:45823|issue:46050", + "issue:43262|issue:43295", + "issue:43452|issue:44451", + "issue:43720|issue:43927", + "issue:43606|issue:43927" + ], + "prepared_review_unit_hash": "ba0f109bd830122856edf37b496a29081f9b34ea64cb20ec149f3a8b2cfd67bb", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36686, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a true duplicate cluster: the items span many unrelated transformer bugs (tokenizers, model loading, training, multimodal processors, device maps, quantization, etc.). All soft edges look like false positives rather than the same underlying bug/change.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue stands out because the set is heterogeneous and the proposed pairings do not describe the same root cause.", + "canonical_pr_reason": null, + "best_issue_reason": "If one issue must represent the cluster, issue 44843 is the strongest standalone candidate because it is concrete, reproducible, and clearly scoped to a specific regression (offline tokenizer loading).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Different subsystems and failure modes: tokenizer slow/fast parity vs a MacOS bus error on a CLIP model." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both involve tokenizer loading, but one is an offline hub access regression and the other is a specific tokenizer failure for a different model family." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Distinct problems: custom model initialization breakage vs FSDP CPU RAM efficient loading regression." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding window behavior and MobileLLM key errors are unrelated bugs." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Different model-specific loading/architecture issues; no shared code-path is evident." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Both are MoE-related, but they concern different implementations and different correctness defects." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization in MoE models is unrelated to a T5 RMSNorm memory leak." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "NaNs appear in different places with different triggers: sliding-window attention on GPU vs classification head init/checkout." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and DiaConfig token-ID persistence are separate issues." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "torch.compile + SDPA on Bamba is unrelated to config token-ID loss on save/load." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "A model-loss normalization bug is not the same as a broad tokenization-version behavior change." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Chunked generation with compiled forward and flash-attention generation regression are different code paths and symptoms." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video input regression for Qwen VL models is unrelated to num_labels propagation in Qwen3.5." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both mention trust_remote_code, but one is missing fields in AutoConfig output and the other is removal of an FX helper breaking models." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "A compile/SDPA model-forward bug is not the same as an MLukeTokenizer AttributeError." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Auto device mapping cuda assertions are unrelated to chunked generation discrepancies." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Flash-attention generation failure and FP8 inference on B200 are different hardware/path-specific bugs." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer NoneType broadcast error and TrainerState token counter not updating are unrelated Trainer bugs." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "Blip2 dtype loading and flash_attention s_aux=None crash are separate issues." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "A dtype persistence issue does not match an MPS attention correctness bug." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy-field model loading and packed-weight unpacking during accelerate loading are different failure causes." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Serialization inconsistency of tied weights is not the same as sinq quantized model reload breakage." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer broadcast failure and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "These are different model-loading/compile bugs with no shared root cause." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Beit legacy-field loading and OmDet-Turbo processor size mismatch are unrelated." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Multi-GPU training for Qwen2-VL has no meaningful overlap with GenerationMixin import failures." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "Key error loading one model is not the same as loading AutoImageProcessor from a URL." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Device-mapped forward mismatch and quantized save/load corruption are different bugs." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio processor chat-template sampling-rate defaulting and processor.tokenizer/image handling regressions are separate issues." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file-related from_pretrained breakage and inability to load a specific model are not the same defect." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "Packed-weight loading in BitNet and config token-ID loss in Dia are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch for bark-small is unrelated to DiaConfig token-ID persistence." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 23, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:33666", + "issue:36331", + "issue:39692", + "issue:43257", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43540", + "issue:43572", + "issue:43606", + "issue:43646", + "issue:43688", + "issue:43720", + "issue:43824", + "issue:43825", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44291", + "issue:44361", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44568", + "issue:44589", + "issue:44625", + "issue:44843", + "issue:44849", + "issue:44964", + "issue:44977", + "issue:45200", + "issue:45229", + "issue:45357", + "issue:45405", + "issue:45520", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:46002", + "issue:46192", + "issue:46209", + "issue:46424" + ], + "soft_pairs": [ + "issue:43994|issue:46192", + "issue:33666|issue:36331", + "issue:43525|issue:43572", + "issue:44849|issue:46209", + "issue:43278|issue:43688", + "issue:43720|issue:44361", + "issue:44625|issue:46002", + "issue:43493|issue:43825", + "issue:28218|issue:33357", + "issue:39692|issue:43493", + "issue:44964|issue:45405", + "issue:44849|issue:45357", + "issue:44568|issue:45812", + "issue:44464|issue:46424", + "issue:45200|issue:45229", + "issue:43540|issue:44977", + "issue:43493|issue:43844", + "issue:44488|issue:44843", + "issue:39692|issue:43295", + "issue:43278|issue:43844", + "issue:43334|issue:43572", + "issue:43646|issue:44466", + "issue:33357|issue:33666", + "issue:43257|issue:43299", + "issue:43334|issue:45759", + "issue:44589|issue:45520", + "issue:43824|issue:44589", + "issue:45229|issue:45823", + "issue:43927|issue:44361", + "issue:43504|issue:43606", + "issue:43572|issue:44291" + ], + "prepared_review_unit_hash": "69c312aa1ab65211528509f9ef9734804ac9c78137ec8db2541f773341ad6165", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37120, + "estimated_input_tokens": 9280, + "estimated_eval_tokens": 18816 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are a heterogeneous mix of unrelated tokenizer, configuration, model-loading, training, and runtime problems. None of the soft pairs look like the same underlying bug or change, so no duplicates should be merged here.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the set does not form one duplicate cluster and the pairs span distinct subsystems and failure modes.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a good global representative because the items are not variants of one bug; even the most similar titles refer to different code paths or models.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "Different problems: SigLIP2 model/pipeline output correctness versus multimodal `return_tensors` handling." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Both involve training, but one is Qwen2-VL multi-GPU training and the other is a `compute_loss` signature break in custom trainers." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Similar symptom class, but different configs and fields: `Llama4Config.pad_token_id` vs `StableLmConfig.pad_token_idx`." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-state behavior is unrelated to FP8 inference failure on B200." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Different bugs: embedding dtype drift during eval versus auxiliary-loss normalization in MoE models." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Unrelated subsystems: BitNet accelerate loading versus tokenizer/task AttributeError for MLuke." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "`num_labels` propagation in Qwen3.5 is not the same as NaNs in classification heads after checkout/init." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation mismatch is unrelated to a pipeline error message about translation tasks." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing regression and MacOS bus error are unrelated failures." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both mention SigLIP2, but one is docs-example issues and the other is an implementation discrepancy; not the same bug." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Model-loading failure for Phi-4 multimodal is unrelated to the PEFT version pin bump." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Different Qwen3.5 regressions: hidden-states output versus incorrect visual encoder keys in save_pretrained." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different models and different failure modes." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Chunked generation with compiled forward is unrelated to cache length handling for quantized cache." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Gemma4 text-only mm_token_type_ids defaulting and Gemma4 multi-GPU OOM are different issues." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3OmniMoe video-input processing and Qwen3.5 flash-attention generation are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 implementation discrepancy versus gradient explosion in ZeRO-3 are not the same code-path problem." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Model load failure for `cjvt/sleng-bert` is not the same as the offline-mode hub call in `_patch_mistral_regex`." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Documentation example errors and a regression in `processor.tokenizer`/image passing are distinct issues." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Embedding dtype mismatch during evaluation is unrelated to ZeRO-3 gradient inflation." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Different missing-field regressions in different models: Qwen3-VL `pad_token_id` vs StableLm `pad_token_idx`." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage is unrelated to device-dependent `lm_head.weight` serialization." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "MacOS bus error on CLIP loading is unrelated to Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen3 MoE loading, but one is accelerate/deepspeed weight conversion and the other is a v5 Qwen3VL loading regression." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Missing `pad_token_id` in Qwen3-VL is not the same as failing to unwrap `text_config` in composite multimodal configs." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Float8 storage deserialization and `flash_attn` import mapping on Python 3.13 are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Importing a missing Qwen2.5-VL class is unrelated to a Float8 storage lookup failure." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma4, but OOM under multi-GPU inference and device mismatch under `device_map=\"auto\"` are different failures." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "DiaConfig save/load token-ID loss is unrelated to MLuke tokenizer task errors." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Legacy-field preset loading for BEiT and CPU-offload device mismatch for Bark are different bugs." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Missing `pad_token_idx` in StableLmConfig and an `init_empty_weights` `_is_hf_initialized` TypeError are different model-loading regressions." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 24, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:15354", + "issue:28218", + "issue:29159", + "issue:29942", + "issue:30333", + "issue:32090", + "issue:34567", + "issue:37428", + "issue:38175", + "issue:42490", + "issue:43065", + "issue:43278", + "issue:43504", + "issue:43540", + "issue:43575", + "issue:43582", + "issue:43606", + "issue:43653", + "issue:43696", + "issue:43701", + "issue:43844", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44112", + "issue:44423", + "issue:44493", + "issue:44521", + "issue:44554", + "issue:44734", + "issue:44792", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45405", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45588", + "issue:45823", + "issue:46050", + "issue:46121", + "issue:46285", + "issue:46369" + ], + "soft_pairs": [ + "issue:43278|issue:43701", + "issue:44493|issue:46285", + "issue:43575|issue:45405", + "issue:43504|issue:43653", + "issue:44792|issue:45405", + "issue:43606|issue:44112", + "issue:43994|issue:44493", + "issue:43873|issue:46050", + "issue:45588|issue:46050", + "issue:44423|issue:44734", + "issue:43065|issue:46369", + "issue:29942|issue:30333", + "issue:42490|issue:43696", + "issue:45406|issue:46121", + "issue:43844|issue:44964", + "issue:28218|issue:34567", + "issue:29159|issue:32090", + "issue:44554|issue:45446", + "issue:45084|issue:45520", + "issue:43582|issue:44977", + "issue:15354|issue:38175", + "issue:43540|issue:45200", + "issue:44521|issue:45290", + "issue:45200|issue:45823", + "issue:45072|issue:45198", + "issue:37428|issue:43957" + ], + "prepared_review_unit_hash": "f3360c0e438f7552f80ad9a387970b0ac18664bb4cc595bd1b5a0bad49d26c90", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35581, + "estimated_input_tokens": 8896, + "estimated_eval_tokens": 18048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it spans tokenizer behavior, multimodal/chat templating, model loading, attention/flash-attn, training, serving, and environment/version issues. None of the proposed soft pairs look like the same underlying bug, so all should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "Issue 43873 is the best broad representative: it is open, heavily discussed, has multiple inbound references, and describes a user-facing quantization/offloading failure that is more central than the other, mostly narrow or CI-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the strongest single issue to anchor triage because it has the most active discussion and a broad symptom statement, but the cluster overall is not a duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Different failures: dtype change during evaluate vs resume-from-checkpoint key mismatch." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Unrelated symptoms: position-id warning vs CLIP class structure change." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Qwen2 tensor-parallel OOM and a PEFT version pin are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy-field model loading bug vs tokenizer special-token registration bug." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure is not the same as a released PEFT version mismatch." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "CPU offload device mismatch and a GraniteSpeech CI stale test are different problems." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 bad outputs and unexpected position-id keys do not share the same code path." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Both involve memory/quantization themes, but one is offloading behavior and the other is broken save/load for sinq quantization." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention s_aux=None crash is unrelated to sinq persistence breakage." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Serve multimodal .to() crash and KV-cache continuation indexing are different serving bugs." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and Sam3 litetext memory overflow are distinct issues." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures and MLflow failure-reporting are unrelated CI/product bugs." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Trailing-slash import warning is not the same as a CUDA OOM on GPT-oss-20b." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer and rope-parameter JSON list handling are separate errors." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient growth and Phi-4 multimodal model loading failure do not indicate the same defect." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer added-space behavior is unrelated to TrainerState token-count tracking." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency and _gpu_broadcast_one NoneType are unrelated subsystems." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness and AuxRequest PyTorch version checking are different code paths." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation error is unrelated to a flash_attn import KeyError on Python 3.13." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon warmup TypeError and Qwen3.5 flash-attention generation bug are not the same bug." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp support and SigLIP2 zero probabilities are wholly unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video-input processing failure and Gemma4 mm_token_type_ids defaults are different multimodal issues." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "All-zero assistant masks for multimodal inputs is not the same as tool-call messages with no content crashing." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Missing mm_token_type_ids defaults and device-map forward mismatch are separate Gemma4 problems." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "Bfloat16 dtype mismatch in inference and Wav2Vec2 save/tokenization failure are unrelated." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Flash-support import-name error and meta-device model loading breakage are different loading/import issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 25, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43295", + "issue:43334", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43950", + "issue:43957", + "issue:43975", + "issue:44186", + "issue:44466", + "issue:44554", + "issue:44610", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44987", + "issue:45127", + "issue:45200", + "issue:45325", + "issue:45356", + "issue:45381", + "issue:45412", + "issue:45520", + "issue:45563", + "issue:45823", + "issue:46178", + "issue:46192", + "issue:46209", + "issue:46285", + "issue:46365" + ], + "soft_pairs": [ + "issue:43493|issue:43701", + "issue:44186|issue:46192", + "issue:43232|issue:44849", + "issue:45823|issue:46192", + "issue:45381|issue:46209", + "issue:43295|issue:43450", + "issue:43295|issue:44610", + "issue:44554|issue:46365", + "issue:39692|issue:46192", + "issue:43572|issue:44987", + "issue:43295|issue:43873", + "issue:44554|issue:45356", + "issue:43572|issue:43957", + "issue:45520|issue:46178", + "issue:44466|issue:45127", + "issue:37428|issue:43824", + "issue:43950|issue:46285", + "issue:44898|issue:45412", + "issue:43550|issue:43844", + "issue:45200|issue:45325", + "issue:43334|issue:43525", + "issue:43450|issue:43575", + "issue:42915|issue:43824", + "issue:42490|issue:43493", + "issue:41720|issue:46209", + "issue:43824|issue:43975", + "issue:44792|issue:45563", + "issue:43388|issue:43526", + "issue:28218|issue:35141", + "issue:38175|issue:43493" + ], + "prepared_review_unit_hash": "f5836354460b9cfae4fe980ef14b5825c9327ffc3bd082fa93d065cd2ccea539", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37015, + "estimated_input_tokens": 9254, + "estimated_eval_tokens": 18764 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous issue cluster. The only plausible duplicates are two small subclusters: SigLIP2 behavior/parity reports, and missing `pad_token_id` on new config classes. The rest are clearly different bugs, import errors, or unrelated model/processor problems.", + "confidence": 0.74, + "canonical_issue_reason": "`issue:43493` is the broadest, most root-cause-oriented report in the SigLIP2 pair; it can plausibly subsume the more specific zero-probability symptom in `issue:38175`.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:43493` is the strongest representative of the only broad behavioral duplicate pair, with the clearest description of the underlying implementation discrepancy.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "Different problems: SigLIP2 parity/behavior vs checkpoint resume key mismatch." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "Both involve tokenization/input handling, but the bugs are different and not the same code-path issue." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Unrelated generation-state update bug vs Qwen3.5 hidden-state output bug." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Different subsystems: device-mapped Gemma4 forward mismatch vs `text_kwargs` ignoring `return_tensors`." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Qwen2.5-VL video position-id bug is unrelated to FP8 inference failure on B200." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Both are processor/multimodal issues, but they affect different code paths and symptoms." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Different processor regressions: `processor.tokenizer`/image passing vs OmDet-Turbo image-size mismatch." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness bug is unrelated to the Trainer import issue." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 doc/example errors are unrelated to silent `return_tensors` handling." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Missing `pad_token_idx` in StableLmConfig is not the same bug as the physical-intelligence/fast load failure." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor regression vs quantization offloading behavior are different issues." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "Attention correctness on MPS is unrelated to Kimi-K2.5 tokenizer regression." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Different model-loading regressions with different failure modes." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "Both are import/config errors, but they involve different backend mappings and unrelated symbols." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "One is serialization of tied weights; the other is LoRA merge collapse after vocab extension." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different missing names and different modules." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption on load is not the same as CLIP structure inconsistency." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver resolution failure and RT-DETR memory release are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile/SDPA failure is unrelated to ZeRO-3 gradient inflation." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Different multimodal bugs: Gemma4 token-type defaults vs Qwen2.5-VL rope/position-id scaling." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": true, + "reason": "Same underlying bug class: new config objects missing `pad_token_id`, causing AttributeError in the generation/loading path." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video processor shape bug is unrelated to tensor-parallel OOM." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "FineGrainedFP8Config failure is unrelated to the Qwen2_5_VL import error." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Trailing-slash local path warning is unrelated to SigLIP2 implementation parity." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3 auto device-mapping assert on A800 is unrelated to B200 FP8 inference failure." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "ImportError for a missing class and tokenizer detokenization regression are different bugs." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Failed janus image-generation test and stale warning logic are not the same issue." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both mention labels, but one is metric gathering and the other is image processor label reduction." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing bug and embedding reinitialization after resize are separate issues." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": true, + "reason": "Same SigLIP2 code-path problem: 38175 is a concrete symptom of the broader HF-vs-JAX implementation discrepancy reported in 43493." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The accepted duplicate pairs are plausible and mergeable into single fixes: the SigLIP2 discrepancy pair (43493/38175) and the missing pad_token_id config pair (43334/43525). The rejected pairs are generally well-justified by distinct titles/symptoms/subsystems, and I don\u2019t see evidence of overclaiming." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 26, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:29942", + "issue:32090", + "issue:37428", + "issue:38175", + "issue:43262", + "issue:43278", + "issue:43335", + "issue:43425", + "issue:43450", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43827", + "issue:43856", + "issue:43901", + "issue:44112", + "issue:44206", + "issue:44442", + "issue:44448", + "issue:44493", + "issue:44514", + "issue:44521", + "issue:44561", + "issue:44610", + "issue:44743", + "issue:44855", + "issue:44857", + "issue:45083", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45245", + "issue:45290", + "issue:45412", + "issue:45520", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:46170", + "issue:46192", + "issue:46234", + "issue:46285", + "issue:46326", + "issue:46402" + ], + "soft_pairs": [ + "issue:43278|issue:43856", + "issue:43526|issue:46192", + "issue:43653|issue:44112", + "issue:45290|issue:46326", + "issue:44442|issue:45198", + "issue:38175|issue:45412", + "issue:43262|issue:43450", + "issue:45520|issue:46170", + "issue:45759|issue:46285", + "issue:44514|issue:44521", + "issue:44610|issue:44857", + "issue:44561|issue:46234", + "issue:44206|issue:45584", + "issue:43819|issue:44112", + "issue:43827|issue:43901", + "issue:43819|issue:45083", + "issue:44857|issue:45593", + "issue:43611|issue:46402", + "issue:43335|issue:43550", + "issue:37428|issue:43425", + "issue:43606|issue:45083", + "issue:29942|issue:32090", + "issue:44448|issue:44743", + "issue:45200|issue:45588", + "issue:45127|issue:45245", + "issue:43335|issue:43653", + "issue:43819|issue:44493", + "issue:44855|issue:45083" + ], + "prepared_review_unit_hash": "de7c78fe8cdce43b4d840a134d0ceee598269e5cffc20a89b98c089056027867", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36338, + "estimated_input_tokens": 9085, + "estimated_eval_tokens": 18426 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose similarity bucket of mostly unrelated Transformers issues; I don\u2019t see any soft pair that looks like the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the items span distinct bugs across models, processors, tokenizers, training, CI, and import/runtime compatibility.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a good global representative for this cluster because the topics are too heterogeneous.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Different problems: dtype drift between train/eval vs MoE training memory usage." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "Different processor bugs: label reduction vs `return_tensors` being ignored when `text_kwargs` is passed." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Unrelated: BigBirdTokenizer special-token registration vs a GraniteSpeech CI device-override test." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both involve chat templating, but the concrete bugs differ: tool-call/empty-content crash vs `continue_final_message` on mistral_common tokenizers." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer failures for different models and code paths (`AutoTokenizer` load vs save/tokenization behavior)." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Completely different areas: SIGLIP output probabilities vs RT-DETR memory release/GC." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Different modality processors and bugs: audio chat-template sample-rate default vs batched video shape handling." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Different environment/import failures: flash_attn lookup on Python 3.13 vs Windows ROCm/torch.distributed dependency issues." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Both are model-structure regressions, but for different model families and different missing/unwrapped components." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Same general API, but different bugs: batched padding=False crash vs all-zero assistant masks for multimodal inputs." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Different detection issues: wrong processor image size vs AMP/CUDA crash in a loss function." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Both mention missing availability helpers, but they are different symbols causing different breakages." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Different speech bugs: unsupported `center` arg in a feature extractor vs empty-transcription generation failure." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "Different code paths: DAC latent reconstruction mismatch vs GraniteSpeech stale CI test." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs-related, but they refer to different pipeline APIs and different documentation mismatches." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "Different subsystems and symptoms: DAC latent behavior vs qwen3_omni_moe feature-length helper semantics." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Different detection-training bugs: AMP crash vs missing auxiliary losses when denoising is off." + }, + { + "left": "issue:43611", + "right": "issue:46402", + "accept": false, + "reason": "Different loading regressions affecting different models and versions." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Unrelated model/config bug vs torch.compile/SDPA failure." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Different compatibility issues: missing flash-attention helper import vs general Torch 2.10 incompatibility." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "Different problems in different areas: CPU-offload device mismatch vs feature-extraction length helper behavior." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Unrelated: Flash Attention 2 test failures vs Trainer `_gpu_broadcast_one` NoneType error." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Different model behaviors: Pegasus output regression vs Qwen recurrent-state reset with cache." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Different multimodal issues: missing default `mm_token_type_ids` vs flash-attention `s_aux=None` crash." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "Unrelated: LoRA merge/tied embeddings collapse vs category-count limit error." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Different model/config bug vs tokenizer special-token registration issue." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "Different failures: DAC latent mismatch vs unexpected key in position-id state dicts." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Different issues entirely: Python 3.13 import/torch.jit parsing vs qwen3_omni_moe helper behavior." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 27, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:28218", + "issue:29127", + "issue:29942", + "issue:32090", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:40444", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43746", + "issue:43825", + "issue:43854", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44360", + "issue:44448", + "issue:44493", + "issue:44561", + "issue:44779", + "issue:44857", + "issue:44977", + "issue:45072", + "issue:45083", + "issue:45229", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45459", + "issue:45478", + "issue:45588", + "issue:45593", + "issue:45823", + "issue:45859", + "issue:45920", + "issue:46192", + "issue:46209", + "issue:46234" + ], + "soft_pairs": [ + "issue:37428|issue:44351", + "issue:43550|issue:44112", + "issue:44206|issue:45823", + "issue:45290|issue:45381", + "issue:43335|issue:43504", + "issue:29127|issue:32090", + "issue:43278|issue:43825", + "issue:45588|issue:45859", + "issue:39692|issue:43526", + "issue:45310|issue:45478", + "issue:45459|issue:46234", + "issue:39692|issue:44493", + "issue:43335|issue:45083", + "issue:43720|issue:44112", + "issue:44448|issue:44779", + "issue:43335|issue:44112", + "issue:43335|issue:45593", + "issue:43606|issue:44857", + "issue:42490|issue:43854", + "issue:34689|issue:35141", + "issue:44360|issue:44561", + "issue:38175|issue:44493", + "issue:44857|issue:45072", + "issue:45920|issue:46192", + "issue:44448|issue:44977", + "issue:44977|issue:45229", + "issue:28218|issue:32090", + "issue:40444|issue:43526", + "issue:43746|issue:46209", + "issue:43701|issue:45083", + "issue:29942|issue:35141" + ], + "prepared_review_unit_hash": "be3f77fb01506dda01a53602ed0d11d8ebd4a78a581a33508732b40439543869", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37049, + "estimated_input_tokens": 9263, + "estimated_eval_tokens": 18782 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous, with one likely duplicate pair centered on the Qwen3.5 MoE `from_pretrained` regression (45310/45478). The rest of the items appear to be unrelated bugs, feature regressions, or documentation/test issues.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 45310 is the best canonical issue for the Qwen3.5 MoE `from_pretrained` bug because it is an open, concrete report and 45478 appears to describe the same underlying failure with essentially the same scope.", + "canonical_pr_reason": null, + "best_issue_reason": "45310 is the best representative issue: it is open, specific, and the earlier of the two highly similar Qwen3.5 MoE `from_pretrained` reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Different missing-import regressions in different code paths (`_flash_supports_window_size` vs `HybridCache`)." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba SDPA/torch.compile failure and GraniteSpeech CI flake are unrelated bugs." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Unsupported `center` arg in LasrFeatureExtractor is unrelated to Gemma4 device-map forward mismatch." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash and Qwen2.5-VL video position-id bug are different failures." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug and BEiT legacy-field load failure are unrelated." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and Trainer `_gpu_broadcast_one` NoneType failure are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "BF16/FP32 dtype drift and pipeline translation-warning behavior are different issues." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Flash-attention `s_aux=None` crash and Qwen3_5Moe CPU-offload incompatibility are distinct code paths." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 docs/example issues do not match the BEiTImageProcessorFast `reduce_labels` bug." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same Qwen3.5 MoE `from_pretrained` regression with the same version scope and no meaningful distinguishing behavior." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer protobuf-error masking and missing `is_falcon_mamba_ssm_available` export are unrelated API issues." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 documentation problems and unexpected position-id keys are different bugs." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "SwitchTransformers config bug and qwen3_omni_moe helper behavior are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Packed-weights loading bug and GraniteSpeech CI device-override flake are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus output regression and DeepSeek tokenizer regression are distinct model-specific problems." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "SwitchTransformers config bug and GraniteSpeech CI failure are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "SwitchTransformers sparse-layer bug and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch and LwDetr AMP crash are separate failures." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Local-path import warning and GLM-4.7 test loading issue are unrelated." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Llama 3.2 Vision loading regression and embedding reinitialization after resize are different bugs." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexer/ReLU discussion and `is_torch_fx_available` removal breakage are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 zero-probability behavior and unexpected position-id keys are distinct bugs." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "LwDetr AMP loss crash and dtype mismatches in SwitchTransformers/TimmWrapper inference are different issues." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "AutoTokenizer wrong-token-ID regression and `return_tensors` being ignored with `text_kwargs` are not the same bug." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus output regression and Qwen3.5 flash-attention generation issue are unrelated." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Flash-attention generation bug and Gemma4 multi-GPU OOM are different problems." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer added-space behavior and Trainer NoneType broadcast error are unrelated." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Qwen2.5-VL multi-image finetuning failure and BEiT reduce_labels bug are unrelated." + }, + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "Local checkpoint PEFT loading and FP8 B200 inference failure are distinct." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Resume-from-checkpoint key mismatch and qwen3_omni_moe helper behavior are unrelated." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 CI failures and embedding reinitialization after resize are different bugs." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster-level conclusion is too strong for the packet. The evidence supports that 45310 and 45478 are highly similar and likely about the same Qwen3.5 MoE `from_pretrained` bug, but the packet does not explicitly confirm them as duplicates. Phrases like \"only confirmed duplicate set,\" \"exact regression,\" and \"later duplicate\" overstate the evidence. The soft-edge rejections are generally conservative, but the duplicate verdict should be downgraded to likely/same-topic rather than confirmed." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 28, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:25251", + "issue:27957", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34689", + "issue:37428", + "issue:41720", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43412", + "issue:43550", + "issue:43575", + "issue:43653", + "issue:43873", + "issue:43927", + "issue:43976", + "issue:44206", + "issue:44448", + "issue:44466", + "issue:44479", + "issue:44561", + "issue:44568", + "issue:44743", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:44987", + "issue:45083", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45406", + "issue:45440", + "issue:45491", + "issue:45563", + "issue:45588", + "issue:45698", + "issue:45803", + "issue:45823", + "issue:46399" + ], + "soft_pairs": [ + "issue:32090|issue:33357", + "issue:44855|issue:46399", + "issue:43550|issue:45083", + "issue:43927|issue:45698", + "issue:42490|issue:43575", + "issue:10000|issue:15354", + "issue:44561|issue:44779", + "issue:37428|issue:44987", + "issue:43873|issue:45229", + "issue:43278|issue:43329", + "issue:41720|issue:45229", + "issue:44857|issue:45588", + "issue:27957|issue:34689", + "issue:44977|issue:45563", + "issue:44568|issue:45491", + "issue:45372|issue:45588", + "issue:43278|issue:43381", + "issue:43412|issue:45588", + "issue:45588|issue:45823", + "issue:45372|issue:45406", + "issue:44448|issue:44466", + "issue:43976|issue:45440", + "issue:45198|issue:45803", + "issue:25251|issue:32090", + "issue:45245|issue:45563", + "issue:43653|issue:44743", + "issue:44206|issue:44479", + "issue:32090|issue:33290" + ], + "prepared_review_unit_hash": "08388b72805ea8acd4665019f04a9816810c2d5ec72aca918c34de2c015d2ff8", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35784, + "estimated_input_tokens": 8946, + "estimated_eval_tokens": 18148 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is highly heterogeneous: the soft pairs share superficial vocabulary or broad subsystem overlap, but none look like the same underlying bug/change. No PRs are present, and I would reject every soft edge as a duplicate candidate.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45588 is the strongest standalone issue: it has a concrete crash, a specific file/function (`integrations/flash_attention.py`), and a clear repro path. It is actionable, but it still does not unify the other items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Both are runtime failures, but one is a Trainer broadcast NoneType error and the other is a MacOS bus error on a community CLIP model; different code paths and symptoms." + }, + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Both mention decorator/script parsing, but one is a Python 3.13 torch.jit import error and the other is a Qwen3.5 registration bug; not the same failure." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Different models and bugs: torch.compile/SDPA failure in Bamba vs a feature-length helper bug in qwen3_omni_moe." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load, but one is custom token IDs being lost and the other is the wrong custom module being loaded; distinct serialization bugs." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "A trailing-slash import warning is unrelated to a tensor-parallel OOM when loading Qwen2-57B-A14B-Instruct." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "Completely different models and failures: DistilBertModel issue vs torch.jit.script GeneratorExp/export problem for ViT." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "One is a removed API breaking trust_remote_code models; the other is tokenizer output corruption in Deepseek. Related release era, not the same bug." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "An import error from flash-attention utils and a model-loading failure for a specific repo are not the same code-path problem." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization/offloading behavior and multi-GPU OOM are different resource-management issues with different likely fixes." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "BF16-to-FP32 dtype drift in eval is unrelated to undefined variables in a video token counting helper." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "A cudaErrorAssert during auto device mapping is not the same as a multi-GPU OOM while loading Gemma4." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "AMP/CUDA loss crash in LwDetrImageLoss is unrelated to an AttributeError in flash_attention handling of `s_aux=None`." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "Both concern model loading/attention, but XLMRoberta Flash Attention 2 and Llama 3.2 90B Vision load break are distinct issues." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation regression and a stale warning for num_return_sequences do not share the same bug surface." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token insertion bug vs Gemma3 NaN embeddings on GPU; different layers of the stack." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 processor loading fails because a dependency import is missing, while the other crashes inside flash_attention on `s_aux=None`; different root causes." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Eval-mode gradient checkpointing restriction is unrelated to embedding dtype changes." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export issues for gemma3n are a different path from a runtime flash_attention AttributeError." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "A flash_attention crash and a Gemma4 PLE device mismatch are separate bugs with different fixes." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Missing `ReasoningEffort` import during processor loading is unrelated to a missing `_tokenizer` attribute in serve." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5 regressions, but one is output divergence and the other is serialization of tied weights; not the same concrete defect." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility and DeepseekV3MoE behavioral divergence are unrelated." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "A save_pretrained/tokenization failure is not the same as a bare-except bug swallowing exceptions." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "top_k output nesting in pipeline is unrelated to a Trainer GPU broadcast NoneType error." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-count limits and a stale warning in paged generate() are different issues." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "Tokenizer special-token registration and recurrent-state reset in a model file are unrelated." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported feature arg in an extractor and a video-input regression in Qwen3.5 are not the same bug." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Unsupported NoneType passed to `_gpu_broadcast_one` is unrelated to an OOM caused by Adafactor in DeepSpeed." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 29, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:35141", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43824", + "issue:43873", + "issue:43994", + "issue:44186", + "issue:44265", + "issue:44279", + "issue:44361", + "issue:44442", + "issue:44462", + "issue:44464", + "issue:44625", + "issue:44871", + "issue:44977", + "issue:45198", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45584", + "issue:45593", + "issue:45712", + "issue:45834", + "issue:46121", + "issue:46369", + "issue:46399" + ], + "soft_pairs": [ + "issue:43065|issue:45412", + "issue:44871|issue:44977", + "issue:44186|issue:45198", + "issue:43611|issue:46121", + "issue:42490|issue:43873", + "issue:33357|issue:39692", + "issue:44361|issue:45198", + "issue:29127|issue:34567", + "issue:43065|issue:43994", + "issue:45459|issue:45712", + "issue:42915|issue:42994", + "issue:43066|issue:44462", + "issue:29942|issue:30064", + "issue:45372|issue:45459", + "issue:45412|issue:46369", + "issue:43278|issue:43575", + "issue:43994|issue:45412", + "issue:44265|issue:45198", + "issue:43381|issue:43701", + "issue:43756|issue:45440", + "issue:43653|issue:44442", + "issue:43421|issue:45491", + "issue:27957|issue:29942", + "issue:29127|issue:33357", + "issue:43824|issue:44279", + "issue:43994|issue:46369", + "issue:45440|issue:45593", + "issue:43526|issue:45834", + "issue:44464|issue:45584", + "issue:46121|issue:46399", + "issue:43526|issue:44186", + "issue:29127|issue:35141", + "issue:44625|issue:45584" + ], + "prepared_review_unit_hash": "18c624fbb8ddcf6058e491a869c7ec5be175777bcfb4d96fbb1ca41c4b90089f", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37492, + "estimated_input_tokens": 9373, + "estimated_eval_tokens": 19002 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is very heterogeneous; most soft pairs are only keyword-similar (often sharing a subsystem like tokenizers, quantization, or attention) but do not describe the same underlying bug. No PRs are present.", + "confidence": 0.69, + "canonical_issue_reason": "issue:44462 is the cleanest, most self-contained regression in the set (AutoTokenizer ignoring tokenizer.json), while most other issues are unrelated, model-specific bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44462 has the clearest repro and narrow scope, making it the best representative anchor for this otherwise mixed cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Different problems: a dummy Conv2d placeholder in Sam3PixelDecoder vs RT-DETR memory not being released." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Same broad area only; eos_token_id mismatch in Gemma-3 is unrelated to Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizers and failure modes: LayoutLMv2 NER/padding crash vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Unrelated code paths: base_model_prefix loading regression vs rope validation type error from JSON-loaded lists." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Path-with-slash import warning and quantization offloading behavior are different bugs." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error with a community CLIP model is not the same as SigLIP2 documentation example errors." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizers and symptoms; MLuke AttributeError is not the same as Wav2Vec2 save/tokenization breakage." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error-message clarity and TrainerState token accounting are unrelated issues." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Different model bugs: Sam3 pixel decoder placeholder vs SigLIP2 producing nonsensical outputs." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer exception masking and leftover dummy classes are separate maintenance bugs." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Both mention quantization, but one is Qwen3Moe FP8 load failure and the other is quantized model saving; not the same code path." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Tokenizer decoder-type migration bug is unrelated to AutoTokenizer ignoring tokenizer.json from the repo." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void segmentation-map processing are different subsystems." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma 4 processor import failure is unrelated to protobuf error masking in tokenizer decoding." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "Both are memory-related, but one is a leak on deletion and the other is an OOM during sam3-litetext usage." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "BF16-to-FP32 embedding dtype change in eval is unrelated to tensor-parallel OOM on Qwen2-57B." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 incorrect outputs and RT-DETR memory release are unrelated bugs." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failure with torch_compilable_check is not the same as Wav2Vec2 save_pretrained/tokenization breakage." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing in eval and resume_from_checkpoint key mismatch are different issues." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Different model-implementation divergences: Smollm3 RoPE-layer dropping vs DeepSeekV3MoE remote/native mismatch." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBird tokenizer special-token registration bug is not the same as AutoTokenizer failing to load FastSpeech2ConformerTokenizer." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Updating tokenizers post-processors at runtime is unrelated to Gemma3 NaN embeddings with sliding-window attention." + }, + { + "left": "issue:27957", + "right": "issue:29942", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and failing FA2 tests overlap only by topic, not by concrete bug." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error-message issue and CLIP bus error on MacOS are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Qwen2_5_VL import error is a specific symbol/version issue, while the other is a broad dependency problem." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "SigLIP2 wrong outputs and SAM3 memory overflow do not share the same failing code path." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepSeekV3MoE divergence and D-FINE auxiliary-loss behavior are different model bugs." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and Kosmos2.5 long-OCR index error are unrelated image-processing failures." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is unrelated to Whisper empty-transcription failure after align_special_tokens." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "Rope-validation list typing and apply_rotary_pos_emb registration errors are separate implementation defects." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "ImageProcessor label handling and tokenizer crashes on NER inputs are different bug classes." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error messaging and post-init embedding reinitialization are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty transcription are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 30, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:33290", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:42175", + "issue:42490", + "issue:42915", + "issue:43262", + "issue:43316", + "issue:43334", + "issue:43381", + "issue:43452", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43611", + "issue:43646", + "issue:43701", + "issue:43720", + "issue:43723", + "issue:43761", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44206", + "issue:44387", + "issue:44442", + "issue:44464", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44977", + "issue:45003", + "issue:45072", + "issue:45245", + "issue:45440", + "issue:45479", + "issue:45491", + "issue:45588", + "issue:45859", + "issue:45907", + "issue:46234" + ], + "soft_pairs": [ + "issue:43761|issue:45859", + "issue:44442|issue:45072", + "issue:43575|issue:43701", + "issue:42915|issue:43931", + "issue:42490|issue:43381", + "issue:43575|issue:44387", + "issue:43526|issue:44857", + "issue:44977|issue:45245", + "issue:43334|issue:43531", + "issue:43262|issue:44206", + "issue:44464|issue:45491", + "issue:27957|issue:29127", + "issue:43452|issue:45588", + "issue:44743|issue:44811", + "issue:27957|issue:33290", + "issue:43646|issue:44077", + "issue:44062|issue:45907", + "issue:33290|issue:34689", + "issue:33290|issue:34567", + "issue:33290|issue:35141", + "issue:43381|issue:43873", + "issue:44442|issue:44898", + "issue:45003|issue:45440", + "issue:43723|issue:45479", + "issue:43761|issue:44079", + "issue:42490|issue:44743", + "issue:43316|issue:43525", + "issue:42175|issue:46234", + "issue:34689|issue:43611", + "issue:43720|issue:44442" + ], + "prepared_review_unit_hash": "31f961d01637e5125352a4c445b9e3bf2cd415ea1069820fbfdcbf8f310526c6", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36630, + "estimated_input_tokens": 9158, + "estimated_eval_tokens": 18572 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: most candidates only share broad themes like loading, memory, or config regressions, but not the same concrete bug. I would keep them separate and use the most central memory/offload regression as the representative issue.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43873 is the most discussion-heavy and broadly phrased memory/offloading regression in the set, so it is the best representative if one issue must stand in for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "43873 is the strongest global issue candidate because it has the most activity and sits closest to the recurring quantization/offloading/memory theme; however, the rest of the set is too diverse for true deduping.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "Different bugs: CLIP hidden-states regression vs Qwen3 MoE CPU-offload incompatibility." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading failure vs bfloat16 dtype mismatch in inference; unrelated code paths." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "OOM while loading a large model with TP vs resume_from_checkpoint key mismatch." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "FineGrainedFP8Config failure vs Qwen3-VL weight-shape mismatch; different loading problems." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Local-path import warning vs gradient-checkpointing-in-eval restriction; unrelated." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both involve memory, but one is TP load OOM and the other is int4 reserved-memory regression." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "ImageProcessor label reduction bug vs AMP/CUDA crash in loss computation." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Flash-attention generation bug vs category-count limit error; different failures." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id on Qwen3-VL config vs sliding_window behavior in Qwen3-MoE." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Chat-template sample-rate default vs unsupported center arg in feature extraction." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency vs NaNs from all-padding sliding-window attention." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "Flash Attention 2 support request vs a LayoutLMv3 error-message clarity issue." + }, + { + "left": "issue:43452", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer/model loading via gguf_file vs flash_attention sinkless AttributeError; different subsystems." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Recurrent-state reset bug vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "Flash-attention model support vs deepspeed Adafactor OOM; not the same bug." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "General custom-init regression vs patchtsmixer-specific post_init policy change." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "AddedToken special-arg duplication vs list_repo_files revision forwarding." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Adafactor/deepspeed OOM vs Llama 3.2 Vision loading breakage; different root causes." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "Memory/OOM issue vs TrainerState token-count update bug." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "OOM in deepspeed vs embedding reinitialization after resize_token_embeddings." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Eval-mode checkpointing restriction vs quantization/offloading memory regression; related area but different bug." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer load failure vs Perceiver image-classification interpolation failure." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe sys.modules access vs DeepseekV3 implementation divergence." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading in v5 vs degenerate zero-loss classification behavior." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIP hidden_states regression vs ModelOutput key assignment bug." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Trailing-slash import warning vs recurrent-state reset in cached generation." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy vs Llama4Config missing pad_token_id; similar family, different attribute." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Package extra/backend dependency issue vs missing import_utils symbol." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Model-loading regression for a specific model vs base_model_prefix-related v5 loading breakage." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight unpacking bug vs tokenizer load failure; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 31, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:28282", + "issue:42617", + "issue:42907", + "issue:43065", + "issue:43232", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43493", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43643", + "issue:43650", + "issue:43701", + "issue:43749", + "issue:43761", + "issue:43825", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44077", + "issue:44188", + "issue:44190", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44496", + "issue:44617", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:44987", + "issue:45137", + "issue:45216", + "issue:45290", + "issue:45632", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:46178", + "issue:46202", + "issue:46365", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:43232|issue:43388", + "issue:43931|issue:45216", + "issue:43761|issue:44188", + "issue:44077|issue:44479", + "issue:44464|issue:46421", + "issue:44987|issue:46178", + "issue:43493|issue:43827", + "issue:44442|issue:45803", + "issue:42907|issue:45698", + "issue:44479|issue:45290", + "issue:43873|issue:44871", + "issue:43749|issue:43931", + "issue:43643|issue:44913", + "issue:43316|issue:45759", + "issue:44190|issue:44496", + "issue:44855|issue:45137", + "issue:43650|issue:46202", + "issue:45632|issue:46424", + "issue:45803|issue:46365", + "issue:25251|issue:28282", + "issue:43295|issue:43825", + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:42617|issue:43575", + "issue:43065|issue:44617", + "issue:16998|issue:22355" + ], + "prepared_review_unit_hash": "2db86bf115ed068c7ff84163fa651cf060f9139a4ce10bdb9e4db8c91398913b", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35402, + "estimated_input_tokens": 8851, + "estimated_eval_tokens": 17958 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose basket of unrelated Transformers issues; none of the soft pairs look like true duplicates. The closest matches are still different bug classes or different code paths.", + "confidence": 0.95, + "canonical_issue_reason": "If one issue must serve as the cluster anchor, issue 43873 is the broadest and most central by activity/inbound refs, but it is not a duplicate of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the best global issue anchor because it is active, broadly scoped, and has the strongest discussion/reference signal in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwarg syncing after sync_gpus vs metric gathering truncating tuple labels; different subsystems and failure modes." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Model weight-shape mismatch on load vs save_pretrained regression for Qwen3.5 checkpoints; related family, different bug." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "CLIPVision hidden_states regression vs attention-kernel divergence under torch.compile; unrelated code paths." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Optional post_init policy for patchtsmixer vs video-input regression in Qwen VL models; no shared concrete defect." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked/compiled behavior, but one is wrong outputs and the other is spurious recompilation from lazy cache init." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Loading a specific model under transformers>=5.1 vs backend mapping missing tensorflow_text; unrelated failures." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation mismatch vs docs still referencing removed pipeline APIs; one is model parity, the other documentation." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer load failure for FastSpeech2ConformerTokenizer vs bare except swallowing errors in Fuyu batch conversion; unrelated bugs." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Save/dequantization problem for specific models vs wrong custom module loaded after save/load; both custom-model related but not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video-input regression for several Qwen VL models vs chat-template crash on tool-call assistant messages; different APIs and code paths." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior vs Gemma eos_token_id config inconsistency; no duplicate evidence." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "Broken FSDP CPU RAM-efficient loading vs a model weight-shape mismatch at load time; different root causes." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields from trust_remote_code AutoConfig result vs rotary_pct not persisting on reload; both config-related but distinct." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Gemma3TextConfig API mismatch vs AutoModelForCausalLM.from_config not unwrapping text_config for Qwen multimodal configs; different model families and fixes." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading failure in an example script vs unrecognized model_type in config.json; unrelated." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 import parsing error from torch.jit.script comment placement vs DeepSpeed ZeRO3 deque underflow; unrelated." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "Placeholder issue vs Gemma pipeline roundtrip failure on a specific token; clearly unrelated." + }, + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "trust_remote_code cache-path collision for local models vs cache max_cache_length only honored for quantized cache; different cache layers." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Bare except in Fuyu tensor conversion vs importing Trainer broken until datasets is uninstalled; unrelated." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k nesting bug vs AutoModel requiring PyTorch not installed; no overlap." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Custom code broke around processor.tokenizer/image passing vs v5 pipeline error-message wording; different regressions." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile + SDPA failure in Bamba vs resume_from_checkpoint key mismatch; unrelated." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch in bark-small vs checkpoint resume key mismatch; unrelated." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel.py execution failure vs tensor-parallel OOM when loading a large Qwen2 model; not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder vs Sam3Video CUDA OOM; same product area, but different concrete issues." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Model_max_length question for DeBERTa-V3 vs missing transformers.onnx import; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 32, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:27957", + "issue:32090", + "issue:41950", + "issue:43295", + "issue:43316", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43452", + "issue:43475", + "issue:43504", + "issue:43540", + "issue:43572", + "issue:43575", + "issue:43646", + "issue:43673", + "issue:43688", + "issue:43716", + "issue:43742", + "issue:43761", + "issue:44077", + "issue:44162", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44743", + "issue:44792", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:45127", + "issue:45200", + "issue:45406", + "issue:45542", + "issue:45584", + "issue:45698", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46234", + "issue:46369", + "issue:46421" + ], + "soft_pairs": [ + "issue:43575|issue:43742", + "issue:43540|issue:44514", + "issue:27957|issue:32090", + "issue:43316|issue:43716", + "issue:44610|issue:46369", + "issue:43316|issue:44877", + "issue:43295|issue:44871", + "issue:44871|issue:45584", + "issue:44162|issue:44792", + "issue:43404|issue:43441", + "issue:44077|issue:44534", + "issue:45803|issue:46170", + "issue:44871|issue:46050", + "issue:41950|issue:43475", + "issue:43761|issue:44479", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:43388|issue:43688", + "issue:45127|issue:45200", + "issue:43673|issue:44743", + "issue:43572|issue:46178", + "issue:44617|issue:45584", + "issue:43504|issue:45698", + "issue:43673|issue:46421", + "issue:44734|issue:45406", + "issue:45542|issue:46234" + ], + "prepared_review_unit_hash": "a485d0044123188b4c3718aefe92f05643deeb5f249902b847776323186df492", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 35010, + "estimated_input_tokens": 8753, + "estimated_eval_tokens": 17762 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are largely unrelated across different models and subsystems. The only plausible duplicate pair is the chunked-prefill cache regression in the generation path; the rest differ in concrete code path, model, or failure mode.", + "confidence": 0.83, + "canonical_issue_reason": "issue:46421 is the best canonical representative for the chunked-prefill cache regression because it names the underlying lazy-initialization problem directly; issue:43673 reads more like the narrower symptom report (missing cache during chunked prefill).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46421 is the strongest issue candidate overall among the soft-edge set: it is specific, actionable, and captures the shared generation/cache code path more clearly than the paired report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "Both are loading failures, but they involve different models and different root causes (TP OOM vs config key error)." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both involve Qwen multimodal/video processing, but one is a video-input processing bug and the other is a batched apply_chat_template/padding bug." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "Unrelated subsystems: FlashAttention support for XLM-Roberta vs Trainer GPU broadcast TypeError." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Different models and different issues: Gemma3TextConfig API mismatch vs Mistral-3 image preprocessor dtype mismatch." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "Both are SAM/vision-related, but one is a processor size mismatch and the other is a memory overflow; not the same bug." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-related, but they concern different models and different failures (API discrepancy vs strict config loading)." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "A regression in processor.tokenizer/image handling is unrelated to the Gemma-3 eos_token_id config inconsistency." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Different models and symptoms: Gemma-3 EOS config mismatch vs Whisper empty transcription failure." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 workflow breakage is unrelated to the janus image-generation test failure." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Same general Mistral family, but the concrete bugs differ: weight tying vs FlashAttention compatibility." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Optional post_init on patchtsmixer is unrelated to non-persistent buffer initialization in v5." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Different bugs and platforms: bare except in Fuyu tensor conversion vs Windows/ROCm torch.distributed import failure." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Gemma-3 EOS token inconsistency is unrelated to quantized model save/load breakage after sinq." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification processor lookup and SAM3 video missing attribute are different code paths and failures." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVision hidden_states regression is unrelated to Qwen VL video-input regression." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Both are model-loading issues, but one is gguf/from_pretrained handling and the other is a missing model_type in config.json." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Transformers 5 custom model initialization breakage is broader and different from GPTNeoXConfig rotary_pct not persisting on reload." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Different training/eval bugs: gather_for_metrics tuple truncation vs auxiliary-loss normalization." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings is unrelated to Gemma 4 mm_token_type_ids defaults." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache/recurrent state, but one is a GenerationMixin chunked-prefill cache initialization issue and the other is a model-specific recurrent-state reset bug." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLm pad_token_idx regression is unrelated to the tensorflow_text backend registration error." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Sam3Video OOM and Whisper empty transcription failure are unrelated symptoms in different models." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Legacy-field preset loading and wrong custom-module selection after save_pretrained are both load-time issues, but they are different bugs." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": true, + "reason": "Both describe the same chunked-prefill cache initialization regression; one reports a missing cache, the other the lazy-init/recompilation consequence, and one fix could plausibly address both." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "KV-cache continuation tensor indexing and Gemma4Processor missing _tokenizer are unrelated serving bugs." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "TensorBoard/TensorFlow backend detection and missing is_falcon_mamba_ssm_available are unrelated import-utils issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative on the soft-edge pairs. The only accepted duplicate pair, issue:43673 and issue:46421, is well supported by the shared chunked-prefill cache regression theme; the rest of the false verdicts are reasonably differentiated by model, subsystem, or failure mode. " + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 33, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:35141", + "issue:41720", + "issue:42490", + "issue:43065", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43644", + "issue:43761", + "issue:43827", + "issue:43931", + "issue:43976", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44190", + "issue:44242", + "issue:44462", + "issue:44464", + "issue:44479", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44792", + "issue:44855", + "issue:44877", + "issue:44913", + "issue:44933", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45245", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45632", + "issue:45907", + "issue:46234", + "issue:46369" + ], + "soft_pairs": [ + "issue:43525|issue:43531", + "issue:45381|issue:46369", + "issue:43761|issue:44242", + "issue:44792|issue:44977", + "issue:45563|issue:45584", + "issue:43931|issue:44060", + "issue:44792|issue:45600", + "issue:44190|issue:44964", + "issue:44534|issue:44913", + "issue:44479|issue:44734", + "issue:35141|issue:42490", + "issue:44855|issue:45588", + "issue:44617|issue:46369", + "issue:43827|issue:43994", + "issue:41720|issue:45245", + "issue:43065|issue:44610", + "issue:44521|issue:45381", + "issue:44464|issue:45563", + "issue:44933|issue:45362", + "issue:43644|issue:44913", + "issue:43976|issue:44855", + "issue:43550|issue:44190", + "issue:44077|issue:44877", + "issue:43540|issue:44734", + "issue:45127|issue:45357", + "issue:44462|issue:45907", + "issue:25251|issue:27957", + "issue:45632|issue:46234" + ], + "prepared_review_unit_hash": "756b80033388320a3318424fd8e44e8c06af536ad67239cd022469c56843969b", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 35901, + "estimated_input_tokens": 8976, + "estimated_eval_tokens": 18208 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a mixed bag of unrelated Transformers bugs and regressions. None of the soft-similarity pairs look like the same underlying issue/change, so no duplicates should be merged here. The most canonical standalone issue is the multimodal chat-template assistant-mask regression, which has clear repros and inbound references.", + "confidence": 0.84, + "canonical_issue_reason": "issue:44521 is the strongest standalone bug report: it is current, well-scoped, has multiple inbound references, and describes a concrete multimodal regression with clear user impact.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the best representative issue because it has the clearest bug description, the highest apparent external interest, and is actionable without needing broader context.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Both involve Qwen3-family models, but one is a missing `pad_token_id` attribute and the other is a `sliding_window` behavior problem; different failure modes and code paths." + }, + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Qwen2.5-VL video `vision_position_ids` corruption and Sam3-litetext memory overflow are unrelated multimodal issues with different symptoms and fixes." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden-states regression and MoE load-balancing loss omission are distinct model behaviors, not the same bug." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "A janus image-generation test failure and Qwen3.5 flash-attention generation failure are different model-specific issues." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "A stale `num_return_sequences` warning in paged generate is unrelated to Whisper empty-transcription failure after token alignment." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Shape mismatch when loading Qwen3-VL weights and an incorrect tied-weights warning in Qwen3-Next are separate loading/config bugs." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "A janus test failure and broken `auto_mappings.py` references to removed configs are not the same underlying problem." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading in an example script and model loading for Phi-4 multimodal instruct are unrelated." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk initialization and GPTNeoX `rotary_pct` reset on reload affect different subsystems." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Both are generation-related, but one is a video-input regression in Qwen multimodal models and the other is a KV-cache continuation indexing crash in `transformers serve`." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Output embedding reinitialization after resize and import warnings from trailing-slash local paths are unrelated bugs." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "A Python 3.13 `torch.jit.script` parsing issue and a flash-attention `AttributeError` on `s_aux=None` are different failure modes." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Sam3Video OOM and Sam3-litetext memory overflow are both memory issues, but they concern different models and likely different causes." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing removed `pipeline()` APIs is not the same as SigLIP2 producing nonsensical outputs through AutoModel/pipeline." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert and the 2^24 category limit error are unrelated problems." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "A dummy `nn.Conv2d` in Sam3PixelDecoder and a processor/model input-size mismatch are different issues, even within Sam3." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal regressions, but assistant-mask generation and video `vision_position_ids` are different pipeline stages and fixes." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency and a stale warning about `num_return_sequences` are not the same bug." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A nonexistent import from `image_utils` and a Qwen3.5 chat crash are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Global non-persistent buffer initialization junk and config reload losing `rotary_pct` are different classes of regression." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Python-version compatibility for Transformers 5.1.0 and a Python 3.13 decorator/comment parsing failure are not the same issue." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure and local dataset loading in an example script are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "Optional `post_init` on patchtsmixer and strict config rejection for granite_speech are distinct config/model-loading concerns." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input validation and serve KV-cache indexing are different code paths." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge collapse with extended vocab and incorrect visual encoder keys on save are separate serialization/training issues." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "AutoTokenizer ignoring repository `tokenizer.json` and `list_repo_files` not forwarding `revision` are unrelated Hub behaviors." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "Pipeline `top_k` nesting behavior and XLMRoberta Flash Attention 2 support are not the same bug or change." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "A local-model cache path collision and a missing `is_falcon_mamba_ssm_available` import utility are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 34, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:27957", + "issue:35141", + "issue:36246", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:43116", + "issue:43295", + "issue:43335", + "issue:43404", + "issue:43479", + "issue:43643", + "issue:43650", + "issue:43673", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44186", + "issue:44190", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44496", + "issue:44683", + "issue:44734", + "issue:44737", + "issue:44938", + "issue:44964", + "issue:44987", + "issue:45003", + "issue:45216", + "issue:45245", + "issue:45276", + "issue:45375", + "issue:45446", + "issue:45542", + "issue:45584", + "issue:45600", + "issue:45632", + "issue:45803", + "issue:46082", + "issue:46121", + "issue:46123", + "issue:46369", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:44112|issue:45600", + "issue:44737|issue:45803", + "issue:43335|issue:44190", + "issue:43116|issue:44190", + "issue:45276|issue:45584", + "issue:43295|issue:45276", + "issue:44938|issue:45542", + "issue:46082|issue:46121", + "issue:45632|issue:46421", + "issue:45245|issue:45584", + "issue:36246|issue:41628", + "issue:42175|issue:44279", + "issue:43404|issue:45216", + "issue:43650|issue:44220", + "issue:44246|issue:45003", + "issue:36246|issue:38617", + "issue:41950|issue:44016", + "issue:44186|issue:44734", + "issue:43673|issue:46424", + "issue:43479|issue:45375", + "issue:44683|issue:45446", + "issue:43643|issue:44496", + "issue:44964|issue:45600", + "issue:43854|issue:44964", + "issue:44016|issue:44220", + "issue:44112|issue:44964", + "issue:46123|issue:46369", + "issue:27957|issue:35141", + "issue:43976|issue:44987" + ], + "prepared_review_unit_hash": "bf4877cbe6ec624cd2224d7240f26e0682e6b132101ff1823774bf3d1a0edf95", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36513, + "estimated_input_tokens": 9129, + "estimated_eval_tokens": 18514 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it mixes unrelated issues across model configs, import/export errors, cache behavior, pipelines, notebooks, and CI/test failures. None of the soft pairs look like the same underlying bug or change, so there is no meaningful canonical duplicate here.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue fits this cluster; the items are not duplicates and span unrelated subsystems.", + "canonical_pr_reason": null, + "best_issue_reason": "No single best issue is suitable because the cluster does not represent one bug family or one fixable change.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both mention CI-like failures, but one is a GraniteSpeech test issue and the other is a broken auto-mapping reference for Sam3LiteText; different bugs." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet CPU device placement and a bare-except bug in Fuyu tensor conversion are unrelated code paths." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformers config construction and an image-classification script failing on local datasets are different problems." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Both are example/script bugs, but one is multi-label classification output handling and the other is local dataset loading in a different script." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Gemma4 embedding resizing and Whisper empty transcription after token alignment are unrelated model behaviors." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "A processor/tokenizer regression and Gemma4 embedding resizing are different failures." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Python 3.14 import failure and missing TensorFlow backend under torch install are distinct dependency issues." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Llama head_dim validation and rope-parameter list handling are unrelated config-validation bugs." + }, + { + "left": "issue:45632", + "right": "issue:46421", + "accept": false, + "reason": "Both involve cache, but one is a trust_remote_code path collision and the other is chunked-prefill recompilation from lazy cache init." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Category cardinality overflow and Whisper empty transcription are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different missing exports/classes and do not look like one shared bug." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "TensorFlow backend packaging under pip extras and a generic dependency issue are not the same concrete problem." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Mistral3 lm_head tying and Qwen3.5 save_pretrained checkpoint correctness are different model-specific bugs." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "A vague data request and an fbank feature extraction issue are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Slow import timing and unsafe sys.modules access are different import-related concerns, not the same bug." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Two separate missing-import symbols from different modules; similar symptom, but not one underlying defect." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "A video-classification pipeline bug and a notebook syntax error are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer crashes on NER padding/truncation and serve KV-cache continuation indexing are different code paths." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache-related, but one is missing cache in chunked_prefill and the other is max_cache_length handling for quantized cache only." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Different multimodal config bugs: default subconfig initialization vs a missing field dropped by strict validation." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both touch flex_attention and torch version checks, but one is a compiled flex_attention failure on torch>=2.9 and the other is an AuxRequest import guard; not clearly the same fix." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "trust_remote_code missing fields and unrecognized model_type are separate config-loading issues." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Both are model-loading problems, but one is Phi-4-multimodal and the other is Sam3LiteText auto-mapping breakage." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Different model-loading failures for different models and different failure modes." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax and audio feature extraction are unrelated." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI test flakiness in GraniteSpeech and Phi-4-multimodal loading are unrelated." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Partial-batch pipeline output dropping and memory overflow in sam3-litetext are different symptoms and code paths." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention support for XLMRoberta and embedding reinitialization after resize are unrelated bugs." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python-version compatibility and loading a specific model on transformers>=5.1.0 are different issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 35, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:28282", + "issue:35141", + "issue:38175", + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43526", + "issue:43531", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44242", + "issue:44279", + "issue:44336", + "issue:44368", + "issue:44496", + "issue:44521", + "issue:44617", + "issue:44792", + "issue:44933", + "issue:45042", + "issue:45103", + "issue:45200", + "issue:45216", + "issue:45325", + "issue:45335", + "issue:45357", + "issue:45372", + "issue:45405", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45600", + "issue:45759", + "issue:46082", + "issue:46209", + "issue:46234", + "issue:46399", + "issue:46424" + ], + "soft_pairs": [ + "issue:42907|issue:43441", + "issue:43531|issue:45600", + "issue:43526|issue:43901", + "issue:44792|issue:46424", + "issue:43295|issue:45335", + "issue:28282|issue:35141", + "issue:44496|issue:45042", + "issue:46082|issue:46399", + "issue:43441|issue:45588", + "issue:45372|issue:46234", + "issue:44016|issue:44062", + "issue:43299|issue:45216", + "issue:43901|issue:44496", + "issue:43824|issue:44933", + "issue:43673|issue:44336", + "issue:43381|issue:45593", + "issue:43976|issue:44279", + "issue:44075|issue:44368", + "issue:44496|issue:45759", + "issue:45335|issue:45584", + "issue:43650|issue:46209", + "issue:43577|issue:45216", + "issue:45103|issue:45372", + "issue:44242|issue:45200", + "issue:38175|issue:44617", + "issue:44521|issue:45325", + "issue:43299|issue:45357", + "issue:44279|issue:45405" + ], + "prepared_review_unit_hash": "1aa4d24ba2e5151a2a8559d0b05d52f1cad3b886e5d273a02cb6bf16ad43174b", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35662, + "estimated_input_tokens": 8916, + "estimated_eval_tokens": 18088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous and does not form one coherent duplicate set. The only strong duplicate-looking pair is the Qwen3.5 `save_pretrained` checkpoint serialization regression; the rest are distinct model/import/config/documentation issues across unrelated code paths.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45357 is the clearest statement of the underlying bug: Qwen3.5 `save_pretrained` is writing incorrect visual-encoder keys, which matches the broader checkpoint corruption report in 45216.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45357 is the best representative of the only real duplicate pair because it pinpoints the concrete faulty save path and the incorrect serialized keys.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Both involve Ministral/flash-attention, but one is about saving a dequantized model and the other is a runtime attention failure; different bugs." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior and a removed Sam3LiteText config reference are unrelated code paths." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and TextClassificationPipeline docs for `return_all_scores` are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "A janus image-generation test failure is not the same bug as cache-length handling for quantized cache." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Custom processor/tokenizer handling regression and t5gemma decoder embedding resize are different failures." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "PyTorch import missing and embedding reinitialization on resize are completely different issues." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model config and incorrect torchvision dependency checks are unrelated." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "LlamaConfig head_dim validation and Qwen3.5 rotary-pos-emb registration are different model/config bugs." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but one is a Ministral model crash and the other is an `s_aux=None` sink-less-model edge case." + }, + { + "left": "issue:45372", + "right": "issue:46234", + "accept": false, + "reason": "Missing `ReasoningEffort` import and missing `is_falcon_mamba_ssm_available` helper are unrelated import errors." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and `AddedToken` keyword collision are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Qwen3.5/Qwen3.6 model-loading failure and incorrect `save_pretrained` checkpoint output are different stages and bugs." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Pipeline docs mismatch and model-type detection failure are not the same issue." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Missing Qwen2.5-VL import and a nonexistent image_utils import are different import-path problems." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Generation cache missing in chunked prefill and ANSI code emission in loading_report are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction and D-FINE auxiliary-loss handling are unrelated training bugs." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python version compatibility breakage and a generic dependency issue are not the same concrete bug." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "SGD optimizer args being ignored and a tie_word_embeddings warning are unrelated." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Unrecognized model config and `from_config` not unwrapping `text_config` are different auto-model loading problems." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "t5gemma embedding resize and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "A vague data request and an FP8 inference-on-B200 bug are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation and Qwen3.5 checkpoint saving are different model behaviors." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring annotation handling and Gemma 4 processor import failure are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "MoE load-balancing loss logic and Gemma 4 multimodal token ids are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "SigLIP2 zero probabilities and Sam3Video OOM are unrelated model/runtime issues." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "All-zero assistant masks and Qwen2.5-VL rope-index scaling are different multimodal bugs." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3.5 loading failure is not the same as incorrect `save_pretrained` serialization output." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Generic dependency issue and a bumped-but-unreleased PEFT minimum version are different packaging problems." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Both report the same Qwen3.5 `save_pretrained` regression producing incorrect checkpoint contents, specifically around serialized keys." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 36, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:33290", + "issue:39692", + "issue:41950", + "issue:42994", + "issue:43295", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43479", + "issue:43525", + "issue:43650", + "issue:43653", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44291", + "issue:44297", + "issue:44351", + "issue:44361", + "issue:44462", + "issue:44466", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:45020", + "issue:45081", + "issue:45276", + "issue:45278", + "issue:45397", + "issue:45405", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712", + "issue:45812", + "issue:45823", + "issue:45834", + "issue:45907", + "issue:46132" + ], + "soft_pairs": [ + "issue:41950|issue:45020", + "issue:44351|issue:45278", + "issue:44779|issue:45812", + "issue:43295|issue:44779", + "issue:39692|issue:43901", + "issue:41950|issue:45823", + "issue:43479|issue:43525", + "issue:44683|issue:45542", + "issue:43976|issue:45405", + "issue:43299|issue:43854", + "issue:44462|issue:45081", + "issue:44297|issue:44462", + "issue:43450|issue:44279", + "issue:43650|issue:46132", + "issue:45081|issue:45907", + "issue:44466|issue:45276", + "issue:44016|issue:44291", + "issue:43441|issue:45834", + "issue:42994|issue:43317", + "issue:43653|issue:45593", + "issue:44623|issue:45397", + "issue:44623|issue:44749", + "issue:45600|issue:45712", + "issue:33290|issue:45656", + "issue:44261|issue:44263", + "issue:44361|issue:45593", + "issue:44671|issue:44779", + "issue:44016|issue:45823" + ], + "prepared_review_unit_hash": "649fca8708ad4cb4775fb907b85ebf5a932b02321ce0f42fedaea65f59834b95", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35383, + "estimated_input_tokens": 8846, + "estimated_eval_tokens": 17948 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose set of Transformers v5 regressions and model-specific bugs, not a set of true duplicates. I would use the broad remote_code regression report as the canonical issue, but no soft pair looks like the same concrete bug/change.", + "confidence": 0.89, + "canonical_issue_reason": "issue:45020 is the broadest and most umbrella-like report here, covering recent-version breakage for model loading/custom code and fitting the general regression theme of the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best single issue to anchor the cluster because it is the most general, well-scoped regression report and can subsume nearby v5 breakage discussions without overfitting to one model.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Different failures: video-classification pipeline processor lookup vs general remote_code loading regressions." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related, but one is HybridCache import and the other is many import errors after upgrade; not the same concrete bug." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Both affect tokenization, but they concern different models and different incorrect outputs, so they are not the same underlying tokenizer bug." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Processor/tokenizer API regression vs DeepSeek tokenization regression; different code paths and symptoms." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues, but one is a SigLIP2 example failure and the other is stale pipeline docs wording; unrelated." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Video pipeline processor bug and Gemma4 device-mapping mismatch are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different config classes and different initialization/attribute problems; not the same bug." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Torch flex_attention compile failure vs TensorFlow/tensorboard backend error; no shared code path." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python version support regression and PEFT version pin issue are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Both are model-loading issues, but they are model-specific regressions with different failure modes." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer repository loading vs Mistral regex patch crash are different tokenizer-loading bugs." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Save-pretrained metadata mismatch is not the same as ignoring tokenizer.json on load." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor shape bug and generic dependency issue do not match." + }, + { + "left": "issue:43650", + "right": "issue:46132", + "accept": false, + "reason": "Unrelated: placeholder data issue vs behavior change in AttentionInterface.register." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regex crash and list_repo_files revision forwarding are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Serialization inconsistency for tied weights vs resize_token_embeddings not updating Gemma4 outputs; different code paths." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error and init_empty_weights argument regression are unrelated." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "FlashAttention failure for Ministral-3 vs long-OCR index error in Kosmos2.5; different model-specific bugs." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Both involve quantization/offload, but one is saving and the other is loading with device_map; not the same bug." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer special-token registration bug and D-FINE denoising-loss behavior are unrelated." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Missing processor save files vs Gemma-4 Zero3 loading are different issues." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor save-pretrained problem and data filtering slowdown are unrelated." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Both are CI/repo hygiene issues, but one is stale auto_mappings entries and the other is leftover dummy classes; different fixes." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both mention DeepSpeed/optimizer behavior, but OOM with Adafactor and optimizer.step being called twice are different concrete bugs." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "Different model internals: q_a_layernorm epsilon config vs torch.split return values in a separate indexer." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer AttributeError on tasks and D-FINE loss behavior are unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Different model families and different failure modes: CamemBERT MLM predictions vs DeepSeek tokenization." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 37, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42491", + "issue:42886", + "issue:43262", + "issue:43299", + "issue:43441", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43643", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44162", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44373", + "issue:44496", + "issue:44521", + "issue:44589", + "issue:44683", + "issue:44734", + "issue:44871", + "issue:45003", + "issue:45103", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593", + "issue:46132", + "issue:46399", + "issue:46424" + ], + "soft_pairs": [ + "issue:43723|issue:44162", + "issue:43441|issue:44683", + "issue:42175|issue:43976", + "issue:43901|issue:43994", + "issue:43262|issue:44521", + "issue:43673|issue:43761", + "issue:44297|issue:45397", + "issue:43824|issue:43976", + "issue:43525|issue:43881", + "issue:43901|issue:44246", + "issue:43525|issue:43643", + "issue:44589|issue:45362", + "issue:41628|issue:43976", + "issue:43784|issue:45542", + "issue:44188|issue:44336", + "issue:45103|issue:46399", + "issue:45561|issue:45563", + "issue:44336|issue:44373", + "issue:44246|issue:44496", + "issue:44295|issue:45397", + "issue:44242|issue:45593", + "issue:42491|issue:43299", + "issue:43531|issue:45362", + "issue:43526|issue:45003", + "issue:41950|issue:44220", + "issue:42886|issue:46132", + "issue:44246|issue:44871", + "issue:44734|issue:46424" + ], + "prepared_review_unit_hash": "0813e514ba7a1ce29c1caab06c5158473d1c46bbfc6770d91ee3daa7b0827de0", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35380, + "estimated_input_tokens": 8845, + "estimated_eval_tokens": 17946 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "A heterogeneous set of Transformers regressions across tokenizer/config loading, multimodal processors, model families, attention/caching, and docs. They share a broad 'v5 broke something' theme, but most pairs are not the same underlying bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43723 is the most general, user-facing loading regression here: it targets a core AutoTokenizer.from_pretrained path and is broad enough to represent nearby import/load failures better than the model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43723 is the best overall representative issue because it describes a common entrypoint failure in v5 and is broad, well-scoped, and close to several other loading-related reports in the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Both are loading-related, but one is a generic AutoTokenizer regression and the other is an ESM2-specific failure; different concrete bug surfaces." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Both involve attention/flash/flex kernels, but one is a Ministral-3 FlashAttention failure and the other is a torch>=2.9 compiled flex_attention regression; not the same bug." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Package/backend dependency issue vs Python version compatibility issue; different underlying causes." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch for return_all_scores vs a SigLIP2 model output problem; unrelated." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Both mention apply_chat_template, but one is about audio sampling rate defaults and the other about multimodal assistant masks; different code paths." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Cache handling in generation vs CLIPVisionModel hidden_states output; unrelated regressions." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer serialization metadata bug vs a gemma-4 from_pretrained/zero3 loading issue; not the same failure." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Specific Qwen2.5-VL import error vs general Python support regression; different problems." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Llama4Config missing pad_token_id vs glm-4v-9b loading failure; different model-specific issues." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Documentation mismatch vs intermittent import slowdown; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Both involve missing config fields, but they are different models and different manifestations; too broad to treat as the same bug." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Float8 storage deserialization issue vs Qwen3.5 chat crash; unrelated." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "AutoImageProcessor import failure vs Python version compatibility; no shared concrete bug." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Different backend/import regressions: sentence-transformers import NameError vs TensorFlow backend detection error." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention kernel divergence on torch.compile vs ANSI codes in loading_report; unrelated." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "Auto-docstring annotation handling bug vs Qwen3.5 kernel registration crash; different subsystems." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "Test harness race condition vs stale generation warning; unrelated." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI output formatting bug vs docstring text error; not the same bug." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import performance issue vs unrecognized model/config problem; different symptoms and causes." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Position_ids buffer access bug vs gemma-4 loading issue; unrelated." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Both concern missing losses, but they are different models and different loss terms; too distinct to merge." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3 MoE models, but one is LoRA compatibility across branches and the other is a model loading regression; not the same concrete bug." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug vs Qwen3.5 chat crash; related family, but different defects." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs unsafe sys.modules access in modeling_utils; unrelated." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification processor mismatch vs audio fbank extraction issue; different modalities and code paths." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline tokenizer cache loading bug vs AttentionInterface.register behavior change; unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowdown vs Gemma-3 eos_token_id configuration mismatch; different bug classes." + }, + { + "left": "issue:44734", + "right": "issue:46424", + "accept": false, + "reason": "Serve KV-continuation indexing crash vs cache max length configuration only applying to quantized cache; different cache issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 38, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:36246", + "issue:42757", + "issue:43257", + "issue:43262", + "issue:43352", + "issue:43475", + "issue:43596", + "issue:43673", + "issue:43901", + "issue:44162", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44351", + "issue:44466", + "issue:44479", + "issue:44485", + "issue:44496", + "issue:44561", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44877", + "issue:44987", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45250", + "issue:45292", + "issue:45335", + "issue:45375", + "issue:45397", + "issue:45412", + "issue:45459", + "issue:45561", + "issue:45584", + "issue:45656", + "issue:45701", + "issue:45759", + "issue:45859", + "issue:46123", + "issue:46170", + "issue:46291", + "issue:46369", + "issue:46399", + "issue:46424" + ], + "soft_pairs": [ + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:44336|issue:44485", + "issue:45003|issue:45561", + "issue:44336|issue:46424", + "issue:44779|issue:45292", + "issue:43475|issue:46369", + "issue:44242|issue:44857", + "issue:44295|issue:44749", + "issue:45859|issue:46399", + "issue:46291|issue:46399", + "issue:36246|issue:43596", + "issue:44162|issue:45200", + "issue:43352|issue:45250", + "issue:43673|issue:44561", + "issue:43901|issue:45003", + "issue:44496|issue:45003", + "issue:43257|issue:45137", + "issue:43262|issue:44246", + "issue:45292|issue:45584", + "issue:45412|issue:46123", + "issue:45375|issue:45759", + "issue:44561|issue:46399", + "issue:45292|issue:45701", + "issue:44877|issue:45375", + "issue:44987|issue:45459", + "issue:44995|issue:45200", + "issue:44466|issue:45335", + "issue:44479|issue:46170", + "issue:45397|issue:45656" + ], + "prepared_review_unit_hash": "dd0288808b2252f111977153a953e315918e453b0dbb14056bf41b331cb960b9", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36687, + "estimated_input_tokens": 9172, + "estimated_eval_tokens": 18600 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a similarity cluster, not a true duplicate set: the issues span unrelated import errors, model-specific bugs, config validation, performance, and docs. I did not find any soft pair that clearly describes the same underlying bug/change.", + "confidence": 0.34, + "canonical_issue_reason": "No strong canonical duplicate exists. If one representative is needed, issue 45003 is the broadest core-library bug and has the most central-looking weak similarity links, but the cluster is heterogeneous overall.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45003 is the best single representative only because it is a concrete internal utility bug with broader applicability than the model-specific reports; however, it is not a true duplicate hub.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols from different packages and unrelated code paths." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both mention import failures in transformers, but the missing classes are different and the reports point to different modules." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI-code output handling and GLM-5 RoPE implementation are unrelated bugs." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "One is an unsafe sys.modules access in modeling_utils; the other is a pytest-xdist file race in testing utilities." + }, + { + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "Loading-report terminal coloring and cache-length behavior are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer correctness and resize_token_embeddings output synchronization are different model behaviors." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "Both are SAM3-related, but one is a missing attribute in vision encoder output and the other is a memory overflow problem." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Load-balancing loss computation and RT-DETR AMP crashes are unrelated." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "A position_ids buffer access error is not the same bug as a slowdown after upgrading transformers." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "Both are Qwen3.5/MoE-related, but one is CPU-offload incompatibility in rotary embeddings and the other is a kernel registration error." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Kernels version/revision installation failure and apply_rotary_pos_emb registration failure are distinct issues." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Qwen2.5-VL import failure and DeepSpeed ZeRO3 BertModel IndexError are unrelated." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 token-type defaults are different model-specific bugs." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "A model not supporting Flash Attention 2.0 yet is not the same as the generic Flash Attention 2.0 issue." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked-prefill cache handling and removal of is_torch_fx_available affect different code paths." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs for text classification pipeline and unsafe sys.modules access are unrelated." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model config and sys.modules access are different failure modes." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both involve DeepSpeed, but one is missing MOE weight conversion and the other is an empty-deque IndexError." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat-template sampling rate defaults and import latency are unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Embedding resize propagation and Whisper empty transcription handling are different bugs." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "RT-DETR memory cleanup and mask pipeline batch-finalization are unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Both touch Qwen multimodal config loading, but one is strict-field dropping and the other is text_config unwrapping; they are not the same concrete bug." + }, + { + "left": "issue:44561", + "right": "issue:46399", + "accept": false, + "reason": "A removed helper breaking trust_remote_code is unrelated to a kernel registration error." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "Embedding resize synchronization and tokenizer version drift are different problems." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both mention strict config handling, but they affect different model configs and different symptoms." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Loading physical-intelligence/fast and hidden tokenizer errors without protobuf are separate import/loading issues." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "A stale indexer cache on second forward and Gemma 4 token-type defaults are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Serialization of tied lm_head weights and t5gemma embedding resize behavior are different bugs." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Qwen VL video regression and Windows/ROCm torch.distributed import issues are unrelated." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma-4 ZeRO3 loading and double optimizer step with deepspeed are different DeepSpeed-related bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 39, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:42886", + "issue:42994", + "issue:43065", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43482", + "issue:43540", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43673", + "issue:43701", + "issue:43931", + "issue:44038", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44483", + "issue:44623", + "issue:44749", + "issue:44869", + "issue:44871", + "issue:44877", + "issue:44977", + "issue:44995", + "issue:45137", + "issue:45229", + "issue:45230", + "issue:45356", + "issue:45397", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45656", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202", + "issue:46421" + ], + "soft_pairs": [ + "issue:45987|issue:46082", + "issue:45137|issue:45600", + "issue:43475|issue:43540", + "issue:44483|issue:46202", + "issue:43653|issue:44162", + "issue:42886|issue:45229", + "issue:44060|issue:44077", + "issue:44995|issue:45588", + "issue:44368|issue:44977", + "issue:43065|issue:43408", + "issue:44368|issue:45563", + "issue:44188|issue:46202", + "issue:44877|issue:46170", + "issue:43931|issue:44368", + "issue:45584|issue:46170", + "issue:44188|issue:46421", + "issue:43065|issue:46123", + "issue:44297|issue:44623", + "issue:44361|issue:44749", + "issue:44869|issue:45356", + "issue:44623|issue:45656", + "issue:43646|issue:44162", + "issue:44749|issue:45656", + "issue:43701|issue:44749", + "issue:43232|issue:43673", + "issue:44871|issue:46132", + "issue:45230|issue:45397", + "issue:42994|issue:43482", + "issue:44038|issue:44623", + "issue:43475|issue:43643" + ], + "prepared_review_unit_hash": "66d33a4200971214ecd412efa7bfe321a85cf110a6625ff8c8d6f4dd349d7294", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35931, + "estimated_input_tokens": 8983, + "estimated_eval_tokens": 18222 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a grab bag of unrelated closed issues; the soft pairs only show superficial embedding similarity, not the same underlying bug. No PRs are present.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43646 is the broadest framework-level regression in the set ('Transformers 5.0.0 breaks custom model initialization') and is the closest thing to an umbrella issue; the others are mostly model-, tokenizer-, or subsystem-specific.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best representative issue for this cluster because it captures a general v5 breakage theme, while the rest are too fragmented to serve as a better canonical duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache return-type bug vs LlamaConfig head_dim validation; different code paths." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error vs stale auto_mappings references; unrelated failures." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "SAM3 encoder missing attribute vs Qwen3Omni video input ValueError; not the same bug." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "Chat/completions request handling vs Gemma pipeline roundtrip; different behaviors." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer special-token registration issue vs ESM2 regression; different subsystems." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline cache tokenizer load failure vs Gemma4 multi-GPU CUDA OOM; unrelated." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Tied-weights warning bug vs patchtsmixer post_init policy change; not the same problem." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Stale cache on second forward vs flash_attention None-aux crash; different code paths." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "Tied-embeddings warning vs flash-attention generation issue; not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Dummy Conv2d in SAM3 decoder vs model-type warning for sam3_video/sam3_tracker; distinct issues." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "Different warnings in different generation/save paths; no shared failure mode." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "torch.compile kernel divergence vs Gemma pipeline token roundtrip; unrelated." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Strict config loading failure vs Windows/ROCm import dependency error; different root causes." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Weight-shape mismatch in Qwen3-VL-30B vs tied-embeddings warning; unrelated." + }, + { + "left": "issue:45584", + "right": "issue:46170", + "accept": false, + "reason": "Whisper empty-transcription generation bug vs Windows ROCm import/type error; different areas." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Attention-kernel divergence under compile vs cache lazy-init recompilation; different defects." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "SAM3 Conv2d placeholder vs MaskGenerationPipeline partial-batch dropping results; not the same bug." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "tokenizer.save_pretrained metadata mismatch vs processor.save_pretrained missing files; same family only, not same bug." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer AttributeError vs global slowdown regression after upgrade; unrelated." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decode crash vs Kimi-K2.5 tokenizer codec regression; different concrete failures." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Missing save_pretrained files vs optimizer step called twice; unrelated." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom model initialization breakage vs ESM2-specific regression; too broad to be the same bug." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Preprocessing slowdown vs double optimizer step under deepspeed; different behavior." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch vs preprocessing slowdown; no shared underlying change." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Generation kwargs update after sync_gpus vs missing cache during chunked prefill; related area but distinct bug." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma EOS config inconsistency vs AttentionInterface.register behavior change; not the same issue." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic bug report vs gemma-4 zero3 loading problem; insufficient overlap and different specifics." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure vs Qwen2.5-GGUF loading failure; different model-loading paths." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe v5 bug vs processor.save_pretrained missing files; unrelated." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "SAM3 video encoder attribute error vs AutoConfig trust_remote_code field loss; different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 40, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42371", + "issue:42673", + "issue:42831", + "issue:42994", + "issue:43232", + "issue:43377", + "issue:43475", + "issue:43504", + "issue:43525", + "issue:43526", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44351", + "issue:44496", + "issue:44792", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45310", + "issue:45375", + "issue:45381", + "issue:45406", + "issue:45478", + "issue:45561", + "issue:45712", + "issue:45753", + "issue:46077", + "issue:46123", + "issue:46202", + "issue:46326", + "issue:46399" + ], + "soft_pairs": [ + "issue:28282|issue:30990", + "issue:43475|issue:45381", + "issue:45406|issue:46202", + "issue:41628|issue:44933", + "issue:42371|issue:45092", + "issue:45375|issue:45478", + "issue:43723|issue:46326", + "issue:42994|issue:43525", + "issue:43784|issue:46399", + "issue:45310|issue:45375", + "issue:45081|issue:45103", + "issue:43873|issue:44945", + "issue:46077|issue:46123", + "issue:42673|issue:43704", + "issue:41628|issue:44351", + "issue:43232|issue:44188", + "issue:44964|issue:45137", + "issue:44222|issue:45310", + "issue:44945|issue:45753", + "issue:42994|issue:43824", + "issue:44038|issue:45310", + "issue:36010|issue:41628", + "issue:43232|issue:44792", + "issue:42831|issue:44222", + "issue:42175|issue:45712", + "issue:44496|issue:45561", + "issue:43526|issue:45561", + "issue:43504|issue:45561", + "issue:42175|issue:44933", + "issue:43377|issue:44242" + ], + "prepared_review_unit_hash": "f6ec3ca0d9a8af03e4cbda2ed4e9b4e7a3e540074a99a999759be964bb136e40", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36057, + "estimated_input_tokens": 9015, + "estimated_eval_tokens": 18286 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a heterogeneous mix of unrelated issues. The only clear duplicate pair in the supplied soft edges is the Qwen3ForCausalLM VRAM leak report (42673/43704), with 42673 the best canonical issue. A few other pairs are superficially similar (imports, Qwen/MoE, FP8, masking) but describe different code paths or symptoms.", + "confidence": 0.9, + "canonical_issue_reason": "issue:42673 is the oldest and most-discussed instance of the only clearly duplicated bug in the set, and it has the stronger duplicate signal versus issue:43704.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 is the best representative of the duplicate cluster because it is the earliest, most active report of the same VRAM leak bug and has an exact-title duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "Different failures: missing PyTorch backend import vs Sentence Transformers hanging while loading." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal bugs, but they affect different models and different failure modes." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Gemma tokenizer/processor problems, but one is a missing attribute and the other is a roundtrip tokenization issue." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import/export problems, but they target different symbols and do not share the same code path." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "Packaging/backend install issue vs remote-code meta-init incompatibility; unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Related Qwen3.5 MoE area, but one is a dropped config field and the other is a from_pretrained load error." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer loading regression vs continue_final_message behavior for mistral_common; different bugs." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized model saving failure vs missing config attribute; different problems." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Sentence-transformers import breakage vs Qwen3.5 kernel registration crash; unrelated." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Same model family, but different bug types: load failure vs missing vision-config field." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Tokenizer regex crash vs auto-docstring crash from future annotations; unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization/offloading behavior vs pipeline-parallel output corruption; different code paths." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Both involve masking, but one is Gemma4 attention masking and the other is pipeline batching state." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same VRAM leak bug in Qwen3ForCausalLM multi-dataloader threads; 42673 is the stronger canonical report." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and different regressions." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs/sync_gpus bug vs torch.compile attention-kernel divergence; not the same issue." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Phi-4 multimodal load failure vs DeepSpeed ZeRO3 deque error; unrelated." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained bug vs Qwen3.5 from_pretrained load error; different failures." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline-parallel correctness issue vs image-specific infinite loop; distinct symptoms and code paths." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized saving vs missing model import; unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Both touch Qwen MoE and transformers v5, but they are different model variants and different failure modes." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they are for different classes and likely different API regressions." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation state bug vs janus image-generation test failure; unrelated." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both mention FP8, but one is an accuracy issue and the other is model saving." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Backend-install packaging issue vs leftover dummy class leakage; different code paths." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Model config loading failure vs xdist testing race; unrelated." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "Beit image-processor label bug vs testing race; unrelated." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Beit pretrained-loading legacy-field bug vs testing race; unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Tensorflow-backend install issue vs image_utils import mismatch; unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "Different model behavior bugs: MIMI batching discrepancy vs MoE load-balancing loss omission." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet. The canonical choice of issue:42673 over issue:43704 is supported by the earlier creation date, higher discussion activity, and identical title. The soft-edge verdicts stay conservative overall and do not overstate duplicates where the packet only shows superficial similarity." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 41, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28282", + "issue:30333", + "issue:30990", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:42222", + "issue:42831", + "issue:42898", + "issue:43278", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:43976", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44079", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44448", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44869", + "issue:44933", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45137", + "issue:45200", + "issue:45245", + "issue:45310", + "issue:45356", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:45812", + "issue:46132", + "issue:46231" + ], + "soft_pairs": [ + "issue:42898|issue:44568", + "issue:43976|issue:44933", + "issue:43278|issue:45137", + "issue:30990|issue:35141", + "issue:44623|issue:45310", + "issue:42222|issue:45200", + "issue:30333|issue:33290", + "issue:43479|issue:44869", + "issue:43479|issue:44079", + "issue:42898|issue:45812", + "issue:43784|issue:43824", + "issue:43723|issue:44568", + "issue:42898|issue:44779", + "issue:42898|issue:45356", + "issue:44016|issue:46231", + "issue:44734|issue:45103", + "issue:43874|issue:45092", + "issue:44295|issue:45656", + "issue:44075|issue:46231", + "issue:42898|issue:44448", + "issue:44060|issue:45125", + "issue:45405|issue:45712", + "issue:44222|issue:44297", + "issue:44933|issue:45712", + "issue:44297|issue:45656", + "issue:44448|issue:46132", + "issue:44222|issue:44623", + "issue:27957|issue:30333", + "issue:30333|issue:34689", + "issue:43479|issue:43644", + "issue:42831|issue:44623", + "issue:30333|issue:36331", + "issue:30333|issue:35141", + "issue:22355|issue:28282", + "issue:45245|issue:46132" + ], + "prepared_review_unit_hash": "d25ada5a1b30b975823c85751b13ded869c2afcf22d78031d8a602dc03134946", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38199, + "estimated_input_tokens": 9550, + "estimated_eval_tokens": 19356 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are mostly unrelated issues that only share broad themes like tokenization, save/load, or deepspeed. I would reject all proposed duplicates; there is no clear same-bug pair here.", + "confidence": 0.92, + "canonical_issue_reason": "If a single issue must represent the set, 42898 is the broadest and most generic tokenizer-regression report among the candidates, but the cluster is not actually a duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "42898 is the strongest general representative of the tokenizer-regression subgroup: concrete, user-visible, and broader than the more model-specific reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer regressions, but they describe different behaviors and different affected models." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python-version incompatibility and a missing import are unrelated failures." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift and a DeepSpeed deque IndexError are separate bugs." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Loading hangs and embedding reinitialization are different code paths and symptoms." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Processor save-pretrained missing files is unrelated to Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Broken vitpose models and Gemma 4 multimodal token defaults are different issues." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow reporting after failure and Adafactor OOM are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Config default initialization and Whisper timestamp decoding are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Multimodal config defaults and ModelOutput key assignment are different subsystems." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are v5 tokenizer regressions, but the affected models and concrete bugs differ." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Importing sentence-transformers and missing Qwen2.5-VL export are unrelated import errors." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer loading failure and missing BOS/EOS addition are not the same bug." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Different tokenizer regressions affecting different models and tokenization logic." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Different tokenizer regression symptoms; one is generic behavior change, the other codec/regex handling." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error and documentation parameter typo are unrelated." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache continuation crash and auto-docstring annotation handling are unrelated." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image-processor method and remote-code meta-init incompatibility are different multimodal bugs." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Position-id buffer access and duplicate optimizer steps are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Optimizer argument handling and a docstring parameter-name mismatch are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Both mention v4-to-v5 output changes, but they are different tokenizer/output regressions." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Incorrect tied-weights warning and missing tensor-parallel plan are distinct Qwen issues." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "Dependency version bump and dummy-class cleanup are unrelated maintenance issues." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 save-pretrained and tokenizer save-pretrained bugs affect different components." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing image_utils import and leftover dummy classes are not the same defect." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer metadata save bug and DeepSpeed optimizer stepping are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Pegasus v4/v5 output regression and AttentionInterface.register behavior change are different bugs." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Model save-pretrained FP8 bug and processor save-pretrained missing files are not the same issue." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention support and MLflow reporting after failures are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "Training-job reporting and Llama 3.2 vision model loading are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Multimodal config defaults and junk-filled buffers are different v5 initialization bugs." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy issues and processor file-saving bugs are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx and missing PyTorch are different optional-dependency import errors." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category-limit runtime error and registered-function behavior change are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 42, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:30990", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42907", + "issue:43122", + "issue:43232", + "issue:43475", + "issue:43644", + "issue:43723", + "issue:43761", + "issue:44190", + "issue:44206", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44734", + "issue:44829", + "issue:44871", + "issue:44977", + "issue:44991", + "issue:45071", + "issue:45092", + "issue:45103", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45834", + "issue:46002", + "issue:46050", + "issue:46082", + "issue:46209", + "issue:46305", + "issue:46328", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:45071|issue:46082", + "issue:43122|issue:46209", + "issue:44297|issue:45310", + "issue:41628|issue:42757", + "issue:43475|issue:46305", + "issue:44190|issue:46082", + "issue:43723|issue:44977", + "issue:44206|issue:45103", + "issue:45381|issue:46305", + "issue:43232|issue:46424", + "issue:44734|issue:46421", + "issue:44222|issue:44484", + "issue:43232|issue:45834", + "issue:44295|issue:44623", + "issue:44263|issue:45092", + "issue:44373|issue:44485", + "issue:43723|issue:44871", + "issue:44483|issue:45406", + "issue:44279|issue:46328", + "issue:43723|issue:44991", + "issue:42907|issue:46050", + "issue:30990|issue:38175", + "issue:44263|issue:44360", + "issue:45092|issue:45357", + "issue:43761|issue:46082", + "issue:43644|issue:46305", + "issue:44829|issue:46002", + "issue:44261|issue:45092", + "issue:42831|issue:45310" + ], + "prepared_review_unit_hash": "0fd8afdc594ca25355caf7b55b9a84e37acbd63ee89195963a967a8e057242a8", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35560, + "estimated_input_tokens": 8890, + "estimated_eval_tokens": 18036 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Loose cluster of mostly unrelated Transformers issues: v5 regressions, tokenizer/config loading failures, cache/serving bugs, and a few model-specific defects. No soft pair appears to be the same underlying bug or change.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45071 is the broadest and most representative item in the cluster: a clear Transformers v5 regression affecting core config behavior. It\u2019s concise, generic, and closest to the cluster\u2019s shared theme of v5 breakage.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45071 is the best single issue representative because it describes a core framework regression rather than a model-specific symptom or a narrow edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "Both are config-related v5 problems, but one is broad type-checking behavior and the other is LlamaConfig validation for head_dim; different code paths and fixes." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Different subsystems and failure modes: tokenizer output changes vs FP8 inference failure on B200." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "One is tokenizer save_pretrained metadata mismatch; the other is a Qwen3.5 from_pretrained model loading error. Related area, not the same bug." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols from different packages/dependencies; not the same underlying issue." + }, + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "Different models and different bugs: missing vision encoder attribute vs a sliding-window mask off-by-one." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Unrelated: local dataset loading in an example script vs LlamaConfig head_dim validation." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer loading failure and flash-attention generation issue are unrelated." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "Different failure modes and code paths: feature extractor argument mismatch vs auto_docstring crash with future annotations." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Both mention positional behavior, but they affect different models and different mechanisms; not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache/generation related, but one concerns updating model kwargs after sync_gpus and the other cache_length semantics for quantized cache." + }, + { + "left": "issue:44734", + "right": "issue:46421", + "accept": false, + "reason": "Serving KV-cache indexing crash vs cache lazy init causing recompilation are separate issues." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "Different save_pretrained topics: FP8 MoE serialization bug vs shard-size default question." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs handling is unrelated to Kosmos2.5 long OCR index errors." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Buffer registration/read issue vs processor save_pretrained missing files; different components and symptoms." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "Different model code paths: DSA indexer split behavior vs remote-code checkpoint meta-init incompatibility." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "A docstring mismatch is not the same as a RoPE implementation bug." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer load regression and eos_token_id config mismatch are distinct problems." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "One is an API request acceptance issue, the other a processor attribute error in serve; not the same bug." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "Generic dependency issue vs a performance bug in SequenceFeatureExtractor.pad." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading related, but they concern different models and likely different causes." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Both involve save/load and quantization, but one is dequantized model saving and the other is sinq round-trip breakage; not the same concrete defect." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Completely different models and symptoms: loading hangs vs zero probabilities." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Same broad area (DSA indexer) but different concrete bugs: split return behavior vs missing ReLU." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Both are model-specific v5 compatibility problems, but they affect different checkpoints and different serialization/init paths." + }, + { + "left": "issue:43761", + "right": "issue:46082", + "accept": false, + "reason": "Vision model hidden_states output bug vs LlamaConfig head_dim validation are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer initialization corruption is unrelated to sliding-window masking." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Different training pathologies with different likely causes; not the same bug." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "RoPE/layernorm precision issue vs remote-code checkpoint meta-init incompatibility are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FP8 accuracy degradation is unrelated to Qwen3.5 from_pretrained loading failure." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 43, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:30990", + "issue:33357", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43479", + "issue:43550", + "issue:43576", + "issue:43646", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:44038", + "issue:44060", + "issue:44117", + "issue:44188", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44368", + "issue:44373", + "issue:44521", + "issue:44568", + "issue:44704", + "issue:44913", + "issue:44945", + "issue:45278", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46132", + "issue:46305", + "issue:46326", + "issue:46369", + "issue:46402" + ], + "soft_pairs": [ + "issue:43479|issue:44913", + "issue:41628|issue:43824", + "issue:43408|issue:46402", + "issue:42831|issue:44038", + "issue:44263|issue:44373", + "issue:39401|issue:44568", + "issue:22355|issue:27957", + "issue:42757|issue:45278", + "issue:27957|issue:30990", + "issue:42886|issue:43066", + "issue:42886|issue:42994", + "issue:44368|issue:46305", + "issue:42831|issue:45753", + "issue:22355|issue:28218", + "issue:43408|issue:46369", + "issue:43232|issue:45593", + "issue:43576|issue:43784", + "issue:43646|issue:46002", + "issue:43756|issue:44060", + "issue:43550|issue:46132", + "issue:44295|issue:44297", + "issue:44117|issue:44704", + "issue:30990|issue:33357", + "issue:42371|issue:44263", + "issue:43122|issue:44521", + "issue:42907|issue:44945", + "issue:44188|issue:44373", + "issue:43122|issue:46326", + "issue:41628|issue:45278", + "issue:42907|issue:43937" + ], + "prepared_review_unit_hash": "6804ae1a4f0db8c3be9bb9383fb06d40d48c0cc9a1b98ec2405f4cb9ae4b1398", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36184, + "estimated_input_tokens": 9046, + "estimated_eval_tokens": 18348 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose semantic grouping of mostly unrelated Transformers issues (imports, tokenizers, model-specific regressions, and training/runtime bugs). None of the proposed soft pairs look like the same underlying bug/change, so all are rejected. The broad import-regression report is the best representative issue, but only as an umbrella, not a duplicate target for the others.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45278 is the broadest and most representative umbrella for the cluster\u2019s recurring import-regression theme after upgrade, even though it does not duplicate the other issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the most suitable cluster representative because it captures a general user-facing regression pattern and can contextualize several related import complaints.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Both concern config handling, but they hit different models and different fields; not the same bug." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and different code paths." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "Same sam3 family, but one is a wrong model-type warning and the other is missing text-encoder weights." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "Different model-specific regressions: FP8 accuracy vs Qwen3-VL-Moe / Transformers 5 compatibility." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "Unrelated problems: tensor split return values versus a docstring issue." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer bugs, but they affect different tokenizers and different behaviors (offset mapping vs BOS/EOS insertion)." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "ONNX import failure and Flash Attention model support are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "There is broad import-error overlap, but one is a specific missing hub export and the other is a general post-upgrade import regression; not the same concrete bug." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "Different subsystems and symptoms: attention backend support vs loading hang." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure is unrelated to wrong tokenizer decoder type in v5." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Tokenizer cache/offline behavior and quantized model saving are different bugs." + }, + { + "left": "issue:44368", + "right": "issue:46305", + "accept": false, + "reason": "A warning about tie_word_embeddings is unrelated to a sliding-window mask off-by-one error." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "Accuracy issue in FP8 is not the same as an infinite-loop multimodal inference bug." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Missing transformers.onnx module and extra-space tokenization behavior are unrelated." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "Both touch sam3, but one is a model-type mismatch warning and the other is a memory overflow." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss logic are different code paths." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "A broken env CLI command is not the same as a NameError during sentence-transformers import." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization breakage and NaNs in classification heads are distinct failures." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Different model-specific issues: RoPE layer retention vs tied-weights warning." + }, + { + "left": "issue:43550", + "right": "issue:46132", + "accept": false, + "reason": "torch.compile SDPA failure is unrelated to AttentionInterface.register changing function behavior." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a registered buffer and tokenizer save metadata mismatch are different bugs." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "One is None handling in tokenizer mapping; the other is missing kwargs propagation in AutoProcessor." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading stuck and a MacOS bus error are different symptoms with different likely causes." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 configuration guidance is unrelated to torch.split return values." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Different tokenizer/template bugs: tokenization drift vs assistant mask generation for multimodal inputs." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Quantized model saving failure and pipeline-parallel wrong output are unrelated runtime issues." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Compiler-dependent attention-kernel divergence is not the same as a docstring problem." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Both involve tokenization, but one is tokenization drift across versions and the other is continue_final_message for mistral_common tokenizers." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Both mention import errors, but they are not the same concrete missing import or code path." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Saving dequantized models and invalid GenerationConfig are different failure modes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 44, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:29159", + "issue:36296", + "issue:39401", + "issue:42175", + "issue:42673", + "issue:42757", + "issue:42907", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43493", + "issue:43784", + "issue:43824", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44222", + "issue:44261", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44938", + "issue:44945", + "issue:44964", + "issue:45072", + "issue:45200", + "issue:45216", + "issue:45290", + "issue:45478", + "issue:45600", + "issue:45753" + ], + "soft_pairs": [ + "issue:39401|issue:44488", + "issue:44261|issue:44373", + "issue:43976|issue:44945", + "issue:44704|issue:45290", + "issue:43784|issue:45600", + "issue:42907|issue:44964", + "issue:39401|issue:44521", + "issue:43116|issue:45200", + "issue:44038|issue:44297", + "issue:43122|issue:44514", + "issue:44625|issue:44829", + "issue:44360|issue:44485", + "issue:44038|issue:44295", + "issue:43408|issue:43475", + "issue:42907|issue:43452", + "issue:43122|issue:44779", + "issue:44038|issue:44222", + "issue:45216|issue:45478", + "issue:43824|issue:44351", + "issue:36296|issue:43116", + "issue:42757|issue:44351", + "issue:22355|issue:29159", + "issue:43278|issue:44945", + "issue:42175|issue:44938", + "issue:43493|issue:45753", + "issue:43066|issue:44779", + "issue:42673|issue:45072", + "issue:22355|issue:29127", + "issue:43992|issue:44704" + ], + "prepared_review_unit_hash": "0cd1e82f7b2798d8db01e6155f1f0ceafb3286877dcd29c680dfb6fcaddf4a2d", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35476, + "estimated_input_tokens": 8869, + "estimated_eval_tokens": 17994 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly a collection of unrelated issues across tokenizer regressions, model-loading bugs, training/inference problems, and packaging/import errors. I found no strong duplicate pair with enough evidence to merge; the best anchor issue is the broad tokenization-regression report in 43122.", + "confidence": 0.84, + "canonical_issue_reason": "Issue 43122 is the clearest broad regression report: it defines a version boundary, a reproducible tokenizer behavior change, and is more general than the model-specific or symptom-specific reports around it.", + "canonical_pr_reason": null, + "best_issue_reason": "43122 is the strongest standalone issue to use as an anchor because it is concrete, versioned, and likely to attract follow-up analysis without being overly niche.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Both are about Qwen-related text behavior, but one is an offset-mapping bug and the other is a model-loading failure; different code paths." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "One is a numerical/config bug in MLA layernorm, the other is a docstring issue for position_ids; not the same defect." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python compatibility/install failure vs incorrect output under pipeline parallelism; unrelated symptoms and fixes." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor kwargs forwarding and chat-template tool-call handling are distinct APIs with different failure modes." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Import-time NameError in sentence-transformers integration vs CI break from stale auto-mapping references; not the same bug." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Save/dequantization failure for Ministral/Devstral is unrelated to loading Phi-4 multimodal." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Both touch Qwen3.5 tokenization/chat templating, but offset_mapping and assistant_masks are different behaviors and code paths." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example script multi-label output bug vs Gemma 4 multimodal token-type defaulting; unrelated functionality." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe loading bug and tokenizer_config save-pretrained mismatch are separate issues." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Versioned tokenizer regression vs batched chat-template crash with padding=False; different failure surfaces." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Config propagation for Qwen3.5 labels is unrelated to flash_attention_3 training degeneration." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "A missing ReLU in a DSA indexer is unrelated to a GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Model-load compatibility issue vs a position_ids buffer-read error; no shared concrete bug path." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Both involve SAM 3 Video, but one is a config-type warning and the other is a missing attribute in encoder output; too different to merge." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Dequantized save failure and gguf_file load failure are different serialization paths." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both mention tokenizer regressions in v5, but Deepseek-specific incorrect outputs are not the same concrete bug as the generic slow/fast-tokenizer inconsistency." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe loading regression and FP8 save_pretrained bug are unrelated model lifecycle failures." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both concern Qwen3.5 and 5.4.0, but save_pretrained corruption and from_pretrained errors are different code paths and not clearly the same defect." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols from different modules and likely different causes." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor-parallel training bug is unrelated to example-script multi-label classification behavior." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Missing huggingface_hub symbol and missing transformers symbol are separate dependency/API breakages." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "Module import failure for transformers.onnx is unrelated to tokenizer slow/fast inconsistency." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Dtype drift between training and evaluate is unrelated to incorrect outputs under pipeline parallelism." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Packaging extras missing TensorFlow and Python 3.14 load failure are different environment issues." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 implementation mismatch and Qwen3_5 infinite-loop-on-image are different model-specific problems." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Wrong tokenizer decoder type and Deepseek tokenization regression may both touch tokenizers, but they are not the same concrete failure." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in multi-threaded Qwen3ForCausalLM use is unrelated to dtype mismatches in bfloat16 inference." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Import error for transformers.onnx and a LayoutLMv3 error-message clarity issue are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Embedding dtype change and pipeline-parallel wrong output are distinct training/inference bugs." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens.weight on load is unrelated to AutoProcessor kwargs forwarding." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 45, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:33453", + "issue:38617", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:43064", + "issue:43116", + "issue:43122", + "issue:43441", + "issue:43450", + "issue:43493", + "issue:43611", + "issue:43673", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:43986", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44188", + "issue:44291", + "issue:44355", + "issue:44380", + "issue:44466", + "issue:44485", + "issue:44779", + "issue:44936", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45083", + "issue:45278", + "issue:45357", + "issue:45405", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:46002", + "issue:46077", + "issue:46231", + "issue:46298", + "issue:46328" + ], + "soft_pairs": [ + "issue:38617|issue:45278", + "issue:43937|issue:44485", + "issue:43611|issue:46002", + "issue:44188|issue:46328", + "issue:44355|issue:46231", + "issue:43116|issue:43493", + "issue:42222|issue:43441", + "issue:42371|issue:43874", + "issue:44075|issue:44936", + "issue:45357|issue:45478", + "issue:44038|issue:45478", + "issue:43976|issue:45446", + "issue:43116|issue:43825", + "issue:45083|issue:45446", + "issue:43064|issue:43673", + "issue:33453|issue:39401", + "issue:43937|issue:45563", + "issue:39401|issue:44779", + "issue:44117|issue:45588", + "issue:44380|issue:46077", + "issue:42673|issue:43906", + "issue:43827|issue:45446", + "issue:43450|issue:43986", + "issue:43673|issue:46328", + "issue:44291|issue:45071", + "issue:44466|issue:46002", + "issue:43937|issue:44977", + "issue:45405|issue:45446", + "issue:44117|issue:45081", + "issue:43122|issue:46298" + ], + "prepared_review_unit_hash": "8e7ad74bfe1e0056e9cdc5fa01c3c248cf6b3886ec704be0788872e4d99cdba2", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36562, + "estimated_input_tokens": 9141, + "estimated_eval_tokens": 18538 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is a grab-bag of mostly unrelated Transformers bugs/regressions. The broadest representative is the v5 import-regression report, while the soft edge candidates are mostly similar only at the subsystem/model-family level and should remain separate.", + "confidence": 0.78, + "canonical_issue_reason": "issue:45278 is the most general and representative bug report in this set: a broad post-upgrade import-regression affecting multiple symbols, rather than a narrow model- or feature-specific failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 best captures the cluster\u2019s dominant theme of v5 upgrade breakage and is broad enough to stand in for the general regression class.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Both are import errors after a version upgrade, but 38617 is a specific missing symbol while 45278 is a broader multi-import regression; they do not look like the same concrete bug." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM-5, but one is a GenerationConfig validation failure and the other is a RoPE implementation discussion; different failures and code paths." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Loading with base_model_prefix and NaNs in classification heads are unrelated symptoms with different root causes." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Attention-kernel divergence under torch.compile and a padding performance issue are unrelated." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled Python-file errors and a documentation parameter-name mismatch are not the same underlying bug." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Example-script multi-label output failure and SigLIP2 implementation discrepancy are different model/problem domains." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "VitPose loading breakage and Ministral FlashAttention failure are separate model-specific regressions." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance and a missing image-processor method are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Optimizer-argument handling and trainer.evaluate() failing after train() are distinct training bugs." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both concern Qwen3.5/MoE, but one is save_pretrained key serialization and the other is from_pretrained loading failure; not the same concrete defect." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Related model family, but these are different Qwen variants and different failure modes." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python-version compatibility and an incorrect PyTorch version check are unrelated." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "An example-script classification result bug is unrelated to a pipeline error-message/docs issue." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Feature-extraction length behavior and a flex_attention import/version check are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Optimizer-state corruption under FSDP2/PEFT and missing GenerationMixin cache in chunked prefill are different code paths." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both are tokenizer-related regressions, but one is loading and the other is offset_mapping correctness; not the same bug." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig invalidity and a stale warning about num_return_sequences are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Different tokenizer regressions in different model families; no sign of the same underlying defect." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping fallback handling and a flash_attention crash on s_aux=None are unrelated." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Both involve attention behavior, but one is backend scaling under SDPA/FlashAttention and the other is Gemma4 bidirectional masking; different concrete bugs." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "VRAM leakage in multi-threaded dataloader use and an isolated reproduction of another issue are not enough to treat as the same bug." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Outdated docs referencing pipeline() and a PyTorch version check bug are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video processor shape handling and a missing torchvision crash are distinct video-processing issues." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Generation cache missing in chunked prefill and a padding performance inefficiency are unrelated." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "init_empty_weights constructor-arg breakage and PretrainedConfig type-checking regression are both v5 regressions but not the same code-path problem." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "lm_head.weight serialization inconsistency and NaNs in classification heads are unrelated." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GenerationConfig validation and a flash-attention generation failure in a different model are separate issues." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "A dependency version pinning problem and an import-version check bug are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping None-handling and a Mistral regex patch crash are both tokenizer-loading adjacent, but they are different failures and should not be merged." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenizer output changes and a noisy logging message are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 46, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42490", + "issue:42617", + "issue:42673", + "issue:42831", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43937", + "issue:44075", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44877", + "issue:44933", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45561", + "issue:45656", + "issue:46178", + "issue:46231", + "issue:46328" + ], + "soft_pairs": [ + "issue:44410|issue:45310", + "issue:45341|issue:45561", + "issue:38617|issue:41628", + "issue:45341|issue:46328", + "issue:42222|issue:46178", + "issue:44829|issue:45216", + "issue:42617|issue:45341", + "issue:45362|issue:46328", + "issue:42673|issue:43582", + "issue:43408|issue:43742", + "issue:39401|issue:43421", + "issue:42490|issue:42673", + "issue:44877|issue:46231", + "issue:42371|issue:44261", + "issue:44484|issue:45310", + "issue:38617|issue:44220", + "issue:43937|issue:44496", + "issue:44933|issue:45278", + "issue:43066|issue:43122", + "issue:42222|issue:45542", + "issue:44410|issue:45478", + "issue:44483|issue:45464", + "issue:42371|issue:42831", + "issue:43012|issue:43577", + "issue:44075|issue:45656", + "issue:42175|issue:42222", + "issue:43097|issue:44493", + "issue:43673|issue:45341", + "issue:41628|issue:45071", + "issue:44263|issue:44485", + "issue:45071|issue:45542" + ], + "prepared_review_unit_hash": "28929648be1f9614d2836d7689ec2cadbe958fd16249b2ce19a36f11804e1e48", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36530, + "estimated_input_tokens": 9133, + "estimated_eval_tokens": 18522 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly false positives across unrelated regressions, import errors, model-specific bugs, and docs/perf issues. Only the tokenizer pair 43066/43122 looks like the same v5 tokenizer regression.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 43066 is the clearest root-cause report for the tokenizer v5 regression; 43122 reads like the same symptom reported from a different angle.", + "canonical_pr_reason": null, + "best_issue_reason": "43066 is the best single issue to keep: it is specific, actionable, and can subsume the duplicate symptom in 43122.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Different model families and failures: missing projections in qwen3next vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both mention testing utils, but one is a vague bug report and the other is a specific xdist race on captured_info.txt; not clearly the same bug." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and code paths." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "Testing utils bug vs a padding performance issue in SequenceFeatureExtractor; unrelated." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "VitPose model breakage and tensorflow_text backend mapping are unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "Flash-attention training degeneration and Qwen3.5 save_pretrained checkpoint corruption are different bugs." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "A script launch failure is unrelated to a testing_utils bug." + }, + { + "left": "issue:45362", + "right": "issue:46328", + "accept": false, + "reason": "Qwen3.5 chat crash vs padding performance; no shared codepath." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leak in dataloader threads and an Apple Silicon warmup TypeError are separate issues." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "Different loading warnings/errors for different model types." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Tokenizer offset mapping bug vs runtime special-token post-processor update; related area but not the same defect." + }, + { + "left": "issue:42490", + "right": "issue:42673", + "accept": false, + "reason": "Local path slash import warning and VRAM leak are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config loading vs a documentation parameter-name mismatch." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API warning and MLA rms_norm_eps precision issue are unrelated." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "Shard-size question vs Qwen3.5 MoE loading error; different concerns." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for a missing config_utils symbol vs an fbank feature-extraction bug." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GenerationConfig validation failure and unrecognized model loading are different issues." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing image_utils import is not the same as a broad many-import-errors regression." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": true, + "reason": "Both point to the same v5 tokenizer regression: changed decoder/type behavior leading to different tokenization." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose model breakage and missing TensorFlow backend after tensorboard install are unrelated." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Same model family, but different concrete failures: missing projections vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Different chat/completions API failures and likely different request/streaming paths." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 behavior warning and FineGrainedFP8 accuracy regression are unrelated." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both involve dtype/precision, but one is a compile warning and the other is incorrect model dtype propagation on load." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "Optimizer args ignored and optimizer step called twice are distinct optimizer bugs." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow backend installation issue vs broken VitPose models; unrelated." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed tie_embeddings API and unexpected key warnings from position ids are different regressions." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing in chunked_prefill is unrelated to a testing_utils bug." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Import error for AutoImageProcessor vs PretrainedConfig type-checking regression; not the same bug." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "torch.split return-value issue in GlmMoeDsaIndexer vs GLM-5 RoPE implementation; different codepaths." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "PretrainedConfig type-checking regression and undefined tf backend error are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The clustering is mostly conservative and grounded in the packet. The only accepted pair, 43066/43122, is plausibly the same tokenizer-v5 regression based on the titles, and the other pair judgments stay on the cautious side of false unless there is stronger body-level evidence not shown here." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 47, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:27957", + "issue:38617", + "issue:42757", + "issue:42890", + "issue:43064", + "issue:43097", + "issue:43335", + "issue:43352", + "issue:43441", + "issue:43452", + "issue:43502", + "issue:43531", + "issue:43673", + "issue:43873", + "issue:43874", + "issue:43931", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44829", + "issue:44933", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45588", + "issue:45753", + "issue:46050", + "issue:46123", + "issue:46172", + "issue:46231", + "issue:46298", + "issue:46369" + ], + "soft_pairs": [ + "issue:27957|issue:43352", + "issue:43874|issue:44261", + "issue:44484|issue:45478", + "issue:44263|issue:46172", + "issue:43931|issue:44410", + "issue:44360|issue:46172", + "issue:43441|issue:44829", + "issue:43931|issue:44038", + "issue:43335|issue:45071", + "issue:43097|issue:43452", + "issue:45003|issue:45341", + "issue:43673|issue:46123", + "issue:43986|issue:45341", + "issue:44933|issue:46231", + "issue:45753|issue:46369", + "issue:44117|issue:44987", + "issue:43986|issue:44484", + "issue:43874|issue:44263", + "issue:43986|issue:45372", + "issue:45440|issue:46172", + "issue:43531|issue:45341", + "issue:43097|issue:46050", + "issue:44261|issue:44485", + "issue:43097|issue:45588", + "issue:43064|issue:44315", + "issue:38617|issue:42757", + "issue:44062|issue:44117", + "issue:44410|issue:45375", + "issue:44038|issue:44410", + "issue:43502|issue:43992", + "issue:42890|issue:46298", + "issue:38617|issue:44351", + "issue:43097|issue:43873" + ], + "prepared_review_unit_hash": "153be6571f80e530131df75892e1404c769ee256f50b6b994c369558a13da3b9", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36772, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly soft-similarity false positives: the paired issues generally share broad themes like FlashAttention, import errors, or model-loading failures, but not the same concrete bug or code path. I do not see a true duplicate pair here.", + "confidence": 0.93, + "canonical_issue_reason": "No single issue is a clear canonical duplicate target: the items span unrelated bugs across different models and subsystems. If a representative issue is needed, it should be a broad, active bug report rather than one of the narrow model-specific cases.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest standalone issue in the set: it is open, has substantial discussion/inbound references, and describes a substantive, broadly relevant bug rather than a narrow one-off failure.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2, but they concern different models and likely different model-compatibility checks." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different problems: a missing image-patch method in GLM46V versus an MLA layernorm config precision issue." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Unrelated: shard-size documentation/question versus a Qwen3.5 MoE loading error." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "Different code paths and models: GlmMoeDsaIndexer split handling versus DeepSeek comb transpose logic." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Both are Qwen-family model bugs, but one is a weight-shape mismatch and the other is missing projections in layer 0." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "DSA indexer ReLU concern is not the same bug as the DeepSeek-v4 comb dtype/transpose issue." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both involve FlashAttention, but one is a model-specific failure and the other is degenerate training with flash_attention_3." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Same general Qwen/VL area, but different failures: shape mismatch versus a Transformers v5 compatibility bug." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and PretrainedConfig type checking are unrelated bugs." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "An API-removal/documentation issue is not the same as gguf_file breaking tokenizer/model loading." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Different files and symptoms: unsafe sys.modules access versus a testing_utils bug." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Both affect generation/pipeline behavior, but the cache-missing chunked_prefill bug is distinct from the final-batch is_last bug." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "A video AutoProcessor import crash is unrelated to the testing_utils issue." + }, + { + "left": "issue:44933", + "right": "issue:46231", + "accept": false, + "reason": "Nonexistent image_utils import and a doc parameter-name mismatch are different classes of issues." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Different models and failures: Qwen3_5 infinite loop on one image versus SAM3-lite memory overflow." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer-mapping None handling is not the same as the physical-intelligence/fast loading failure." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "Completely different topics: torchvision-dependent processor loading versus shard-size behavior." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Different GLM46V vs GlmMoeDsaIndexer failures; same broad area only by naming." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Both are loading/import failures, but they involve different optional dependencies and different processors/models." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "Native DeepseekV3 divergence is a model-implementation mismatch, not the DeepSeek-v4 comb transpose bug." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the testing_utils bug." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "A v5 API removal issue and a quantization save/load regression are different bugs." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Both are model-specific implementation discussions, but the missing rms_norm_eps bug is distinct from GLM-5 RoPE implementation." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Tie-embeddings removal and flash_attention.py crashing on sink-less models are unrelated." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is a different trainer bug from Liger Kernel not being applied with model_init." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are ImportError reports, but they are for different missing symbols in different packages." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "AddedToken special-argument conflict and tokenizer-mapping None handling are separate tokenizer bugs." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Different Qwen3-family config/model issues: missing projections versus a silently dropped config field." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Both involve Qwen3 variants, but one is a Transformers v5 loading break and the other is missing layer projections." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "local_files_only network leakage and missing embed_tokens.weight are different from load-path bugs." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "Test flakiness from missing set_seed() is unrelated to a noisy logging message." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Two different import failures; they do not point to the same missing symbol or code path." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Tie-embeddings removal and quantization offloading problems are different subsystems and bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 48, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:36683", + "issue:38617", + "issue:42490", + "issue:42757", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43824", + "issue:43825", + "issue:43906", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44117", + "issue:44162", + "issue:44222", + "issue:44230", + "issue:44368", + "issue:44380", + "issue:44483", + "issue:44484", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:44704", + "issue:44936", + "issue:44938", + "issue:45103", + "issue:45478", + "issue:45479", + "issue:45600", + "issue:45712", + "issue:45753", + "issue:46132", + "issue:46172", + "issue:46178", + "issue:46298", + "issue:46365", + "issue:46421" + ], + "soft_pairs": [ + "issue:44704|issue:45103", + "issue:44380|issue:46421", + "issue:43986|issue:44117", + "issue:43824|issue:43986", + "issue:44938|issue:46178", + "issue:43066|issue:43906", + "issue:43208|issue:43550", + "issue:43992|issue:44484", + "issue:43012|issue:43550", + "issue:43012|issue:44368", + "issue:43825|issue:44655", + "issue:45712|issue:46298", + "issue:42907|issue:44655", + "issue:42907|issue:45753", + "issue:43097|issue:46298", + "issue:43352|issue:44380", + "issue:43352|issue:43976", + "issue:44162|issue:46172", + "issue:44380|issue:45479", + "issue:44380|issue:46132", + "issue:43064|issue:44380", + "issue:44483|issue:44514", + "issue:44222|issue:45478", + "issue:36683|issue:43582", + "issue:43208|issue:43526", + "issue:44936|issue:46365", + "issue:42490|issue:43012", + "issue:43012|issue:44683", + "issue:42757|issue:43824", + "issue:38617|issue:43502", + "issue:45600|issue:46298", + "issue:44038|issue:44230" + ], + "prepared_review_unit_hash": "c132f81ef0451e713aba86b97a079d4487e77844d69ddb1229c8902e47073b7b", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37093, + "estimated_input_tokens": 9274, + "estimated_eval_tokens": 18804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft pairs are overwhelmingly false positives: they span unrelated import errors, model-specific config regressions, backend/compile issues, and save/load behaviors. I don\u2019t see any pair that clearly shares the same underlying bug or change.", + "confidence": 0.93, + "canonical_issue_reason": "Issue #43824 is the cleanest high-signal regression report here: a specific ImportError, a clear user-facing failure mode, and multiple inbound references. It is more concrete and representative than the other candidates.", + "canonical_pr_reason": null, + "best_issue_reason": "#43824 is the strongest single issue to anchor this set because it has a precise failure, visible impact, and enough discussion/references to be actionable.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Different failures: kwargs propagation into cached_file vs an auto-docstring crash under future annotations." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "Both touch model/runtime behavior, but one is SDPA/FlashAttention scaling semantics and the other is cache lazy initialization/recompilation." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "One is an AutoProcessor crash without torchvision; the other is TOKENIZER_MAPPING_NAMES returning None. Different code paths." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Both are import/loading-related, but one is a missing model class export and the other is a video-model AutoProcessor crash tied to torchvision." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Different platform/backend issues: Python 3.14 import failure vs an invalid BACKENDS_MAPPING entry for tensorflow_text." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "No evidence they share the same bug; 43906 is only a reproduction issue, not the same tokenizer-decoder regression." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training bugs and Bamba torch.compile/SDPA failure are unrelated model-specific issues." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing embed_tokens weights during from_pretrained is unrelated to the max_shard_size default in save_pretrained." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "A PyTorch precision warning on bfloat16 is not the same as a compile-time SDPA failure." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "A warning about float32 precision is unrelated to tie_word_embeddings/tie_embeddings warnings in Qwen3.5 fine-tuning." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "One is a misleading pipeline task message; the other is failure to save Pipeline objects. Different bugs." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Left is leftover dummy class cleanup; right is noisy logging. Same area of hygiene, not same defect." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals and saving Pipeline objects are different save_pretrained failures." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "A dequantized save failure is unrelated to an infinite loop on a specific image input." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder behavior is a functional API change, not a generic noisy logging issue." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Nemotron FlashAttention-2 support and GPT2 attention scaling under SDPA/FlashAttention are different backend/model problems." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "FlashAttention support for a model is unrelated to Python version compatibility in Transformers 5.1.0." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage and a deepseek-v4 tensor expression bug are different model-specific regressions." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention scaling under backend dispatch is unrelated to a zero-loss classification bug." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "GPT2 scaling under SDPA/FlashAttention is not the same as AttentionInterface.register changing registered-function behavior." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to attention scaling being ignored under backend kernels." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat/completions request rejection and Qwen2_5_VLProcessor batching crash are different API surfaces and failures." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5 MoE, but one is FP8 save_pretrained and the other is from_pretrained loading; not the same concrete bug." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config missing vocab_size and Apple Silicon allocator warmup TypeError are unrelated." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training blockers and BeitImageProcessorFast reduce_labels returning one label are unrelated." + }, + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "trainer.evaluate() after train() failing and inability to import Trainer with datasets installed are different trainer/import issues." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Trailing-slash local model path import warning is unrelated to a bfloat16 precision warning." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "Precision warning on bfloat16 and compiled flex_attention failing on torch>=2.9 are different bugs." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Missing is_offline_mode and missing Qwen2_5_VLForConditionalGeneration are unrelated import errors." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "layer_type_validation import removal and API requests despite local_files_only=True are different regressions." + }, + { + "left": "issue:45600", + "right": "issue:46298", + "accept": false, + "reason": "Broken auto_mappings CI references and noisy logging message are not the same issue." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Qwen3-VL-Moe bug in Transformers 5.0 and FP8 inference support for Qwen3.5 MoE are different requests/failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 49, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:42175", + "issue:42757", + "issue:42994", + "issue:43010", + "issue:43064", + "issue:43066", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43611", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44075", + "issue:44077", + "issue:44242", + "issue:44355", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44492", + "issue:44625", + "issue:44655", + "issue:44855", + "issue:44938", + "issue:45216", + "issue:45276", + "issue:45278", + "issue:45290", + "issue:45310", + "issue:45375", + "issue:45584", + "issue:45593", + "issue:45698", + "issue:46077", + "issue:46082", + "issue:46172", + "issue:46202", + "issue:46209", + "issue:46365" + ], + "soft_pairs": [ + "issue:45698|issue:46172", + "issue:43618|issue:44380", + "issue:43618|issue:43761", + "issue:44483|issue:45290", + "issue:43299|issue:43867", + "issue:43756|issue:44485", + "issue:36683|issue:43475", + "issue:42994|issue:44655", + "issue:43066|issue:46209", + "issue:45276|issue:46077", + "issue:43901|issue:44655", + "issue:36010|issue:42757", + "issue:43611|issue:44655", + "issue:43352|issue:43937", + "issue:36683|issue:44077", + "issue:43867|issue:45216", + "issue:44625|issue:46077", + "issue:43976|issue:44938", + "issue:45584|issue:46365", + "issue:42175|issue:46365", + "issue:43352|issue:45593", + "issue:46077|issue:46082", + "issue:36683|issue:43525", + "issue:43618|issue:45216", + "issue:44077|issue:46365", + "issue:36683|issue:44242", + "issue:44075|issue:46365", + "issue:43992|issue:45310", + "issue:43064|issue:44492", + "issue:43010|issue:45278", + "issue:44355|issue:44855", + "issue:43867|issue:46202", + "issue:43867|issue:45375" + ], + "prepared_review_unit_hash": "49e00c70d9d71d2393921f79d237f13bd3eeb6af3403af66826955288f8c5c60", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37169, + "estimated_input_tokens": 9293, + "estimated_eval_tokens": 18842 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous: many items are Transformer v5 regressions, but they touch different models, configs, and code paths. I did not find any soft pair that looks like the same underlying bug/change.", + "confidence": 0.29, + "canonical_issue_reason": "Issue 45698 is the broadest and most actively discussed remaining issue, centered on save/load behavior for custom modules after `save_pretrained`/`from_pretrained`. It is the best anchor only for that subset, not for the whole mixed cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "45698 is the strongest standalone issue in the set: open, high discussion activity, and a concrete reproducible save/load regression with clear user impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45698", + "right": "issue:46172", + "accept": false, + "reason": "Different bugs: custom module reload after saving vs a DeepSeek-V4 tensor transpose/dtype issue." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Both mention attention, but one is missing CLIPOutput attentions and the other ignores GPT-2 attention scaling under SDPA/FlashAttention; different code paths." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Same model family, but separate regressions: CLIPOutput.attentions not assigned vs CLIPVisionModel.hidden_states becoming None." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "API request validation failure vs chat template tokenization crash; not the same bug." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3VL MoE loading failure and a generic sorted-state_dict load error are different root causes." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Smollm3 RoPE layer-count mismatch and GLM-5 RoPE implementation discussion are not the same concrete defect." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Different missing attributes on different objects/models: `Gemma3Config.vocab_size` vs `Sam3VisionEncoderOutput.fpn_position_embeddings`." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model saving failure and pipeline `save_pretrained` failure are related only broadly to saving, but not the same bug." + }, + { + "left": "issue:43066", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer decoder-type regression vs FP8 inference failure on B200; unrelated symptoms and subsystems." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Both are Gemma4 issues, but one is resize_token_embeddings propagation and the other is bidirectional attention-mask construction." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Documentation mismatch for `return_all_scores` is not the same as a runtime pipeline serialization bug." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols from different packages (`transformers.generation` vs `huggingface_hub`)." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Different loading/saving problems: `base_model_prefix` handling vs pipeline serialization." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Flash Attention support gating for Nemotron is unrelated to GLM-5 generation-config validation." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Missing config attribute vs a restriction on optional `post_init`; not the same failure." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both involve model loading/saving, but one is state_dict ordering and the other is incorrect Qwen3.5 checkpoint contents after `save_pretrained`." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "`num_labels` propagation in Qwen3.5 is unrelated to Gemma4 attention-mask behavior." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Different Python-version compatibility bugs on different interpreter versions and with different causes." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper generation after special-token alignment is unrelated to `Trainer` import being blocked by `datasets`." + }, + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Dependency/backend installation problem vs import conflict with `datasets`; not the same underlying bug." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "Flash Attention unsupported warning and missing auxiliary losses in D-FINE are unrelated." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Gemma4 attention-mask bug vs LlamaConfig head-dim validation are different model/config problems." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Both are missing config attributes, but they affect different models and different fields; too broad to merge." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions missing and Qwen3.5 saved checkpoint corruption are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "`patchtsmixer` init-policy change is unrelated to a `Trainer` import conflict caused by `datasets`." + }, + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Missing `vocab_size` and missing load-balancing loss are different bugs in different subsystems." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "Optimizer argument handling has nothing to do with `Trainer` import resolution." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Different models and different errors: UMT5 missing weights vs Qwen3.5 MoE `from_pretrained` failure." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Distributed optimizer-state bug and a typo in cache strategies are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "`@torch.no_grad` decoration for cache updates is not the same as broad import failures after upgrade." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Compiled-file import errors and a Python 3.13 decorator/comment parsing regression are different root causes." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Generic sorted-state_dict load failure vs Gemma pipeline roundtrip on the underscore token are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "State_dict ordering issue and missing `deepstack_visual_indexes` config field are different bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 50, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:28282", + "issue:30333", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42548", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43065", + "issue:43232", + "issue:43502", + "issue:43575", + "issue:43576", + "issue:43618", + "issue:43723", + "issue:43867", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44246", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44625", + "issue:44749", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:45071", + "issue:45125", + "issue:45357", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45561", + "issue:45600", + "issue:45850", + "issue:46132", + "issue:46285", + "issue:46365" + ], + "soft_pairs": [ + "issue:41669|issue:44246", + "issue:44938|issue:45071", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:43010|issue:45850", + "issue:44625|issue:46365", + "issue:43992|issue:45478", + "issue:43867|issue:45357", + "issue:43010|issue:43976", + "issue:36296|issue:43575", + "issue:43618|issue:44077", + "issue:28282|issue:33453", + "issue:43618|issue:46285", + "issue:43618|issue:44857", + "issue:44360|issue:45440", + "issue:43502|issue:44062", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:42890|issue:45600", + "issue:43723|issue:44556", + "issue:44387|issue:44556", + "issue:44222|issue:45125", + "issue:43010|issue:43065", + "issue:43576|issue:46132", + "issue:43502|issue:44485", + "issue:42548|issue:44351", + "issue:45125|issue:45478", + "issue:44355|issue:44749", + "issue:30333|issue:33453", + "issue:43010|issue:43232", + "issue:45125|issue:45375" + ], + "prepared_review_unit_hash": "c02bf1e946eac8e1896227e8dae61b7a841a41067e6edafb5af497350b3c06a9", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36541, + "estimated_input_tokens": 9136, + "estimated_eval_tokens": 18528 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose mix of unrelated Transformers v5 regression reports. The soft-edge pairs share broad themes like import/loading, quantization, tensor parallelism, or test failures, but none look like the same concrete bug or change.", + "confidence": 0.95, + "canonical_issue_reason": "issue:42548 is the strongest representative because it describes a broad, user-facing core import/export break in the package, which is more central than the narrower model- or test-specific reports here.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42548 is the best single issue to anchor the cluster since it reflects a high-impact API regression and is easy to recognize as a canonical core-package failure mode.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Both concern import-related performance, but one is about `import *` slowing import time and the other about intermittent slow `import transformers`; not the same bug." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 load failure and `PretrainedConfig` type-checking break are different compatibility issues." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "A missing `vocab_size` on `Gemma3Config` and a FSDP2/PEFT optimizer-state mismatch are unrelated problems." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Seed-dependent integration flakiness and a pytest-xdist race in testing utils are different test failures." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "Decorating cache/layer `update()` with `torch.no_grad` is a specific autograd fix; \u201cbreaks too many models\u201d is far broader and not the same defect." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Missing `num_labels` propagation in Qwen3.5 is a model config bug; inability to import `Trainer` until uninstalling `datasets` is an import dependency issue." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "`UMT5Encoder` missing `embed_tokens.weight` and Qwen3.5 MoE `from_pretrained` error are distinct model-loading failures." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "A load error when state dict keys are sorted and incorrect visual encoder keys saved by Qwen3.5 are not the same bug." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "`torch.no_grad` on cache updates and Python-version support in 5.1.0 are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug and OOM when loading a specific Qwen2 model with TP are related at most broadly, but not the same concrete issue." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "Missing `CLIPOutput.attentions` assignment and `patchtsmixer` post_init policy are separate model API changes." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch-missing import error for `AutoModel` and tokenizer-loading regression are different code paths." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "Both mention CLIP-related structures, but one is an output field regression and the other is inconsistent model class structure; not the same bug." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "CLIP attentions regression and `LwDetrImageLoss` AMP crash are unrelated model bugs." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "A discussion about a DSA indexer activation function and a DeepSeekV3 implementation divergence are different topics." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "`local_files_only=True` still making network requests and `AddedToken()` duplicate `special` kwarg are unrelated tokenization/loading issues." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Missing `set_seed()` in tests and failure to load a specific multimodal model are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "A flaky test due to missing seeds and a CI stale device-override failure are different test issues." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Seed-related integration test instability and auto_mappings references to removed configs are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer loading regression and checkpoint reload failure after upgrade are different loading problems." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization memory growth/OOM and checkpoint reload failure are not the same bug." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 `save_pretrained` behavior for MoE and missing `_tp_plan` for tensor parallelism are separate features/bugs." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "`torch.no_grad` decoration for cache updates and a dummy `nn.Conv2d` in Sam3PixelDecoder are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "A broken `transformers env` command and `AttentionInterface.register` behavior change are distinct issues." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Unexpected network requests with `local_files_only=True` and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Both are import failures, but one is `PreTrainedModel` export/import and the other is `HybridCache`; different symbols and likely different causes." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Missing `_tp_plan` and Qwen3.5 MoE `from_pretrained` failure are related only by model family, not the same concrete bug." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file errors and a slowdown when filtering data after upgrading are not the same issue." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLflow reporting on failed training jobs and tokenizer loading regression are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache/layer update autograd decoration and `_update_model_kwargs_for_generation` after `sync_gpus` are different generation-path bugs." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Missing tensor-parallel plan and a missing `deepstack_visual_indexes` field dropped by strict config handling are different config/model issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 51, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:27957", + "issue:30333", + "issue:30990", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:43299", + "issue:43502", + "issue:43526", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:43901", + "issue:44038", + "issue:44222", + "issue:44230", + "issue:44273", + "issue:44315", + "issue:44355", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44908", + "issue:44938", + "issue:44961", + "issue:45071", + "issue:45125", + "issue:45230", + "issue:45362", + "issue:45468", + "issue:45542", + "issue:45850", + "issue:46095", + "issue:46143", + "issue:46144", + "issue:46209" + ], + "soft_pairs": [ + "issue:30990|issue:31795", + "issue:43502|issue:44222", + "issue:43299|issue:44556", + "issue:43502|issue:44961", + "issue:30333|issue:36296", + "issue:42831|issue:43874", + "issue:44355|issue:45071", + "issue:43827|issue:44961", + "issue:44355|issue:44655", + "issue:44315|issue:44360", + "issue:33453|issue:34689", + "issue:27957|issue:33453", + "issue:42757|issue:43576", + "issue:43576|issue:45071", + "issue:43576|issue:43824", + "issue:31795|issue:38617", + "issue:44961|issue:45468", + "issue:44222|issue:44230", + "issue:43576|issue:44938", + "issue:41628|issue:42548", + "issue:46143|issue:46144", + "issue:44908|issue:44961", + "issue:44485|issue:44961", + "issue:36296|issue:43901", + "issue:44315|issue:46209", + "issue:36296|issue:43526", + "issue:43673|issue:44492", + "issue:45362|issue:45850", + "issue:44623|issue:45230", + "issue:44273|issue:45542", + "issue:44038|issue:45125", + "issue:44230|issue:46095", + "issue:44393|issue:45071" + ], + "prepared_review_unit_hash": "1153933e6143ed68598457792aa529f41851a6def46cf4d1a4cc52fdd092e015", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 36712, + "estimated_input_tokens": 9178, + "estimated_eval_tokens": 18612 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues, with a few broad v5-regression reports mixed in. None of the soft pairs look like true duplicates under a strict same-bug/same-change standard.", + "confidence": 0.58, + "canonical_issue_reason": "issue:45850 is the broadest umbrella in the set: it explicitly summarizes the general v5 breakage theme and can absorb the various model-loading/regression complaints better than the narrower, model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the best global representative for the cluster because it is the only high-level issue that spans the repeated 'v5 breaks X' pattern rather than a single model or subsystem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Sentence Transformers loading hangs and forward-docs confusion are unrelated problems." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "local_files_only network leakage and FP8 moe save_pretrained are different code paths." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are load regressions, but they affect different models and likely different failure points." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "API requests despite local_files_only and a unrelated racoon-named issue are not the same bug." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLflow reporting on failure and tensor-parallel training are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "FineGrainedFP8 accuracy and a missing image-patch helper in GLM46V are distinct defects." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled Python file errors and PretrainedConfig type checking are different regressions." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Pipeline docs cleanup and the racoon issue share no underlying change." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled-file handling and Pipeline.save_pretrained are separate failures." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel application via model_init and a DSA ReLU discussion are unrelated." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer-loading regression and Llama 3.2 vision model loading are not the same concrete bug." + }, + { + "left": "issue:27957", + "right": "issue:33453", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to tokenizer loading regression." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "missing is_offline_mode import and a broken env command are different issues." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "The env CLI breakage and PretrainedConfig type checking have no common code-path." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Broken env command and missing Qwen2.5-VL export are unrelated." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Documentation ambiguity and a missing symbol import are not duplicates." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "Racoon and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained for moe and FP8 inference support are different features/failures." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "env CLI breakage and Python 3.14 load failure are distinct regressions." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Missing AutoImageProcessor and missing PreTrainedModel exports are similar in shape but not clearly the same underlying bug." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "One is kwargs not being forwarded, the other is the RoFormer attention implementation not using the attention interface." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "inverse_sqrt scheduler kwargs handling and racoon are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation and racoon have no shared failure mode." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "tensor-parallel training and TextClassificationPipeline docs mismatch are unrelated." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Liger Kernel not applied with model_init and FP8 inference on B200 are different bugs." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "tensor-parallel training and BeitImageProcessorFast reduce_labels are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "chunked_prefill cache handling and a cache-strategies typo are not the same issue." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A single Qwen3.5-35B chat crash is narrower than the broad v5 breakage umbrella." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "processor.save_pretrained missing files and a generic bug report title are not enough to match as duplicates." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "lazy loading and a TensorFlow backend installation error are different problems." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "generic Qwen3-VL-Moe breakage and missing _tp_plan for tensor parallelism are not the same concrete defect." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support and a DeepSeek weight-conversion question are unrelated." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output issues and PretrainedConfig type checking do not share a code path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 52, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:34689", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43576", + "issue:43673", + "issue:43704", + "issue:43827", + "issue:43976", + "issue:44077", + "issue:44246", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:45003", + "issue:45071", + "issue:45092", + "issue:45230", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45478", + "issue:45753", + "issue:45850", + "issue:46018", + "issue:46172", + "issue:46218", + "issue:46328" + ], + "soft_pairs": [ + "issue:44273|issue:45003", + "issue:43576|issue:44556", + "issue:44393|issue:45753", + "issue:43525|issue:45850", + "issue:41669|issue:43673", + "issue:44556|issue:45092", + "issue:43572|issue:45850", + "issue:44492|issue:45850", + "issue:45278|issue:45850", + "issue:42371|issue:44492", + "issue:41669|issue:46218", + "issue:42890|issue:44908", + "issue:43408|issue:44315", + "issue:44297|issue:45230", + "issue:44246|issue:44273", + "issue:43352|issue:44393", + "issue:44295|issue:45230", + "issue:43576|issue:44908", + "issue:44704|issue:44908", + "issue:43976|issue:44393", + "issue:44908|issue:45071", + "issue:41669|issue:44351", + "issue:44492|issue:46328", + "issue:41669|issue:45341", + "issue:43976|issue:45850", + "issue:44393|issue:45478", + "issue:44077|issue:44861", + "issue:43976|issue:46218", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:44315|issue:44829", + "issue:43576|issue:44273", + "issue:44655|issue:44861", + "issue:45230|issue:45310", + "issue:44360|issue:45468", + "issue:44273|issue:44908", + "issue:43576|issue:46218", + "issue:43519|issue:44485", + "issue:46018|issue:46172", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689" + ], + "prepared_review_unit_hash": "8b6312b4a650ed82ca56e8b52b95e28da0dc0d58394fca4cd0ad22e9ef59e618", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 42, + "serialized_chars": 39546, + "estimated_input_tokens": 9887, + "estimated_eval_tokens": 20030 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a mixed set of Transformers v5-era regressions and adjacent docs/perf/model-specific bugs. The items are thematically related by version upgrades, loading/import behavior, and model-specific breakages, but there are no strong duplicate pairs with the evidence provided.", + "confidence": 0.71, + "canonical_issue_reason": "Issue 45278 is the best representative of the cluster because it states a concrete, broad v5 upgrade regression with many import failures, which matches the cluster\u2019s dominant theme better than the narrower or more vague reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45278 is the most suitable global representative: it is concrete, user-facing, and broadly captures the upgrade breakage pattern without being as vague as the more generic complaints.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Both touch lazy-loading/import behavior, but one is about lazy loading not working and the other about unsafe sys.modules access/performance; not the same bug." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "One is a broken `transformers env` command, the other is checkpoint reload failure after upgrading; different surfaces and code paths." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Both are Qwen3-VL problems, but one is 2D box hallucination/error and the other is an infinite loop on a specific image; too different to merge." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "A missing `pad_token_id` in `Llama4Config` is a specific config bug; the other is a broad complaint that v5 breaks too many models." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Removing `import *` for import-time cost is unrelated to missing `GenerationMixin` cache during chunked prefill." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are upgrade incompatibilities, but they affect different models and failure modes; not the same concrete loading bug." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "A missing `pad_token_idx` in one config is not the same as a broad statement that many models break in v5." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A typo in cache strategies is a narrow docs/code issue, not the same as a general v5 breakage report." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Both are broad regression reports, but they are not the same underlying failure; one is import errors on upgrade, the other is a generic complaint about many models." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API settings and a cache-strategy typo are unrelated changes." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "One is about import-star cleanup and import-time cost; the other is a missing `Qwen3_5ForCausalLM` import error." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Missing `set_seed()` in tests is unrelated to scheduler kwargs being ignored." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "A model-type warning for SAM3 is unrelated to Liger Kernel not being applied during `model_init`." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer class mismatch on save is a specific serialization bug; the other is an unspecified generic bug report." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Import slowness and lazy-loading failure are related at a high level, but not clearly the same defect." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support gating for Nemotron is unrelated to Qwen3-VL bounding-box errors." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Reading a buffer-registered `position_ids` and a generic bug report do not describe the same issue." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "A broken env command and an inverse_sqrt scheduler kwargs bug are unrelated." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "`AutoProcessor.from_pretrained` kwargs propagation and scheduler kwargs handling are different code paths." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version support breakage is unrelated to Qwen3-VL 2D box output errors." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Ignored scheduler kwargs and PretrainedConfig type checking are unrelated regressions." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-star cleanup/import cost is not the same as failing to import `HybridCache`." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "A cache-strategy typo and a SequenceFeatureExtractor.pad performance problem are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Model import cleanup/perf is unrelated to a bug in `testing_utils.py`." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Python compatibility breakage is not the same as the vague claim that too many models break." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen model families and different symptoms; not the same bug." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "A `post_init` policy issue in patchtsmixer is unrelated to `_get_tied_weight_keys` crashing on a list." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python version incompatibility and a missing Qwen3_5 import are different problems." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing `pipeline()` after removals is unrelated to scheduler kwargs." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion about forward args is unrelated to a tokenizer loading regression." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel not being applied and flash_attention_3 causing degenerate training are different implementation issues." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "A broken env command and lazy loading not working are both v5-adjacent, but not the same specific bug." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Saving Pipeline objects and tied-weight key extraction are unrelated failures." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "A generic bug report and a specific Qwen3.5 MoE from_pretrained failure are not the same issue." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer ReLU discussion is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy-loading issues and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "A broken env command is unrelated to a Qwen3_5 import failure." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are different model-specific bugs." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "Both are DeepSeek-V4 related, but one reports gating semantics and the other a tensor transpose/casting issue; not clearly the same bug." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "A multimodal timestamp bug and a scheduler kwargs bug are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star cleanup/import time and a VRAM leak in dataloader threads are different problems." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Docs wording around `input_ids`/`past_key_values` is unrelated to a model-loading regression for Llama 3.2 Vision." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 19, + "review_unit_index": 53, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-450", + "nodes": [ + "issue:30333", + "issue:31515", + "issue:31795", + "issue:33453", + "issue:36296", + "issue:37934", + "issue:38617", + "issue:42371", + "issue:42913", + "issue:42936", + "issue:43066", + "issue:43240", + "issue:43298", + "issue:43352", + "issue:43502", + "issue:43519", + "issue:43525", + "issue:43572", + "issue:43606", + "issue:43618", + "issue:43668", + "issue:43704", + "issue:43717", + "issue:43813", + "issue:44031", + "issue:44230", + "issue:44261", + "issue:44263", + "issue:44351", + "issue:44485", + "issue:44492", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45290", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46144", + "issue:46218" + ], + "soft_pairs": [ + "issue:31515|issue:31795", + "issue:43704|issue:44485", + "issue:44485|issue:45468", + "issue:43298|issue:46095", + "issue:45850|issue:46218", + "issue:43502|issue:43519", + "issue:42371|issue:43704", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:43606|issue:45468", + "issue:44263|issue:45468", + "issue:44261|issue:45468", + "issue:44351|issue:46218", + "issue:43352|issue:46218", + "issue:42913|issue:43066", + "issue:44908|issue:44998", + "issue:44492|issue:44998", + "issue:44998|issue:45850", + "issue:43618|issue:44998", + "issue:44861|issue:45290", + "issue:44998|issue:45341", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:30333|issue:31515", + "issue:31515|issue:36296", + "issue:44230|issue:45310", + "issue:31515|issue:33453", + "issue:43298|issue:44230", + "issue:43717|issue:46144", + "issue:42936|issue:46018", + "issue:43668|issue:44031", + "issue:42936|issue:43240", + "issue:43240|issue:46144", + "issue:42936|issue:44031", + "issue:42936|issue:46144", + "issue:42936|issue:43813", + "issue:43240|issue:43668", + "issue:43668|issue:43813", + "issue:37934|issue:43668", + "issue:42936|issue:43668" + ], + "prepared_review_unit_hash": "62348282138d5ec4e8049bcb03c84c756f1e0ebd35842297ba63c46edc6c2764", + "component_budget": { + "node_count": 450, + "item_count": 450, + "soft_pair_count": 1664, + "serialized_chars": 861347, + "estimated_input_tokens": 215337, + "estimated_eval_tokens": 430930 + }, + "budget": { + "node_count": 44, + "item_count": 44, + "soft_pair_count": 42, + "serialized_chars": 37123, + "estimated_input_tokens": 9281, + "estimated_eval_tokens": 18818 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Loose cluster of mostly unrelated Transformers bug reports, with a weak theme around model/tokenizer loading regressions. None of the soft pairs are true duplicates under a strict triage standard.", + "confidence": 0.41, + "canonical_issue_reason": "issue:33453 is the clearest concrete regression report in the tokenizer-loading family and is more representative than the broader or more vague items.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:33453 is the strongest standalone issue to anchor this cluster: it is specific, user-facing, and closest to the small tokenizer-regression subtheme here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Different bugs: slow checkpoint loading vs documentation confusion about input_ids/past_key_values." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "Different code paths: VRAM leak in dataloader-thread usage vs GLM-5 RoPE implementation question." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Both concern model internals, but the specific bugs are unrelated: RoPE math vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Feature/support request mismatch: continuous batching for audio models vs DeepSeek-V4 weight conversion tooling." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "Too broad vs specific import failure; not the same underlying bug." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Local-files-only network leakage vs incorrect timestamp calculation in a processor." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "Unrelated areas: TF32 API behavior vs VRAM leak in dataloader-threaded inference." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Different processor bugs: timestamp calculation vs missing files from save_pretrained." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Different failures in different code paths: processor timestamp math vs multi-thread VRAM leak." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU-offload device mismatch is not the same bug as Gemma-4 audio positional encoding." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "Different model logic: torch.split return handling vs Gemma-4 audio positional encoding." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "RMS norm epsilon precision issue is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "Importing HybridCache and importing Qwen3_5ForCausalLM are separate import-path failures." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash-Attention support gating for Nemotron is unrelated to a Qwen3_5 module import error." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Related tokenizer-v5 theme, but one is a behavior regression and the other is a wrong decoder type; not clearly the same bug." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Inverse-sqrt scheduler kwargs bug has nothing to do with the unrelated unemployment issue." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "Typo in cache strategies is unrelated to the unemployment placeholder issue." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "Unemployment placeholder vs broad model-breakage complaint; no shared code path." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to the unemployment placeholder issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weights utility crash vs chat-template crash with tool calls; different functions and symptoms." + }, + { + "left": "issue:44998", + "right": "issue:45341", + "accept": false, + "reason": "Unemployment placeholder vs a testing_utils bug; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is unrelated to _get_tied_weight_keys crashing on list input." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated to tied-weight key handling." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Checkpoint loading slowdown and missing layer_type_validation import are unrelated issues." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support request 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 checkpoint loading speed." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint loading slowdown vs tensor-parallel training bug; different failure modes." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 support request is not the same as the Qwen3.5 MoE from_pretrained loading error." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Both mention loading, but one is a performance complaint about checkpoints and the other is a tokenizer-loading regression." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models and FP8 inference support are distinct feature requests." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "init_weights behavior in Mamba models is unrelated to RoFormer attention-interface usage." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2former ignore_value handling and DeepSeek-V4 shared-expert gating are different model-specific bugs." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "Incorrect type hint for norm_eps is unrelated to tokenizer regex warning noise." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2former ignore_value usage and kwargs not passed to loss calculation are separate issues." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss-function kwargs plumbing is unrelated to RoFormer attention-interface implementation." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Different problem areas: segmentation ignore_value vs tokenizer regex warnings." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Different model bugs with no shared concrete code path." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Mask2former ignore_value bug is unrelated to a quantization-operation typo." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss kwargs propagation and a config type-hint typo are unrelated." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "Norm_eps type hint and quantization-operation typo are different, isolated fixes." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Llama4TextL2Norm/RMS-norm question is unrelated to ModernBERTConfig type-hint correctness." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Ignore_value handling and norm_eps typing are unrelated model/config 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": 19, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10519-3", + "nodes": [ + "issue:10519", + "issue:11005", + "issue:11307" + ], + "soft_pairs": [ + "issue:10519|issue:11307", + "issue:10519|issue:11005", + "issue:11005|issue:11307" + ], + "prepared_review_unit_hash": "ff007f9e3a3f74e44d0abc898d94eb227657988274d7e73c30b9c000c749dd0e", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2877, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2877, + "estimated_input_tokens": 720, + "estimated_eval_tokens": 1696 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster does not appear to contain duplicates: the three issues are about unrelated topics (truncation behavior, ReduceLROnPlateau-like scheduling, and Wav2Vec2 word time offsets).", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: the items describe different feature requests in different parts of the library, so none is a true representative of the others.", + "canonical_pr_reason": null, + "best_issue_reason": "If one must be selected, issue 11307 is the most discussion-heavy, but it is still not a good cluster representative because it is unrelated to the other two.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different requests: token truncation direction vs Wav2Vec2 word boundary offsets; no same underlying bug/change." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Completely different features: truncation option vs ReduceLROnPlateau-like functionality." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "Unrelated topics: learning-rate scheduling behavior vs audio token/word offset extraction." + } + ] + }, + "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": 19, + "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": "a7f570ce78e4ef1966634cd978ab3fcd91a368937819f00523bd91b16cc1eec0", + "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": "The cluster is not a true duplicate set: one issue is about tokenization behavior, one about private-model auth/download errors, and one about a general model-loading failure. The closest canonical issue is the model-loading report, but the others are distinct bugs.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 16618 ('Can't load the model for 'bert-base-uncased'.') is the broadest and most central model-loading failure in the set, so it is the best anchor for the cluster even though the other issues are not true duplicates of it.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 16618 is the best representative because it concerns a concrete loading failure and is the most likely duplicate target among these reports; the others are narrower and clearly about different code paths.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tokenization mismatch vs. model loading failure are different underlying bugs and different code paths." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Private-model download/auth error is not the same as failing to load bert-base-uncased; they may both involve loading, but the failure mode and fix are different." + } + ] + }, + "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": 19, + "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": "5a4d3ca3ecf13c79859842add75908b045e1024341ef657119dffc2fd50219f9", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are both broad community-contribution requests, but they target different features: Flash Attention 2 support versus GGUF support. They do not describe the same underlying bug or change, so they should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different feature requests in different code paths (Flash Attention 2 vs GGUF). The shared wording about adding support for more architectures is too generic to indicate the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 19, + "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-26413-2", + "nodes": [ + "issue:26413", + "issue:31501" + ], + "soft_pairs": [ + "issue:26413|issue:31501" + ], + "prepared_review_unit_hash": "7f63e11b860fea25c0fe90c33eeafbc80aaeb7fd31a1d2c4f8c69a9f291e694a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1594, + "estimated_input_tokens": 399, + "estimated_eval_tokens": 1054 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1594, + "estimated_input_tokens": 399, + "estimated_eval_tokens": 1054 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues both concern training flow failures, but they describe different symptoms and code paths: one is a checkpoint-resume iterator crash, the other is a training job that never starts for FineWeb SLM. I would not treat them as duplicates.", + "confidence": 0.48, + "canonical_issue_reason": "Issue 26413 is the more specific and actionable report, centered on a concrete `resume_from_checkpoint` failure and a clear error message.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 26413 is the best representative because it names a precise failure mode and likely narrower root cause than the broad 'training doesn't start' report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "The reports do not describe the same underlying bug: one is a resume/checkpoint epoch-iterator crash, the other is a broader training-start failure on FineWeb. Shared training context alone is not enough to merge them." + } + ] + }, + "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": 19, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-27088-8", + "nodes": [ + "issue:27088", + "issue:28936", + "issue:38435", + "issue:43834", + "issue:44365", + "issue:44370", + "issue:46222", + "issue:46226" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:27088|issue:38435", + "issue:27088|issue:28936", + "issue:28936|issue:43834", + "issue:28936|issue:44365", + "issue:28936|issue:38435", + "issue:28936|issue:44370", + "issue:38435|issue:44370", + "issue:38435|issue:44365", + "issue:38435|issue:43834", + "issue:38435|issue:46222", + "issue:38435|issue:46226", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:43834|issue:46222", + "issue:43834|issue:46226", + "issue:44365|issue:44370", + "issue:44365|issue:46222", + "issue:44365|issue:46226", + "issue:44370|issue:46222", + "issue:44370|issue:46226", + "issue:46222|issue:46226" + ], + "prepared_review_unit_hash": "5dda4b39e45831de65a7d22e4605e8c18380a20be289630bf13240754dd3bdb1", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13524, + "estimated_input_tokens": 3381, + "estimated_eval_tokens": 7018 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13524, + "estimated_input_tokens": 3381, + "estimated_eval_tokens": 7018 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of distinct i18n translation-request issues for different languages, so they should not be merged across languages. The only true duplicates are the near-identical generic placeholder issues ([i18n-]...) and the identical Thai pair.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue for the whole cluster: most items are separate language-specific translation requests. The generic placeholder issues form a smaller duplicate subgroup only.", + "canonical_pr_reason": null, + "best_issue_reason": "43834 is the best representative of the generic placeholder duplicate subgroup because it is one of the earliest near-identical template issues, but it is not a global canonical for the full cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different target languages; same translation-request template, but not the same underlying issue." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different target languages; not the same underlying translation request." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Turkish request vs generic placeholder/template issue; not the same concrete issue." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Different target languages; these are separate translation requests." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish vs Spanish translation request; distinct issues." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish request vs generic placeholder/template issue; not the same underlying issue." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different target languages; separate translation requests." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish vs Romanian translation request; distinct issues." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Different target languages; not duplicates." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian request vs generic placeholder/template issue; not the same issue." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian request vs generic placeholder/template issue; not duplicates." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian request vs generic placeholder/template issue; distinct translation tasks." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Different target languages/content; not the same underlying request." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Different target languages/content; not duplicates." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Near-identical generic placeholder/template issue; same underlying translation-request stub." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Near-identical generic placeholder/template issue; same underlying translation-request stub." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Generic placeholder/template issue vs different Thai-language issue; not the same underlying task." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Generic placeholder/template issue vs different Thai-language issue; not duplicates." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Near-identical generic placeholder/template issue; same underlying translation-request stub." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Generic placeholder/template issue vs different Thai-language issue; distinct issues." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Generic placeholder/template issue vs different Thai-language issue; not duplicates." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Generic placeholder/template issue vs different Thai-language issue; not the same underlying request." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Generic placeholder/template issue vs different Thai-language issue; not duplicates." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": true, + "reason": "Identical Thai text and same body shape; true duplicate issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is conservative and grounded in the packet. It correctly treats the language-specific translation requests as distinct issues, and it only groups the clearly near-identical generic placeholder issues plus the identical Thai pair as duplicates. The canonical-issue conclusion is appropriately cautious, and the soft-edge false/true verdicts do not overstate evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 19, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-36817-16", + "nodes": [ + "issue:36817", + "issue:36879", + "issue:37015", + "issue:41115", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322", + "issue:46133" + ], + "soft_pairs": [ + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:36817|issue:36879", + "issue:36879|issue:41211", + "issue:36879|issue:43909", + "issue:36879|issue:44121", + "issue:36879|issue:43837", + "issue:36879|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:36879|issue:37015", + "issue:37015|issue:42503", + "issue:41115|issue:43837", + "issue:41115|issue:44121", + "issue:41115|issue:42503", + "issue:41115|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42738|issue:45027", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43671|issue:43909", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:44121|issue:45322", + "issue:43837|issue:45027", + "issue:37015|issue:45027", + "issue:43837|issue:45177", + "issue:45322|issue:46133" + ], + "prepared_review_unit_hash": "cab8c2e546cd855918ab27f1f2a13c38f758e61ca5a26d9ce99d569fcc5bad97", + "component_budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21659, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21659, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of distinct model-request issues with similar wording, not true duplicates. The overlap is broad template similarity around \"add/support model X\" rather than the same underlying bug or change.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue: the items request different models, architectures, and modalities, so they are not duplicates of one underlying ticket.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43837 is a clear, moderately discussed support request and a reasonable representative of the cluster, but it is still not a duplicate target for the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different models (EuroBert vs DEIMv2); only generic request-template similarity." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "Different model families and modalities; not the same change." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert vs RF-DETR are unrelated model requests." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "Different models and likely different integration paths." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR and LFM2.5 Audio are unrelated requests." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "Object detection model vs sparse transformer request; not same underlying task." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "Different model requests (RF-DETR vs Qwen3-ASR)." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are different modalities and code paths." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are unrelated model additions." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "Different model families; only broad 'add model' similarity." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT vs LFM2.5 Audio are separate requests." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are distinct models with no shared concrete fix." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are different model integrations." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 vs Qwen3-ASR are different model requests." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and weight-sparse transformer are not the same underlying change." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are separate model-add tickets." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 and Voxtral TTS are different model targets." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated additions." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT and Qwen3-ASR are different model families and tasks." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are not the same code-path change." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT vs BERT-like models with RoPE is a broad thematic overlap, not a duplicate." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "RoPE support request vs Voxtral TTS support request are different changes." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Both are Qwen3 speech-related, but TTS and ASR are different concrete model integrations." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are different models and implementations." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Different models and domains; no shared duplicate bug." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "TTS support request vs sparse transformer request are unrelated." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model requests." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "ASR support and LFM2.5 Audio are different model additions." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE target different model families and tasks." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "ASR support request vs weight-sparse transformer request are unrelated." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio and SAM3-LiteText are different model requests." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Sparse transformer and EUPE are different concrete model changes." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are distinct speech-model integrations." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS are unrelated requests." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "ASR support and DeepSeek-OCR2 are different model additions." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are both vision-language related, but they are different models and changes, not the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 19, + "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": "3fb562e696d96d146870bfd902d591beefcb11bc0f24d0d1a7d033da25125203", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1479, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are not duplicates: one requests block decoding behavior, while the other asks for DeepSeek v3.2 model support. They describe different features/problems with no shared targets or code path.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No canonical issue stands out because the issues are about unrelated functionality rather than the same underlying bug or feature request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different requests: block decoding vs DeepSeek v3.2 support. No evidence they refer to the same bug, change, or code path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 9, + "component_count": 19, + "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": "5e286194a90ffcbc2b4a7f2c542f0e1916bd06993b08db088c7b714030ebb4f4", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are related to Trainer extensibility, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. They should not be merged as duplicates.", + "confidence": 0.92, + "canonical_issue_reason": "Issue #43598 is the more concrete and narrowly scoped request, centered on a specific Trainer behavior (`remove_unused_column`) and its customizability.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #43598 is the best issue to keep as the representative item because it has a clear, actionable change request tied to one Trainer code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "They concern different Trainer behaviors: column removal customizability vs FLOPS estimation for custom trainers. Shared theme is extensibility, but not the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 19, + "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": "e9206a51b1e18b30d13ed162778a8992d3e4c12ce37a4939cd43171183491bb1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related Qwen feature requests, but they are not duplicates: one asks for registering Qwen3-Omni with AutoModel/AutoModelForConditionalGeneration, while the other requests AutoModelForSequenceClassification support for Qwen3.5 (Qwen3_5Config). They target different model families and different auto-class APIs.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical duplicate here; the requests are about different Qwen variants and different auto-model classes.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #43684 is the broader, more foundational registration request and is the better representative of this small cluster\u2019s theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying requests: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration vs. Qwen3.5 sequence classification support. Different model/config targets and different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 19, + "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": "2adaf9613d744269dfa53ce4f61ddd96931983a92b51391988eb966ca4e92992", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4266, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4266, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three near-identical closed issues titled \"huggingface\" look like duplicates of the same report; the \"Fett\" issue appears unrelated. I\u2019d pick the most active \"huggingface\" issue as the canonical representative.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43859 is the best representative among the duplicate-looking \"huggingface\" issues because it shares the same title and has the most discussion activity.", + "canonical_pr_reason": null, + "best_issue_reason": "43859 is the strongest issue candidate: same apparent topic as the others and slightly more activity, making it the best cluster representative.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no evidence they describe the same bug or request." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no concrete shared problem shown." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different title; no sign this is the same underlying issue." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Same title, created back-to-back, and no distinguishing evidence suggests they are separate reports." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "Identical title and very similar metadata strongly suggest duplicate reports of the same issue." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Identical title and near-identical timing make them likely duplicates of the same report." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: 43845/43846/43859 share the same title and close creation times, while 43835 differs in title and has no evidence tying it to the others. The canonical choice of 43859 is conservative and supported by its slightly higher discussion activity. No PR-related claims were made." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 19, + "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": "b9bd213bee093e8f6842b80ef5768a39bc5aeee011b5b22cfea493178d62ee43", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster looks like unrelated low-information/spam issues rather than a duplicate set. One item is a documentation suggestion, one mentions \"Google LLC\", and the rest are generic spam placeholders.", + "confidence": 0.91, + "canonical_issue_reason": "No clear canonical issue: the items do not describe the same underlying bug or change, so there is no valid representative duplicate to merge around.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44418 is the only semantically readable title, but it still does not match the other spam-like items and is not a good duplicate representative.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are spam-labeled placeholders, but there is no evidence they report the same underlying issue or change." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Identical placeholder-style titles are not enough to treat them as duplicates; no shared bug/change is described." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "A spam placeholder and a documentation enhancement suggestion are clearly different kinds of issues." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "No sign these describe the same problem; one is a spam placeholder and the other is a distinct titled item." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Documentation enhancement and spam placeholder are unrelated; no shared underlying change is evident." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Documentation enhancement and spam placeholder are unrelated; not the same issue." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Same placeholder title alone is insufficient; there is no content tying them to the same underlying bug or request." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "No concrete overlap is shown between a spam placeholder and the \"Google LLC\" issue." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "No concrete overlap is shown between a spam placeholder and the \"Google LLC\" issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 19, + "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": "f12390b74d023fc2eb88cd190687a802b139dcb41ffd1658646cd52a2c218826", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues have the same title and very similar body size, so they are likely duplicate or near-duplicate reports. #45145 is the better canonical issue because it was created first and already has an inbound reference.", + "confidence": 0.84, + "canonical_issue_reason": "#45145 is earlier and has an inbound reference, so it is the more established duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the best single issue to keep because it appears to be the earliest and most representative report of the pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": false, + "reason": "They look likely related or duplicate, but the packet only provides soft similarity plus basic metadata; that is not enough to confidently treat them as the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The packet supports a likely duplicate/near-duplicate relationship: the titles match exactly, the bodies are similar in size, and #45145 is earlier with one inbound reference. However, the soft-edge verdict is too strong for the evidence available. The packet only provides soft_similarity plus metadata, so the conclusion should stay conservative (e.g., 'likely duplicate' or 'probably same report'), not 'same underlying issue rather than merely related reports.'" + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 19, + "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": 15, + "component_count": 19, + "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": "2b455ed2231a85a0e27d0db2c41d37847a11bb50729e400c6936a236e129b06e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1670, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1670, + "estimated_input_tokens": 418, + "estimated_eval_tokens": 1092 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are about 4-bit quantized inference, but they describe different regressions in different code paths: one is a Gemma4/audio-module uint8 vs. torch.finfo TypeError, the other is a BitsAndBytesConfig skip-modules regression that removes the default lm_head exclusion and triggers an AssertionError.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45674 is the cleaner standalone report: it names the configuration knob, the exact regression, and the resulting AssertionError, making it the better representative issue in this small cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Not the same underlying bug: 45672 concerns Gemma4/audio-module uint8 weights hitting torch.finfo during NF4 inference, while 45674 concerns llm_int8_skip_modules clearing the default lm_head exclusion and causing an AssertionError. They share only the broad 4-bit inference context." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 19, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45762-2", + "nodes": [ + "issue:45762", + "issue:45763" + ], + "soft_pairs": [ + "issue:45762|issue:45763" + ], + "prepared_review_unit_hash": "558f54fd9b8d868ea55224a139067871178efe8bb04ebe6b9a4dc069ea63d630", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1477, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated closed items with only superficial title similarity (personal names + dates). There is no evidence they describe the same bug or request, so they should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Different named subjects and different dates; no shared problem statement, target, or code context. The similarity is incidental, not duplicate evidence." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 17, + "component_count": 19, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45764-2", + "nodes": [ + "issue:45764", + "issue:45765" + ], + "soft_pairs": [ + "issue:45764|issue:45765" + ], + "prepared_review_unit_hash": "c9f858fcd367c816893b2f046e70e47f7abc68c12a559cfb4c641720cd8aa929", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two closed issues are just separate personal-name/date entries with no shared bug context, so they should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "Issue 45764 is the earliest-created item and, given the lack of any substantive body or context in either issue, is the most reasonable placeholder canonical issue.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45764 is the least arbitrary representative for the same reason: earliest created and equally sparse as the other issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "The titles refer to different people and dates, with no shared problem statement, code path, or explicit linkage; they do not appear to describe the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 18, + "component_count": 19, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45810-2", + "nodes": [ + "issue:45810", + "issue:46295" + ], + "soft_pairs": [ + "issue:45810|issue:46295" + ], + "prepared_review_unit_hash": "014f8f99a0ecbf77a1aff23c8ef51296065ea325282fed5ddeaea0682ab9172e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are not duplicates: one requests a new Qwen3_5 token-classification/value-model class, while the other asks for FP8 training support. They concern different features and code paths.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #45810 is the more concrete, implementation-specific request; it names a missing model class and a clear target use case.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #45810 is the best representative issue because it is narrowly scoped and directly tied to an actionable model-class addition.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "These are unrelated feature requests: one is for adding Qwen3_5ForTokenClassification, the other is for FP8 training support. They do not describe the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 19, + "component_count": 19, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45869-5", + "nodes": [ + "issue:45869", + "issue:45870", + "issue:45871", + "issue:45872", + "issue:45873" + ], + "soft_pairs": [ + "issue:45869|issue:45870", + "issue:45869|issue:45871", + "issue:45870|issue:45871", + "issue:45870|issue:45872", + "issue:45870|issue:45873", + "issue:45871|issue:45872", + "issue:45871|issue:45873", + "issue:45872|issue:45873" + ], + "prepared_review_unit_hash": "12b022c514266ee11c7a182d270a830930e482c70066c872d3b2e9c473dd12d2", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5724, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster appears to split into two highly similar issue pairs\u201445870/45871 and 45872/45873\u2014plus one unrelated dataset promotion issue (45869). Based on the available metadata, I would not call any pair proven exact duplicates; the evidence is suggestive but not conclusive.", + "confidence": 0.84, + "canonical_issue_reason": "No single issue cleanly represents the whole cluster because it contains at least two distinct topic groups and an unrelated dataset post. If a representative is needed, 45870 is the best anchor for the first DataForge example pair because it is the earlier of the two nearly identical reports.", + "canonical_pr_reason": null, + "best_issue_reason": "45870 is the strongest overall issue representative only as the earliest, cleanest member of the most tightly matched pair. It is not a perfect global representative because 45872/45873 cover a different DataForge topic and 45869 is unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Dataset promotion vs DataForge prediction example; different topics." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Dataset promotion vs DataForge prediction example; different topics." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": false, + "reason": "Same title and similar metadata, but the packet does not prove an exact duplicate or same underlying issue." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Both mention finance/DataForge, but one is a prediction example and the other is a scraping-service topic; distinct requests." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Both mention finance/DataForge, but one is a prediction example and the other is a scraping-service topic; distinct requests." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Prediction example vs scraping-service topic; not the same underlying issue." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Prediction example vs scraping-service topic; not the same underlying issue." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": false, + "reason": "Same title and similar metadata, but the packet does not prove an exact duplicate or same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary overstates the evidence. The packet only provides titles, body lengths, and soft-similarity scores; it does not prove any pair is an exact duplicate. In particular, calling 45870/45871 and 45872/45873 \"clear\" or \"exact\" duplicates is stronger than the available metadata supports, and every pair in the packet has deterministic_accept=false. A more conservative summary would say the cluster appears to split into two highly similar topic pairs plus an unrelated dataset post, without asserting exact duplication." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 33, + "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-39895-16", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048", + "pull_request:46267", + "pull_request:46344", + "pull_request:46367" + ], + "soft_pairs": [ + "pull_request:39895|pull_request:43424", + "pull_request:41855|pull_request:43424", + "pull_request:41855|pull_request:43823", + "pull_request:42668|pull_request:43823", + "pull_request:43251|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:43424|pull_request:44827", + "pull_request:43424|pull_request:43823", + "pull_request:43823|pull_request:45655", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:46344", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:46267", + "pull_request:44827|pull_request:46367", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:44827|pull_request:45655", + "pull_request:46048|pull_request:46367", + "pull_request:44827|pull_request:46048", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "cba00d6424d124a7562e958998ca7cc20b0987b39b7d0ac50401fffcbeed37d2", + "component_budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 22, + "serialized_chars": 43088, + "estimated_input_tokens": 10772, + "estimated_eval_tokens": 21800 + }, + "budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 22, + "serialized_chars": 43088, + "estimated_input_tokens": 10772, + "estimated_eval_tokens": 21800 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Only one true duplicate pair appears here: the two `fixed_cross_entropy` PRs both address issue #43240 by adding supported kwargs to the same loss call. The rest are distinct model additions, refactors, or unrelated fixes that merely share broad subsystem similarity.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43251 is the strongest canonical representative: it has the explicit issue target (#43240) and the fuller implementation of the same `fixed_cross_entropy` fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #43251 is the best overall representative because it directly fixes the tracked issue and cleanly captures the underlying change in one place.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Videoprism model addition vs executorch dynamic-shape export test; unrelated changes." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Mistral tokenizer method additions vs executorch export tests; different code paths and goals." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Tokenizer API additions vs adding MobileLLM model support; not the same underlying change." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor auto-loading robustness work vs new MobileLLM model files; broad shared area only." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss-function kwargs fix vs new model implementation; unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Cross-entropy kwargs fix vs Mistral4 test/model tweaks; no shared bug or code path." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch dynamic-shape test vs Mistral4 test fixes; both are tests, but not the same issue." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Executorch export test vs MobileLLM model addition; unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "Model addition vs config-resolution ordering fix; distinct changes." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "MobileLLM implementation vs Falcon refactor for standardized output capture; different models and purposes." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon model refactor vs model-creation skill/docs; not the same code-path problem." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon output-interface refactor vs Mistral4 test fixes; unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46344", + "accept": false, + "reason": "Mistral4 test fixes vs VitPose graph-break fix; different model subsystems and bugs." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test fixes vs layrnorm/layernorm rename sweep; unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 test fixes vs chat-response parsing overhaul; different feature areas." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46367", + "accept": false, + "reason": "Mistral4 test fixes vs deepseek/xpu kernel integration; different model and kernel work." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 test fixes vs model-creation skill content; completely unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "New MobileLLM model support vs repository skill/docs change; not the same artifact." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 test fixes vs config-resolution order fix; different problems." + }, + { + "left": "pull_request:46048", + "right": "pull_request:46367", + "accept": false, + "reason": "TDT loss kernel integration vs finegrained-fp8/xpu kernel integration; same general integration area, but different kernels and code paths." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 test fixes vs TDT loss kernel addition; unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs patch the same `fixed_cross_entropy` function to pass through supported `cross_entropy` kwargs for issue #43240, and could plausibly be merged into one change." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet. The only clearly mergeable duplicate pair is the two `fixed_cross_entropy` PRs (#43251 and #43254), which both target issue #43240 and patch the same loss helper in compatible ways. The other pair verdicts are conservative and consistent with the listed filenames/diffs, and no unsupported broad claims stand out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-3", + "nodes": [ + "pull_request:40546", + "pull_request:42781", + "pull_request:46331" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781", + "pull_request:42781|pull_request:46331" + ], + "prepared_review_unit_hash": "ad88523de97426c411f1255c18c2f8fadd252b7f4ff61517a621acdd53719b03", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10379, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10379, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster contains three distinct open PRs for speech-model additions: a VibeVoice implementation, a VibeVoice Realtime extension, and a Parakeet-RNNT addition. They overlap in auto-mapping/doc plumbing, but they are not the same underlying change.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #40546 is the cleanest canonical representative: it is the original VibeVoice implementation, has substantial review activity, and is more central than the later draft realtime extension.", + "best_issue_reason": null, + "best_pr_reason": "PR #40546 is the best single PR for the cluster because it is the core VibeVoice add-on; #42781 is a different realtime feature set, and #46331 is for a different model family entirely.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Both are VibeVoice-related and share doc/auto-registration files, but #42781 adds VibeVoice Realtime with extra acoustic-tokenizer and generation changes, so it is a separate feature rather than the same concrete fix." + }, + { + "left": "pull_request:42781", + "right": "pull_request:46331", + "accept": false, + "reason": "These target different model families (VibeVoice Realtime vs Parakeet-RNNT). Shared auto-mapping files are broad plumbing overlap, not evidence of 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": 3, + "component_count": 33, + "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": "12ca58dec53934a4d49a2c62113833ca2dc03c6bc3a753c2f0576a43c38991ce", + "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": "These two pull requests are unrelated: one adds TimesFM covariate forecasting support, while the other adds DINOv3 image classification support. They touch different models, code paths, and docs, so they should not be clustered as duplicates.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different underlying changes for different model families: TimesFM forecasting with covariates vs DINOv3 image classification support. No shared bug, feature, or mergeable code-path overlap beyond generic docs/tests/model plumbing." + } + ] + }, + "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": 33, + "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": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41797-3", + "nodes": [ + "pull_request:41797", + "pull_request:45186", + "pull_request:45630" + ], + "soft_pairs": [ + "pull_request:41797|pull_request:45186", + "pull_request:45186|pull_request:45630" + ], + "prepared_review_unit_hash": "45fb9d56dd0cee1594598486fd6dd977b9667ade073c9aabed567d7e5691ee23", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10435, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10435, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three open PRs add unrelated new models (DeepSeek OCR, Isaac, Kimi2-6). They overlap on onboarding boilerplate, but not on the same underlying change or code path.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: these are distinct model additions for different model families and implementation paths, not duplicates.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for deduping this cluster; the shared files are mostly registration/docs boilerplate, while the substantive model code differs across all three PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different model additions (DeepSeek OCR vs Isaac). Shared auto-mapping/docs boilerplate is not enough to indicate the same underlying change." + }, + { + "left": "pull_request:45186", + "right": "pull_request:45630", + "accept": false, + "reason": "Different model families and code paths (Isaac vs Kimi2-6). They share onboarding scaffolding, but not a concrete 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": 33, + "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": "1f2ddb817b71bde1ce4fd72d2f6dfd96701b01e18f923bf17d8218858bb6677a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate GitHub Actions maintenance PRs: one upgrades general workflow actions for Node 24 compatibility, the other bumps docker/build-push-action to v6 in docker-related workflows. They overlap only at the workflow-infrastructure level, not the same concrete change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: 42975 and 42976 touch different action families and solve different upgrade tasks, so they should not be deduplicated.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the broader workflow-actions upgrade and is the more representative umbrella change, but it is still not a duplicate of 42976.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Different upgrades and different action ecosystems: Node 24 compatibility bumps core workflow actions, while the other only updates docker/build-push-action to v6. They are not the same underlying PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 33, + "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:46411" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:43451", + "pull_request:42978|pull_request:45490", + "pull_request:42978|pull_request:46411", + "pull_request:43448|pull_request:43451", + "pull_request:43448|pull_request:45490", + "pull_request:43448|pull_request:46411", + "pull_request:43451|pull_request:45490", + "pull_request:43451|pull_request:46411", + "pull_request:45490|pull_request:46411" + ], + "prepared_review_unit_hash": "332ef72d80aab67c606a602127a60e24a1da61e5466869ae28c5c1c41baf4388", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 10, + "serialized_chars": 17822, + "estimated_input_tokens": 4456, + "estimated_eval_tokens": 9168 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 10, + "serialized_chars": 17822, + "estimated_input_tokens": 4456, + "estimated_eval_tokens": 9168 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All five PRs are separate model-addition efforts for different model families (ViT NEPA, Molmo, Molmo2, CTSM, MiniMax M2.7). They share a common scaffold of docs/auto-registration changes, but they do not look like duplicates of the same underlying bug or feature.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR stands out for deduplication because each PR adds a different model/package and touches different model-specific code paths.", + "best_issue_reason": null, + "best_pr_reason": "No PR is a good global canonical duplicate target; the PRs are unrelated model integrations rather than alternative fixes for one change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different models and code paths: ViT NEPA vs Molmo. Shared docs/auto-registration patterns are not enough to call them the same change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different model families: ViT NEPA vs Molmo2. They add separate implementations, not the same underlying feature." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "ViT NEPA and CTSM are unrelated model additions, despite both updating docs and auto mappings." + }, + { + "left": "pull_request:42978", + "right": "pull_request:46411", + "accept": false, + "reason": "ViT NEPA and MiniMax M2.7 are clearly different model integrations with no shared concrete bug/change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are related by name, but they are distinct model releases with different files and capabilities, so not the same change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM are different model additions; overlap is only in generic Transformers plumbing." + }, + { + "left": "pull_request:43448", + "right": "pull_request:46411", + "accept": false, + "reason": "Molmo and MiniMax M2.7 target different model families and code paths." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are separate model integrations, not one shared fix or feature." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46411", + "accept": false, + "reason": "Molmo2 and MiniMax M2.7 are unrelated model additions." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46411", + "accept": false, + "reason": "CTSM and MiniMax M2.7 are distinct model-specific PRs with no evidence of a common underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-4", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815", + "pull_request:46380" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:46380", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:46380", + "pull_request:44815|pull_request:46380" + ], + "prepared_review_unit_hash": "192b572cc904d08ee5acfb45771970515bf3889215407ad45f7c07a0f9333565", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 11145, + "estimated_input_tokens": 2787, + "estimated_eval_tokens": 5830 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 11145, + "estimated_input_tokens": 2787, + "estimated_eval_tokens": 5830 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These four PRs are not duplicates of one another: they target unrelated code paths (Whisper decoding, test-only edits, FP8 dequantization, and generation config validation). No soft-edge pair looks like the same underlying fix or change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46380 is the strongest representative of the cluster because it contains a concrete runtime bug fix in core generation config validation, with broader but coherent follow-up test coverage.", + "best_issue_reason": null, + "best_pr_reason": "PR #46380 is the best standalone PR here: it fixes a clear validation/type-conversion bug in an actual code path rather than only adjusting tests or touching an unrelated subsystem.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Whisper tokenizer decode logic and test-only CLIP/ViT edits are unrelated; they do not look like the same bug or change." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper decoding and FP8 dequantization are different subsystems with different failure modes and code paths." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "Both touch a Whisper file, but one fixes timestamp decoding while the other fixes generation watermarking config validation; not the same underlying bug." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "A test-only PR with placeholder asserts is not the same change as an FP8 dequantization fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "Test placeholder edits across CLIP/ViT are unrelated to generation config validation and watermarking handling." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "FP8 dequant/padding changes and generation config validation are different concrete problems; shared broad ML subsystem is not enough." + } + ] + }, + "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": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-2", + "nodes": [ + "pull_request:43488", + "pull_request:43532" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532" + ], + "prepared_review_unit_hash": "c01b27e2d94d5d4b2b724689706b57ff6c1179903409e61d8392f7bb75d6a5d7", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4562, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4562, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs are not duplicates: one is a small repo-bot/setup/version sanity check, while the other rewires CI/comment workflows and testing utilities. They only overlap on a shared Vit test file, which is not enough to indicate the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43532 is the better representative of the cluster because it contains the broader, concrete code/workflow change set; PR #43488 is a narrow bot-check/style test PR and does not share the same change intent.", + "best_issue_reason": null, + "best_pr_reason": "PR #43532 is the most substantial and generalizable artifact here, touching CI and test infrastructure across multiple files. PR #43488 is too specialized and unrelated to serve as the cluster representative.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Different goals and code paths: #43488 changes packaging/versioning for a bot/check, while #43532 updates CI/workflow and testing utility behavior. Shared Vit test file alone is not enough to merge them 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": 10, + "component_count": 33, + "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": "279d5dc56023b89d5195bc4accce091ce0af63f160fd24985be800c27bf972d9", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6083, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two separate PRs add different new segmentation pipelines with similar scaffolding, but they target different tasks/models and are not duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43613 is the broader/more representative pipeline addition in this cluster: it adds Promptable Visual Segmentation and wires in support across multiple related model docs/tests, making it the stronger canonical example of the shared pipeline-creation pattern.", + "best_issue_reason": null, + "best_pr_reason": "PR #43613 is the best representative of the cluster because it is a complete, self-contained pipeline feature addition with substantial integration work; however, it should not subsume PR #43612 since that PR adds a different pipeline.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Both PRs follow the same template, but they implement different concrete features: Promptable Concept Segmentation vs Promptable Visual Segmentation. They touch different task pages, pipeline modules, and model/test sets, so they are not the same underlying change and should not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 33, + "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": "3965e6b52e223fce6934c3136e678633f3f28406edd910dde0b4dfc2a1c10641", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs both touch `trainer.py`, but they fix different trainer extension points: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers/PEFT. They are not duplicates and should not be merged together.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the changes target different code paths and different customization hooks in `Trainer`, so they are not the same underlying fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the more self-contained behavior change and directly addresses a concrete training-step loss-scaling issue; PR #43663 is a separate refactor/extension point for column filtering.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different problems and merge targets: #43651 changes loss normalization during training, while #43663 adds a signature-column helper for data filtering. They are not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-2", + "nodes": [ + "pull_request:43656", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842" + ], + "prepared_review_unit_hash": "524ef028715d04e10b5efcfa22239f8dbb320321e204f10a1ecb11c7e2bd7e60", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3663, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs target the same import-time NameError in `src/transformers/cli/serve.py` around `TypeAdapter`/type annotations; PR 43842 is the cleaner, more targeted fix, while PR 43656 is a broader annotation-runtime cleanup that appears to address the same bug.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43842 is the best canonical PR because it directly fixes the reported `TypeAdapter` NameError with the smallest, most specific change (guarding the import under `TYPE_CHECKING` and quoting the annotation).", + "best_issue_reason": null, + "best_pr_reason": "PR 43842 is the strongest representative of the cluster since it cleanly addresses the concrete CLI runtime error without mixing in broader typing refactors.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": true, + "reason": "Same underlying code-path problem in `cli/serve.py`: runtime evaluation of annotations causing a `TypeAdapter`-related NameError. The two PRs are plausible alternative fixes for the same bug and could have been merged into one change." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays within the packet: both PRs touch `src/transformers/cli/serve.py`, both are about annotation/runtime `TypeAdapter` NameError fixes, and PR 43842 is the more targeted change. The soft-edge acceptance is conservative enough because the two diffs look compatible and plausibly mergeable into one bugfix PR." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 33, + "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": 33, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44066|pull_request:44068", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44018|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44085", + "pull_request:44018|pull_request:44086", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request:44098|pull_request:44101", + "pull_request:44059|pull_request:44722", + "pull_request:44018|pull_request:44071", + "pull_request:44017|pull_request:44722", + "pull_request:44017|pull_request:44161", + "pull_request:44017|pull_request:44073", + "pull_request:44068|pull_request:44071", + "pull_request:44071|pull_request:44086", + "pull_request:43997|pull_request:44073", + "pull_request:44017|pull_request:44071", + "pull_request:44073|pull_request:44101", + "pull_request:44059|pull_request:44071", + "pull_request:44018|pull_request:44059", + "pull_request:44073|pull_request:44161", + "pull_request:44056|pull_request:44101", + "pull_request:44059|pull_request:44068", + "pull_request:44101|pull_request:44161", + "pull_request:44017|pull_request:44018", + "pull_request:43997|pull_request:44056", + "pull_request:44071|pull_request:44076", + "pull_request:44074|pull_request:44101", + "pull_request:43997|pull_request:44018", + "pull_request:44056|pull_request:44161", + "pull_request:43997|pull_request:44161", + "pull_request:44074|pull_request:44161", + "pull_request:44018|pull_request:44161", + "pull_request:43997|pull_request:44071", + "pull_request:44010|pull_request:44071", + "pull_request:44018|pull_request:44129", + "pull_request:44010|pull_request:44073", + "pull_request:43997|pull_request:44074", + "pull_request:44071|pull_request:44129", + "pull_request:44010|pull_request:44018", + "pull_request:43997|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154" + ], + "prepared_review_unit_hash": "b6ca91714e984d79dc5a72a2ecd33ecdb505f9e987d29483288b6738db6d1771", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84894, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75088, + "estimated_input_tokens": 18772, + "estimated_eval_tokens": 37800 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a model-by-model output-tracing refactor sweep, not a single duplicate bug. The only clear duplicate groups are the GPT-Neo pair and the GPT-J/CodeGen family; the rest are parallel refactors in different subsystems.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44066 is the strongest representative: it explicitly targets the shared output-tracing refactor and covers the central GPT-J/CodeGen code path that later follow-up PRs overlap with.", + "best_issue_reason": null, + "best_pr_reason": "PR 44066 is the best single PR to represent this cluster because it is the broadest concrete implementation in the repeated GPT-J/CodeGen tracing-refactor family and has the clearest issue linkage.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model/code path; same issue theme is not enough to make these duplicates." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Both are output-tracing refactors, but they touch unrelated model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Same GPT-J output-tracing refactor family, with overlapping gptj/modeling_gptj.py changes." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "ResNet and EfficientNet are separate code paths; these are parallel refactors, not the same fix." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "Different models and different tracing code paths; shared theme alone is insufficient." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen vs EfficientNet are unrelated subsystems." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa are unrelated implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "Different model families and different concrete code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen output tracing and MPT output tracing are separate model-specific changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both are GPT-Neo output-tracing refactors in the same code path." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo is a different model/code path from GPT-J/CodeGen." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are distinct implementations." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated subsystems." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "Different model code paths; both mention output tracing but do not fix the same bug." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are different concrete implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are unrelated; shared issue target does not imply duplication." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "Different model families and no shared code path." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated concrete changes." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "Different model families and different output paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are separate model implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT2 and MGP-STR are different model families." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are unrelated implementations." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different model-specific output changes." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT2 and GPT-J/CodeGen are distinct code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J/CodeGen are unrelated subsystems." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are separate model implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are unrelated code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model families." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated fixes." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet and VisualBert are different model implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are unrelated code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT2 and GPT-Neo are different model-specific output tracing changes." + }, + { + "left": "pull_request:44101", + "right": "pull_request:44161", + "accept": false, + "reason": "Flaubert/XLM and LongT5 are separate subsystems." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44018", + "accept": false, + "reason": "SegFormer and GPT-Neo are unrelated models." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are separate model families." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are unrelated concrete code paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet and Flaubert/XLM are different model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are unrelated changes." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are different model-specific refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are unrelated code paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are separate models." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44161", + "accept": false, + "reason": "GPT-Neo and LongT5 are unrelated." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44071", + "accept": false, + "reason": "RegNet and MPT are different model families." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are unrelated implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are separate model-specific changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "FNet and LongT5 are unrelated code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "FNet and SegFormer are different model implementations." + }, + { + "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 separate fixes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen output-tracing refactor, with overlapping model file changes." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both target the same GPT-J output-tracing code path and would plausibly merge into one PR." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are unrelated model-specific changes." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "FNet and ResNet are different concrete code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "FNet and SpeechT5 are unrelated models." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa v2 and SpeechT5 are separate model implementations." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa v2 and ViTS are unrelated fixes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: the accepted soft edges stay within the same concrete GPT-Neo or GPT-J/CodeGen code paths, and the rejected edges correctly avoid conflating unrelated model implementations that only share the high-level output-tracing theme. The chosen representative PR (44066) is also defensible from the evidence because it is issue-linked and central to the GPT-J/CodeGen family." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 33, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44024", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44024|pull_request:44076", + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "4e9450537d6878aec8f5053c6bc2749a287e08f8ae519e3db67976210ab49ed1", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84894, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17127, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a set of model-specific refactors to standardized output tracing/capture_outputs across different architectures. They are related by theme, but they are not duplicates of the same underlying bug or change because each PR touches a different model code path.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44129 is the strongest representative: it has an explicit issue target, a clear refactor title, and a more substantive diff showing the standardized output-capture pattern applied to SpeechT5.", + "best_issue_reason": null, + "best_pr_reason": "PR 44129 is the best standalone PR in the set because it is the most clearly scoped and explicitly linked to an issue, with a concrete model-specific implementation change rather than only a generic decorator migration.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are standardized-output refactors, but they affect different models (FocalNet vs ImageGPT) and different forward paths; this is a shared pattern, not the same change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBert and TextNet are separate model implementations with different output-tracing code paths, so these are not the same concrete fix." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet and VisualBert are unrelated model-specific refactors; similar decorator usage does not make them duplicates." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "These PRs touch different models and different tracing logic, so they are not mergeable as one duplicate change." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are distinct architectures; the PRs share a migration theme but not the same underlying bug or code path." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 are separate model refactors; the shared output-capture pattern is too broad to treat as one 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": 15, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030" + ], + "prepared_review_unit_hash": "e22d35c90497aa6473076010e9c989e6d8a6e46eac7b99bc5152c4c477dbecc0", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39774, + "estimated_input_tokens": 9944, + "estimated_eval_tokens": 20144 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39774, + "estimated_input_tokens": 9944, + "estimated_eval_tokens": 20144 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a set of similarly named but distinct model-specific output-tracing refactors, not duplicate bugfixes or one mergeable PR.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical duplicate: each PR edits a different model path. If one representative is needed, 44025 is the broadest example because it spans two related files, but it does not subsume the others.", + "best_issue_reason": null, + "best_pr_reason": "44025 is the best anchor for the cluster because it is the most representative multi-file output-tracing refactor, but it is still model-specific rather than a duplicate of the rest.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model code paths (`timm_backbone` vs `upernet`); same refactor theme only." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Different models (`timm_backbone` vs `vision_text_dual_encoder`); not the same underlying change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Separate model implementations (`timm_backbone` vs `univnet`); too broad to treat as duplicates." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "`timm_backbone` and `falcon_mamba/mamba` are unrelated forward paths; same pattern, different change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different architectures (`timm_backbone` vs `rwkv`); not one mergeable fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "`mobilenet_v1` and `univnet` are distinct model paths; same output-tracing refactor theme only." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different files and models (`mobilenet_v1` vs `upernet`); not the same concrete bug/change." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "`mobilenet_v1` and `depth_anything` refactor different model forwards; too different to merge as duplicates." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "`mobilenet_v1` vs `vision_encoder_decoder`; same subsystem theme, different code path." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "`mobilenet_v1` and `speech_encoder_decoder` are separate model wrappers; not one underlying change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Different model families (`vision_text_dual_encoder` vs `univnet`); same refactor pattern, not duplicate." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "`vision_text_dual_encoder` and `upernet` touch different forward paths; no shared concrete fix." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "`vision_text_dual_encoder` vs `falcon_mamba/mamba`; unrelated model implementations." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "`vision_text_dual_encoder` and `depth_anything` are different model-specific refactors; not mergeable as one PR." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "`univnet` and `upernet` are different code paths; same output-tracing theme only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "Different model implementations (`univnet` vs `falcon_mamba/mamba`); not the same bug/change." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "`univnet` and `depth_anything` refactor distinct forwards; too broad to combine." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "`upernet` and `rwkv` are unrelated model paths; similar refactor, not duplicate." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "`upernet` vs `falcon_mamba/mamba`; separate model-specific changes." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "`falcon_mamba/mamba` and `codegen` target different code paths; not the same concrete fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "`timm_backbone` vs `codegen`; same refactor motif but unrelated implementations." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "`univnet` and `codegen` are unrelated forward paths; not one underlying change." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "`upernet` and `codegen` touch different model internals; not a duplicate pair." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "`vision_text_dual_encoder` vs `codegen`; similar output-tracing refactor, different model change." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Different wrappers (`depth_anything` vs `vision_encoder_decoder`); not the same code-path problem." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "`depth_anything` and `speech_encoder_decoder` are separate model-specific refactors." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "`depth_anything` vs `superpoint`; same theme, different concrete model path." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "`depth_anything` and `rwkv` are unrelated implementations; not mergeable as one PR." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "`depth_anything` vs `dpr`; different model code paths and scope." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are different wrappers; similar refactor only." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "`vision_encoder_decoder` vs `superpoint`; not the same underlying bug/change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "`vision_encoder_decoder` and `rwkv` are unrelated model paths." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "`vision_encoder_decoder` vs `dpr`; distinct model-specific changes." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "`speech_encoder_decoder` and `superpoint` are separate implementations; not a duplicate fix." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "`speech_encoder_decoder` vs `rwkv`; same refactor pattern, different concrete change." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "`speech_encoder_decoder` and `dpr` are unrelated model paths." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "`superpoint` and `rwkv` are different code paths; too dissimilar for a duplicate." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "`superpoint` vs `dpr`; distinct model-specific refactors." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "`rwkv` and `dpr` are unrelated implementations; not one mergeable 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": 16, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-2", + "nodes": [ + "pull_request:44662", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401" + ], + "prepared_review_unit_hash": "2c7fcb183c37c13bdd48e87f62b3e17b1ca8c00b65f28318eed786226d496963", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6552, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6552, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are both new model integrations, but for different models and different code paths: PenguinVL vs Voxtral TTS. The overlap is only in generic Transformer scaffolding (docs, auto registries, model init), so they are not duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: 44662 adds PenguinVL while 45401 adds Voxtral TTS, so they do not represent the same underlying change.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for duplication purposes; both are standalone model-addition PRs and neither subsumes the other.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different target models and implementations; shared files are broad registry/docs updates, not the same concrete bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-4", + "nodes": [ + "pull_request:44979", + "pull_request:45363", + "pull_request:45666", + "pull_request:46339" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363", + "pull_request:45363|pull_request:46339", + "pull_request:45666|pull_request:46339" + ], + "prepared_review_unit_hash": "b817b71c7caaaa5f0f85555ee5731d2d1e7279c2e02596b1b2b651a32ed6beb0", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 9633, + "estimated_input_tokens": 2409, + "estimated_eval_tokens": 5074 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 9633, + "estimated_input_tokens": 2409, + "estimated_eval_tokens": 5074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All items are related to the same kernel/module fusion area, but they look like successive iterations rather than duplicate PRs. PR 44979 introduces the base Module Fusion API, 45363 adds n-to-1 fusion via KernelConfig, 45666 extends that path, and 46339 further extends/simplifies the same feature set with additional loading/refactor changes.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46339 is the broadest and most recent implementation in this line, touching the core loading and hub-kernel code paths with follow-up simplification/refactoring on top of the KernelConfig fusion work.", + "best_issue_reason": null, + "best_pr_reason": "PR 46339 best represents the cluster because it is the latest, most complete kernel-fusion iteration and subsumes the earlier n-to-1 fusion direction more than the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Related feature area, but 44979 is the foundational Module Fusion API while 45363 is a separate n-to-1 KernelConfig implementation; they are not the same concrete change." + }, + { + "left": "pull_request:45363", + "right": "pull_request:46339", + "accept": false, + "reason": "46339 appears to be a later extension/simplification of the same fusion machinery, not the same single PR-level change; the shared files alone are not enough to treat them as duplicates." + }, + { + "left": "pull_request:45666", + "right": "pull_request:46339", + "accept": false, + "reason": "Both work on n-to-1 KernelConfig fusion, but 46339 is a follow-up refactor/extension with additional loading changes, so they are consecutive iterations rather than 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": 18, + "component_count": 33, + "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": "e0939d904095dbc9451ca32b12fa6e427575b43a194b432b70c79a1b6dd3805c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests fix the same AttributeError in `_patch_mistral_regex` by switching from `tokenizer.backend_tokenizer.pre_tokenizer` to `tokenizer.pre_tokenizer`; 45317 is the more complete version because it includes tests.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the better canonical PR because it contains the same core fix as 45086 plus test coverage, making it the more complete representative of this change.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the strongest overall PR candidate because it directly addresses the bug, includes validation in tests, and appears to subsume the earlier simpler patch.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same concrete bug fix in the same code path, with identical key edits in `tokenization_utils_tokenizers.py`; 45317 simply adds test coverage, so these are duplicate/overlapping implementations of 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": 19, + "component_count": 33, + "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": "b457cb0bb457a684fea8123e943e10cf7cdd77edeabbeb9968695466a06e42e8", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are related NemotronH architecture work, but they target different changes: one is a conversion-mapping/refactor PR for inheriting GraniteMoeHybrid-compatible weights, while the other is a broader split into NemotronH dense/sparse variants with docs and auto-mapping updates. They are not the same fix and should not be merged as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45569 is the better representative of the cluster because it is the broader, more complete NemotronH rework and includes the model split plus supporting documentation and registration changes.", + "best_issue_reason": null, + "best_pr_reason": "PR 45569 best captures the main final direction of the NemotronH changes; PR 45115 is a narrower conversion-mapping refactor that overlaps only partially on shared model files.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "They share NemotronH files, but the concrete changes differ: 45115 focuses on checkpoint conversion and GraniteMoeHybrid-compatible renaming, while 45569 introduces dense/sparse model variants, docs, and auto/model registration updates. Same subsystem, not the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 33, + "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": "67b4f5d135617a858ff7d516494b441bc7a55ca90a373be98f82dc878aef120d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5466, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs in the same heterogeneity feature area: one is a broad implementation of heterogeneous model support, while the other is a narrower per-layer configuration PR. They overlap in config scaffolding but do not look like duplicate fixes/changes.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45332 is the broader umbrella change: it adds heterogeneous model support across config, modeling, cache handling, and tests. It subsumes the per-layer config groundwork and is the better representative of the cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR #45332 is the most complete and central implementation in this cluster, so it is the best candidate to keep as the primary PR for the feature.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Related feature work, but not the same concrete change. #45333 is configuration-only groundwork, while #45332 extends into modeling and cache behavior; they could coexist as layered PRs rather than being 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": 21, + "component_count": 33, + "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": "c56d93992db94261afe194e1b37407cc1516061a53ec316522767714e9cbf8bf", + "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": "These are both typing-related PRs, but they address different changes: one is a broad type-checking sweep across many files, while the other adds typing helper protocols and targeted fixes around `modeling_utils`/flash-attention integration.", + "confidence": 0.72, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the better representative because it is a focused typing-infrastructure change with a clearer, narrower scope. PR 45415 is a broad repo-wide typing pass that bundles many unrelated file edits.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the most suitable single representative for the cluster because its change set is more concrete and self-contained, whereas PR 45415 is a wide typing sweep with many touched files and mixed cleanup work.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Both are about typing, but they do not appear to fix the same concrete bug or implement the same change. 45415 is a broad type-checking enablement across many modules; 45425 adds specific typing protocols and related fixes in a smaller area. They would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 22, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45513-2", + "nodes": [ + "pull_request:45513", + "pull_request:46084" + ], + "soft_pairs": [ + "pull_request:45513|pull_request:46084" + ], + "prepared_review_unit_hash": "baf48fe75836276b9d91b2731ff9080f9950162981d3a1fee2fc960d2d071a62", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4857, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4857, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs both touch cached multi-token forward paths, but they target different model families and different bugs. They are superficially similar yet not the same underlying change, so the soft edge should be rejected.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical duplicate PR: #45513 is a broad Qwen3.5/Qwen3-next/OLMo-hybrid cache fix, while #46084 is a focused Mamba2 seq_len>1 cache-path fix for a different code path.", + "best_issue_reason": null, + "best_pr_reason": "#46084 is the cleaner standalone PR: it is tightly scoped, explicitly tied to issue #46032, and fixes one concrete cache/seq_len bug in Mamba2.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45513", + "right": "pull_request:46084", + "accept": false, + "reason": "Both involve use_cache and multi-token cached forwarding, but they address different architectures (Qwen/OLMo hybrid vs Mamba2) and different implementation bugs. They do not appear to be the same fix and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 23, + "component_count": 33, + "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": "5bf18bcf0dea43cc3a836c3a5fee2b5261e2e8d474c8f499127efd0318ee9799", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs both involve expert parallelism, but they address different concrete bugs: one fixes EP + ZeRO-3 loading under accelerate launch, the other fixes EP + FSDP2 expert parameters being overwritten by rank-0 broadcast. They share a broad subsystem and one file, but not the same underlying change.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45662 is the better canonical representative for the cluster because it fixes the more explicit EP runtime corruption bug and includes the broader code-path changes across loading and MoE/tensor-parallel integration.", + "best_issue_reason": null, + "best_pr_reason": "PR #45662 is the strongest standalone candidate: it targets a concrete failure mode (experts overwritten by rank-0 broadcast) with a coherent fix across the relevant loading and distributed parameter paths.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Different bugs and different fix paths: #45548 is about EP + DeepSpeed ZeRO-3 loading via accelerate launch, while #45662 is about EP + FSDP2 rank-0 broadcast overwriting expert parameters. Shared EP context and one shared file are not enough to treat them as the same PR-worthy 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": 33, + "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": "c1b0045dc92905bd5ad14ac59dd0c3c91f7fa498441ec9f94954f5ee11da0ef5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4590, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4590, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same underlying xdist collision in captured_info debug artifacts; #45645 is the broader, more complete follow-up that also updates CI log collection, while #45639 is the smaller core fix in testing_utils.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45645 is the better canonical PR because it covers the core xdist-safe captured_info fix plus the downstream workflow/log-handling changes needed to preserve debug output in CI.", + "best_issue_reason": null, + "best_pr_reason": "PR #45645 is the best overall representative since it fully addresses the artifact-collision problem and includes the CI-side adjustments, not just the library-side change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Both PRs fix the same concrete bug: xdist workers colliding on captured_info debug output. #45645 appears to extend the same fix with CI artifact aggregation, so they are plausibly 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": 25, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45670-3", + "nodes": [ + "pull_request:45670", + "pull_request:46320", + "pull_request:46401" + ], + "soft_pairs": [ + "pull_request:45670|pull_request:46320", + "pull_request:46320|pull_request:46401" + ], + "prepared_review_unit_hash": "42790d45d71d8710b14213bb15d55f197d452de65d29dbf9f053fb8de94907fc", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 7028, + "estimated_input_tokens": 1757, + "estimated_eval_tokens": 3770 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 7028, + "estimated_input_tokens": 1757, + "estimated_eval_tokens": 3770 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs form a loose topic cluster around model/auto-model support and attention dispatch, but they are not duplicates of the same underlying change. The only potentially similar pair is the RecurrentGemma and BLIP-2-family SDPA/attn dispatch work, which targets different model code paths and would not plausibly merge into one PR. The GLM ASR auto-mapping PR is unrelated.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46401 is the best canonical representative only in the sense that it is the clearest, most self-contained implementation PR in the cluster: it adds SDPA dispatch for the BLIP-2 family with code changes and tests. However, it is not a duplicate of the RecurrentGemma or GLM ASR PRs.", + "best_issue_reason": null, + "best_pr_reason": "PR #46401 is the strongest standalone PR here because it has the clearest concrete code-path fix, spans the main implementation plus tests/docs, and is more substantial than the small auto-model mapping tweak in #45670 or the narrower RecurrentGemma dispatch change in #46320.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45670", + "right": "pull_request:46320", + "accept": false, + "reason": "Different underlying changes: auto-model registry mapping for glmasr vs RecurrentGemma attention-dispatch support. Same broad Transformers area, but not the same bug or mergeable change." + }, + { + "left": "pull_request:46320", + "right": "pull_request:46401", + "accept": false, + "reason": "Both mention attn_implementation dispatch, but they fix different model families and code paths (RecurrentGemma vs BLIP-2/InstructBLIP). They are related themes, not duplicates of one concrete change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 26, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45720-2", + "nodes": [ + "pull_request:45720", + "pull_request:46259" + ], + "soft_pairs": [ + "pull_request:45720|pull_request:46259" + ], + "prepared_review_unit_hash": "4c1941fb403da8d2182f6de834754e58e4c48a81daee9362ab79ad8b4ab63d70", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7189, + "estimated_input_tokens": 1798, + "estimated_eval_tokens": 3852 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7189, + "estimated_input_tokens": 1798, + "estimated_eval_tokens": 3852 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two unrelated linter/modeling-rule PRs: one implements TRF003 for `return_dict`/tuple-return handling in model forwards, the other adds TRF016 for preprocessing flag consumption. They share some touched model files but not the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "No single PR is a good canonical representative because these are distinct rule additions/fixes affecting different code paths and different linter behavior.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "Different changes: TRF003 modifies model forward signatures/tuple-return behavior, while TRF016 adds a rule about dead `do_*` preprocessing flags. Shared model filenames are incidental and do not indicate the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 27, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45737-2", + "nodes": [ + "pull_request:45737", + "pull_request:45738" + ], + "soft_pairs": [ + "pull_request:45737|pull_request:45738" + ], + "prepared_review_unit_hash": "cc2ade73a49ec6a8883a7a24c433988645588bcd049e8aad2bce56b1b24daa1c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same MusicgenMelody cache initialization bug in `forward()`: replacing `EncoderDecoderCache(DynamicCache, DynamicCache)` with a single `DynamicCache`. PR #45737 is the more complete duplicate because it bundles the fix with a regression test, while #45738 is the fix-only variant.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45737 is the best canonical PR because it contains the same code fix as #45738 plus a regression test, making it the most complete representation of the bug fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45737 is the strongest overall candidate: it fixes the concrete cache bug and adds a targeted regression test, whereas #45738 only changes the implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45737", + "right": "pull_request:45738", + "accept": true, + "reason": "Same underlying code-path bug in `musicgen_melody` cache setup; the diffs match on the core fix and the test-only addition in #45737 could plausibly live in the same PR as #45738." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: both PRs make the same `musicgen_melody` cache initialization change, and #45737 only adds a regression test on top of that fix. The soft-edge accept is conservative enough because the diffs are mergeable into one PR for the same bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 28, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45932-2", + "nodes": [ + "pull_request:45932", + "pull_request:46100" + ], + "soft_pairs": [ + "pull_request:45932|pull_request:46100" + ], + "prepared_review_unit_hash": "26f25861a59b30af689700ec2c1e11a92e1884bdb364ae72c64eea37f6b79dc0", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4300, + "estimated_input_tokens": 1075, + "estimated_eval_tokens": 2406 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4300, + "estimated_input_tokens": 1075, + "estimated_eval_tokens": 2406 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are not duplicates: one is a broad code/config/workflow change across multiple model files, while the other is a test-only PR adding deliberately failing assertions in ViT tests. They do not address the same underlying bug or change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45932 is the substantive cluster member, touching actual model/config/workflow code across multiple files; PR 46100 is only a test-migration placeholder with failing assertions and is not the same change.", + "best_issue_reason": null, + "best_pr_reason": "PR 45932 is the best representative because it is the real implementation change; PR 46100 is clearly test-only and not a candidate for duplicate consolidation with it.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Reject: 45932 changes model/config/workflow plumbing across several files, while 46100 only edits ViT tests with dummy failing assertions. They do not fix the same code-path problem and would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 29, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46073-3", + "nodes": [ + "pull_request:46073", + "pull_request:46154", + "pull_request:46196" + ], + "soft_pairs": [ + "pull_request:46073|pull_request:46154", + "pull_request:46073|pull_request:46196", + "pull_request:46154|pull_request:46196" + ], + "prepared_review_unit_hash": "e254aee5bacdb3f280e0af5528a85fc8262fef2b4e50552dccdef2fa7eb1417e", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8083, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8083, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Three separate model-specific PRs add SDPA dispatch support in different code paths (CTRL, T5, and wav2vec2/SeamlessM4T variants). They share a theme but do not look mergeable into one single duplicate PR.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46196 is the broadest representative of the cluster: it covers multiple model families and the same SDPA-dispatch theme, making it the best single PR to anchor the cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR 46196 is the best representative PR because it spans the widest set of affected models and most clearly captures the shared feature direction across the cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46073", + "right": "pull_request:46154", + "accept": false, + "reason": "Both add SDPA support, but for different architectures (CTRL vs T5) with different model code paths and tests; not the same underlying change." + }, + { + "left": "pull_request:46073", + "right": "pull_request:46196", + "accept": false, + "reason": "These target different model families and distinct attention implementations; they are related feature work, not duplicate fixes or one mergeable PR." + }, + { + "left": "pull_request:46154", + "right": "pull_request:46196", + "accept": false, + "reason": "T5 SDPA dispatch and wav2vec2/SeamlessM4T SDPA dispatch are separate model-specific changes with different affected files and logic." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 30, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46126-2", + "nodes": [ + "pull_request:46126", + "pull_request:46269" + ], + "soft_pairs": [ + "pull_request:46126|pull_request:46269" + ], + "prepared_review_unit_hash": "fdd46eda2eb3f503daa9baf11ace80366f82182947d87d64b848c884ab035226", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5571, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5571, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are PRs in the distributed stack, but they are not the same change. PR 46126 is a focused EP/Trainer integration patch, while PR 46269 is a broad distributed branch base that spans many subsystems.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46126 is the more focused, concrete change set and better represents a single underlying fix/feature than the broad base branch in PR 46269.", + "best_issue_reason": null, + "best_pr_reason": "PR 46126 is the best representative PR for triage because it targets a specific EP routing/Trainer integration path and is narrowly scoped.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Reject: the overlap is only superficial. 46126 implements EP/Trainer integration around tensor-parallel routing, while 46269 is a large distributed branch base touching many unrelated files; they do not appear to fix the same concrete code-path problem and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 31, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46152-2", + "nodes": [ + "pull_request:46152", + "pull_request:46207" + ], + "soft_pairs": [ + "pull_request:46152|pull_request:46207" + ], + "prepared_review_unit_hash": "30e14149d8fdf8fd0be3feddf93f0b290fe8c1dd1fd0b044d4c6d997713ff1b2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are about removing brittle source-file open() usage, but they fix different code paths in different modules: one in import_utils for resolving package traversables, the other in modeling_utils for attention-implementation detection. They are thematically related but not duplicates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46207 is the more concrete and self-contained fix: it updates the _can_set_attn_implementation path in modeling_utils to use inspect.getsource-style source inspection instead of file open(), with tests. It targets a single behavior change and is clearly scoped.", + "best_issue_reason": null, + "best_pr_reason": "PR 46207 is the best representative of the cluster because it addresses one specific bug-prone source-reading path directly. PR 46152 is broader and centered on a utility helper in import_utils, so it is less canonical for this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46152", + "right": "pull_request:46207", + "accept": false, + "reason": "Same broad theme (avoid open() on source), but they modify different modules and different concrete mechanisms. 46152 adds a traversable resolver in import_utils; 46207 changes model capability detection in modeling_utils. Not the same underlying bug or mergeable into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 32, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46337-4", + "nodes": [ + "pull_request:46337", + "pull_request:46362", + "pull_request:46366", + "pull_request:46368" + ], + "soft_pairs": [ + "pull_request:46337|pull_request:46368", + "pull_request:46337|pull_request:46362", + "pull_request:46362|pull_request:46368", + "pull_request:46362|pull_request:46366" + ], + "prepared_review_unit_hash": "538fa3eed621a1be5726697d03b63ee2c8c18f113f0df68c3690d958ab26e4d7", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 8228, + "estimated_input_tokens": 2057, + "estimated_eval_tokens": 4370 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 8228, + "estimated_input_tokens": 2057, + "estimated_eval_tokens": 4370 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster is superficially similar because all four PRs touch model test files and three of them are test skips, but they address different models and different concrete behaviors. There is no true duplicate pair here.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 46366 is the most substantive and self-contained change in the set: it adds backend detection and a clearer error path for the deepseek_v4 distributed test. Even so, it is not a duplicate of the other PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46337", + "right": "pull_request:46368", + "accept": false, + "reason": "Both are skip-style test changes, but for different architectures and different model-specific failures: cohere2_moe test sizing/offload behavior vs nemotron_h quant-cache incompatibility." + }, + { + "left": "pull_request:46337", + "right": "pull_request:46362", + "accept": false, + "reason": "cohere2_moe adjusts test configuration to fit accelerate splitting, while deepseek_v4 skips a quant-cache test; these are unrelated code paths and bugs." + }, + { + "left": "pull_request:46362", + "right": "pull_request:46368", + "accept": false, + "reason": "They share the same test name pattern, but the failures are model-specific quant-cache incompatibilities in different models, so they are not the same underlying change." + }, + { + "left": "pull_request:46362", + "right": "pull_request:46366", + "accept": false, + "reason": "One PR skips a quant-cache test; the other adds CUDA/XPU backend selection for distributed execution. Different functionality, not one mergeable fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 33, + "component_count": 33, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46376-2", + "nodes": [ + "pull_request:46376", + "pull_request:46417" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417" + ], + "prepared_review_unit_hash": "082c3754f1ae3cf547f9c1eb05f06f159c9f86af014f648bb0cc3ed501665595", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5960, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5960, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are unrelated: one adds multimodal support to continuous batching, the other adds a new HunYuan VL model integration and docs. No duplicate/merge candidate here.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical duplicate PR exists in this cluster; the PRs target different code paths and feature areas.", + "best_issue_reason": null, + "best_pr_reason": "PR 46417 is the clearer standalone representative because it is a complete model integration with tests/docs, but it is not a duplicate of 46376.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Different underlying changes: continuous batching multimodal/cache/scheduler work vs a HunYuan VL model add-on with docs, auto-mappings, and tests. No shared concrete bug or fix path, 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 + } + ] +} diff --git a/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/manifest.json b/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..25fe29b30c71d19cbe7e733ed822eff9b31ff6a9 --- /dev/null +++ b/snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260604t180024z", + "artifacts": { + "hybrid": "snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-06-04T18:11:32Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260604T180024Z", + "variant": "hybrid" +} diff --git a/snapshots/20260604T180024Z/manifest.json b/snapshots/20260604T180024Z/manifest.json index 10e24a36b5d8a938ebb8fedc2f4b31816ca77145..4b1ca09748d93e7da0a6684b81e5016e55058ea2 100644 --- a/snapshots/20260604T180024Z/manifest.json +++ b/snapshots/20260604T180024Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 2031 }, "extracted_at": "2026-06-04T18:00:24Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260604t180024z", + "runs": { + "hybrid-model-20260604t180024z": { + "analysis_id": "hybrid-model-20260604t180024z", + "artifacts": { + "hybrid": "snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260604T180024Z/analysis-runs/hybrid-model-20260604t180024z/manifest.json", + "model": null, + "published_at": "2026-06-04T18:11:32Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260604T180024Z", "watermark": {