diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 463256e9b94f2be0a873f7e44e891d1b279da243..e6b0d9164f38c6c332471cea0b352bee6a71c8c4 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": "20260605T000034Z", - "generated_at": "2026-06-05T00:09:29Z", + "snapshot_id": "20260605T060128Z", + "generated_at": "2026-06-05T06:10:45Z", "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": 0.506, - "size_similarity": 0.925, - "file_overlap": 0.286, - "area_overlap": 0.545, - "patch_similarity": 0.297, - "shared_filenames": [ - "src/transformers/integrations/hub_kernels.py", - "src/transformers/modeling_utils.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/integrations/hub_kernels.py", - "left_ranges": [ - [ - 11, - 17 - ], - [ - 22, - 35 - ], - [ - 505, - 716 - ] - ], - "right_ranges": [ - [ - 14, - 52 - ], - [ - 285, - 294 - ], - [ - 521, - 685 - ] - ] - }, - { - "filename": "src/transformers/modeling_utils.py", - "left_ranges": [ - [ - 3738, - 3747 - ] - ], - "right_ranges": [ - [ - 1346, - 1354 - ], - [ - 3785, - 3811 - ], - [ - 4245, - 4256 - ], - [ - 4650, - 4663 - ] - ] - } - ] - }, - { - "left_pr_number": 45666, - "right_pr_number": 46339, - "code_similarity": 0.399, - "size_similarity": 0.322, - "file_overlap": 0.667, - "area_overlap": 0.247, - "patch_similarity": 0.318, - "shared_filenames": [ - "src/transformers/core_model_loading.py", - "src/transformers/integrations/hub_kernels.py", - "src/transformers/modeling_utils.py", - "tests/kernels/test_kernels.py" - ], - "shared_file_areas": [ - { - "filename": "src/transformers/core_model_loading.py", - "left_ranges": [ - [ - 19, - 25 - ], - [ - 80, - 86 - ] - ], - "right_ranges": [ - [ - 19, - 25 - ], - [ - 80, - 86 - ] - ] - }, - { - "filename": "src/transformers/integrations/hub_kernels.py", - "left_ranges": [ - [ - 11, - 53 - ], - [ - 286, - 295 - ], - [ - 521, - 1077 - ] - ], - "right_ranges": [ - [ - 14, - 52 - ], - [ - 285, - 294 - ], - [ - 521, - 685 - ] - ] - }, - { - "filename": "src/transformers/modeling_utils.py", - "left_ranges": [ - [ - 1330, - 1338 - ], - [ - 3763, - 3789 - ], - [ - 4223, - 4234 - ], - [ - 4628, - 4641 - ] - ], - "right_ranges": [ - [ - 1346, - 1354 - ], - [ - 3785, - 3811 - ], - [ - 4245, - 4256 - ], - [ - 4650, - 4663 - ] - ] - }, - { - "filename": "tests/kernels/test_kernels.py", - "left_ranges": [ - [ - 19, - 54 - ], - [ - 228, - 281 - ], - [ - 294, - 317 - ], - [ - 605, - 995 - ] - ], - "right_ranges": [ - [ - 21, - 28 - ], - [ - 33, - 39 - ], - [ - 90, - 103 - ], - [ - 217, - 323 - ], - [ - 336, - 359 - ] - ] - } - ] - } - ] - }, { "cluster_id": "cluster-43240-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", @@ -1784,7 +1426,7 @@ "cluster_id": "cluster-45561-3", "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", "status": "open", - "confidence": 0.8, + "confidence": 0.55, "canonical_issue_number": 45561, "canonical_pr_number": 45645, "issue_numbers": [ @@ -1795,9 +1437,7 @@ 45645 ], "evidence_types": [ - "closing_reference", - "shared_issue_target", - "soft_similarity" + "closing_reference" ], "pr_comparisons": [ { @@ -1972,35 +1612,6 @@ } ] }, - { - "cluster_id": "cluster-44002-2", - "summary": "Cluster of 2 related pull requests linked by soft_similarity.", - "status": "open", - "confidence": 0.5, - "canonical_issue_number": null, - "canonical_pr_number": 44002, - "issue_numbers": [], - "pr_numbers": [ - 44002, - 44029 - ], - "evidence_types": [ - "soft_similarity" - ], - "pr_comparisons": [ - { - "left_pr_number": 44002, - "right_pr_number": 44029, - "code_similarity": 0.135, - "size_similarity": 0.43, - "file_overlap": 0.0, - "area_overlap": 0.0, - "patch_similarity": 0.329, - "shared_filenames": [], - "shared_file_areas": [] - } - ] - }, { "cluster_id": "cluster-44018-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", @@ -2332,15 +1943,6 @@ "target_issue_number": 43979, "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." }, - { - "cluster_id": "cluster-44002-2", - "canonical_pr_number": 44002, - "duplicate_pr_numbers": [ - 44029 - ], - "target_issue_number": null, - "reason": "PRs in cluster-44002-2 are treated as duplicates because they share soft_similarity evidence." - }, { "cluster_id": "cluster-44018-2", "canonical_pr_number": 44068, @@ -2359,16 +1961,6 @@ "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, @@ -2376,7 +1968,7 @@ 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." + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference evidence." }, { "cluster_id": "cluster-45647-3", diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index e73f5fc11bc4a04eb772b2f241defd7c1caa3be7..50895c2be33219e55b1d7d6c50d8fccb69def6db 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": "20260605T000034Z", - "generated_at": "2026-06-05T00:09:29Z", + "snapshot_id": "20260605T060128Z", + "generated_at": "2026-06-05T06:10:45Z", "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-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:24643", "issue:30277", @@ -133,334 +133,418 @@ "issue:41762|issue:43596", "issue:39290|issue:41720" ], - "prepared_review_unit_hash": "50bfea56a64ba632f40d890e975a7f65fd2cc664722718084808f796ad965790", + "prepared_review_unit_hash": "1ca6031677298406b7f90b52eb84cd1dccebc39d81e66856e238e260ccb6c383", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 64, - "serialized_chars": 46875, - "estimated_input_tokens": 11719, - "estimated_eval_tokens": 23694 + "serialized_chars": 46879, + "estimated_input_tokens": 11720, + "estimated_eval_tokens": 23696 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a mixed cluster of unrelated, model-specific runtime/docs issues. Only one clear duplicate pair stands out: the two ZeRO-3/BERT IndexError reports. The rest share broad subsystems (DeepSpeed, Qwen, docs, dtype/device issues) but not the same concrete bug.", - "confidence": 0.77, - "canonical_issue_reason": "issue:43596 is the best canonical issue for the only clear duplicate pair: it predates issue:43638 and describes the same ZeRO-3/BERT IndexError failure more cleanly and broadly.", + "summary": "Mostly a heterogeneous cluster of unrelated bug reports. The only clear duplicate pocket is the DeepSpeed ZeRO-3 / BERT index-out-of-bounds reports; most other soft links are just broad topical similarity and should not be merged.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43638 is the best canonical issue: it most clearly states the shared DeepSpeed ZeRO-3 + non-pretrained BERT failure, and it cleanly absorbs the near-duplicate issue:43596.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43596 is also the best overall issue to keep: earliest, concise, and the clearest representative of the shared concrete failure.", + "best_issue_reason": "issue:43638 is the strongest representative for the cluster\u2019s only true duplicate pair and is specific enough to anchor triage.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different failures: missing config attribute vs streaming chat/completions error." + "reason": "Different Qwen3.5 failures: missing config attribute vs streaming chat/completions API error." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Different hardware/backend problems on different models; only broad runtime similarity." + "reason": "Different models and backends: AMD GPU failure for GPT-OSS vs MPS histogram op not implemented for Int." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Image preprocessor dtype mismatch is not the same as AMD GPU execution failure." + "reason": "Both are model/runtime issues, but the model, hardware, and code path are unrelated." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Both are dtype/runtime issues, but on different models and code paths." + "reason": "Different root causes: autocast dtype mismatch in Phi-tiny-MoE vs GPT-OSS running on AMD GPUs." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Different model/backend failures; not the same concrete bug." + "reason": "Both touch MoE-ish models, but the reported failures and code paths are different." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Different Qwen/GPT-OSS inference failures with no concrete shared code path." + "reason": "Separate model/runtime failures with no clear shared bug mechanism." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch and ONNX export are unrelated." + "reason": "Unrelated failures: DeepSpeed collective mismatch vs ONNX export error." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Export-to-ONNX bug vs FP8 model failure; different issues." + "reason": "Different model families and different failure modes during export vs FP8 inference." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "Different model/runtime failures; no shared concrete bug." + "reason": "ONNX export error is not the same as Qwen3 auto device mapping cuda assert." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Table QA crash and outdated RAG docs are unrelated." + "reason": "One is a task crash, the other is stale RAG documentation; not the same issue." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are v5 docs cleanup items, but they reference different removed docs content." + "reason": "Both are docs cleanup, but they target different removed v5 references and are not the same concrete change." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export error and GenerationMixin import error are different failures." + "reason": "Different breakages: ONNX export vs import failure for GenerationMixin." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "TRL SFT embedding unpacking and training-resume/trust-remote-code are different problems." + "reason": "Different training bugs: TRL SFT unpacking vs resume-with-local-model/trust-remote-code." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Mask/index shape mismatch is not the same as Gemma4 device-map mismatch." + "reason": "Both are tensor/device errors, but the models and failure mechanisms differ." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both are shape/index errors, but the model/setup and failure path differ." + "reason": "ZeRO-3/BERT index error is not the same as the generic mask shape mismatch." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different error signatures and code paths; not the same bug." + "reason": "Same as above: related runtime style, but different bug and repro." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask shape mismatch is unrelated to video token/template mismatch." + "reason": "Different model/input pipeline failures; no shared underlying bug." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Different model features and failure mechanisms." + "reason": "Generic shape error vs GPT-OSS MoE tensor-parallel issue; not the same defect." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video apply_chat_template failure and training resume issue are unrelated." + "reason": "Video chat-template StopIteration is unrelated to resume-training / trust-remote-code." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "Different training/setup issue; no concrete overlap beyond general model loading." + "reason": "DeepSpeed/BERT zero-sized tensor error is unrelated to resume-training with a local model." }, { "left": "issue:43572", "right": "issue:46165", "accept": false, - "reason": "StableLmConfig pad token regression is unrelated to training-resume behavior." + "reason": "Config field regression vs resume-training behavior; different problem." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice preset bug and ONNX export error are unrelated." + "reason": "BarkProcessor voice preset bug is unrelated to ModernBERT ONNX export." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both mention DeepSpeed, but the failures are different and model-specific." + "reason": "Collective mismatch under Jamba/ZeRO-3 is not the same as BERT zero-sized tensor indexing." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same ZeRO-3/BERT IndexError during model init/loading; same concrete failure path, with 43596 being the cleaner canonical report." + "reason": "Same underlying DeepSpeed ZeRO-3/BERT empty-tensor index-out-of-bounds bug; wording differs but the repro is the same class of failure." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger cross_entropy routing and GPT-OSS tensor-parallel failure are different bugs." + "reason": "Both involve model execution, but the Liger Kernel cross-entropy call and GPT-OSS TP issue are different bugs." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Detokenization regression and LoRA merge/save/load inconsistency are different issues." + "reason": "Incorrect detokenization vs LoRA merge/save/load mismatch; separate issues." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen-3 NVFP4 runtime failure and LoRA save/load mismatch are unrelated." + "reason": "Qwen3 NVFP4 runtime error is unrelated to the LoRA merge/save problem." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation-checkpoint/FSDP tensor-size mismatch is a different bug from mask/index errors." + "reason": "Different distributed-training bugs: checkpointing/FSDP mismatch vs generic mask shape mismatch." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "Different DeepSpeed failures: 2-D weight error vs collective mismatch." + "reason": "DeepSpeed training weight-shape error vs collective mismatch; not the same root cause." }, { "left": "issue:45859", "right": "issue:45902", "accept": false, - "reason": "Same Qwen3.5-MoE rotary-embedding area, but CPU-offload incompatibility and meta\u2192to_empty uninitialized-memory are not clearly the same bug." + "reason": "Same component family, but CPU-offload incompatibility and meta->to_empty materialization are different bugs." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "BERT/ZeRO-3 loading issue and GPT-OSS tensor-parallel issue are unrelated." + "reason": "BERT zero-sized tensor index error is unrelated to GPT-OSS tensor-parallel behavior." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Activation checkpointing FSDP mismatch and custom compute_loss signature error are unrelated." + "reason": "Activation-checkpoint/FSDP recompute mismatch is unrelated to CustomTrainer compute_loss signature regression." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "Different models and failure mechanisms; no shared bug." + "reason": "Different model families and different distributed/runtime failures." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Different model families and distinct runtime errors." + "reason": "Different models and failure modes: dtype mismatch in Mistral-3 vs invalid method call in GraniteMoEHybrid." }, { "left": "issue:45874", "right": "issue:46294", "accept": false, - "reason": "Input-embeds reverse lookup cost and video token/template mismatch are unrelated." + "reason": "Gemma4 reverse embedding lookup overhead is unrelated to Qwen3.5 video token/template mismatch." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3 vLLM config regression and Mistral regex patch crash are different bugs." + "reason": "Gemma3/vLLM attribute regression is unrelated to the Mistral tokenizer regex crash." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Device-map mismatch and expensive reverse embedding lookup are different forward-path issues." + "reason": "Device-map forward mismatch vs expensive reverse embedding lookup are different Gemma4 issues." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Torch/version compatibility issue and Qwen NVFP4 runtime failure are unrelated." + "reason": "Torch-version incompatibility is not the same as Qwen-3 NVFP4 runtime failure." }, { "left": "issue:34928", "right": "issue:41720", "accept": false, - "reason": "FSDP activation-checkpoint mismatch and Qwen3 auto device-map assert are different bugs." + "reason": "FSDP activation-checkpointing mismatch and Qwen3 auto device map cuda assert are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Different models and different bugs: autocast dtype mismatch vs GPT-OSS tensor-parallel issue." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral grouped_mm tracing error is unrelated to the Mistral-3 image preprocessor dtype mismatch." }, { "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Qwen-3 NVFP4 runtime failure." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Mistral-3 image dtype issue is unrelated to an MPS histogram operator limitation." + }, + { + "left": "issue:44918", "right": "issue:45161", "accept": false, - "reason": "Phi dtype/autocast issue and GPT-OSS tensor-parallel failure are unrelated." + "reason": "TRL SFT embedding unpacking bug is unrelated to GPT-OSS TP failure." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Different models and different training/runtime paths; not a duplicate." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "Both are MoE-related only at a broad level; the concrete failures are different." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different model-specific runtime errors with no shared code path." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Cross-entropy/Liger Kernel issue is unrelated to TRL SFT input-embedding unpacking." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed ZeRO-3, but one is Gemma3 loading and the other is BERT index-out-of-bounds; not the same bug." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 docs cleanup, but they concern different removed references and are not the same concrete documentation change." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Mistral-3 image dtype mismatch is unrelated to Qwen3.5 video StopIteration." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Different backends and errors: autocast dtype mismatch vs MPS Int histogram limitation." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "TRL trainer embedding issue is unrelated to MPS MoE backend failure." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different models and code paths; only broad training/runtime similarity." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch and compute_loss signature mismatch are different code paths." + "reason": "Collective mismatch under Jamba/ZeRO-3 is unrelated to the trainer signature regression." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "FSDP recomputed-size mismatch and GenerationMixin import error are unrelated." + "reason": "FSDP checkpointing mismatch is unrelated to the GenerationMixin import failure." }, { "left": "issue:45237", "right": "issue:46015", "accept": false, - "reason": "GPU/backend runtime failure and outdated RAG docs are unrelated." + "reason": "AMD GPU model-runtime failure vs outdated RAG examples; not the same issue." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger kernel cross-entropy routing and tensor mask shape mismatch are different issues." + "reason": "Liger Kernel cross-entropy call is not the same as the generic mask/tensor shape mismatch." }, { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Corrupted checkpoint and missing text-encoder weights are not the same bug." + "reason": "Corrupted checkpoint vs missing text encoder weights are different model-loading problems." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "General torch/transformers incompatibility and USE_DISTRIBUTED=0 regression are different compatibility issues." + "reason": "Both mention version incompatibility, but the affected versions and failure conditions differ." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "PaddleOCR-VL/vLLM load failure and distributed-build regression are unrelated." + "reason": "PaddleOCR-VL/vLLM load failure is unrelated to USE_DISTRIBUTED=0 support regression." }, { "left": "issue:41762", "right": "issue:43596", "accept": false, - "reason": "Same broad ZeRO-3 IndexError theme, but different model family and report context; not clearly the same bug." + "reason": "Same broad ZeRO-3 area, but Gemma3 loading and BERT zero-sized tensor errors are different bugs." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3 vLLM config regression and Qwen3 auto device-map assert are unrelated." + "reason": "Gemma3/vLLM config regression is unrelated to Qwen3 auto device mapping cuda assert." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted duplicate pair, issue:43596 vs issue:43638, is well supported by the shared ZeRO-3/BERT IndexError wording, and the remaining soft-edge rejections do not overclaim sameness." + "feedback": "Grounded and conservative overall. The only accepted duplicate pair (issue:43596 / issue:43638) is well supported by the titles and cluster signals, and the rest of the soft-edge rejections stay cautious rather than overclaiming sameness. The summary accurately characterizes the cluster as mostly heterogeneous." }, "evaluator_used": true, "retried": false, @@ -479,7 +563,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36010", "issue:39290", @@ -573,20 +657,20 @@ "issue:43366|issue:45464", "issue:43531|issue:45070" ], - "prepared_review_unit_hash": "1309680ea656fd2dc8d79781e0ee3f14e32da7737c8a10662811bad205f7887b", + "prepared_review_unit_hash": "f256caf58c2a293616cbc0c532214f0e1390e12d4b2d822f8df0cfc2f1cebee5", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39796, + "serialized_chars": 39793, "estimated_input_tokens": 9949, "estimated_eval_tokens": 20154 }, @@ -595,265 +679,268 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous cluster of mostly unrelated Transformers issues. The soft edges are largely superficial similarities (same model family, same generic error class, or same subsystem) rather than the same underlying bug/change, so none should be merged as duplicates.", - "confidence": 0.79, - "canonical_issue_reason": "If one issue must be treated as the representative, issue #45812 is the broadest v5 regression here because it affects multiple model families and has the clearest cross-cutting user impact. That said, the cluster does not form a single duplicate family.", + "summary": "The tokenizer regression pair is a clear duplicate; the ZeRO-3/Bert index-error pair is related but not a safe duplicate under a conservative standard. Most other links are only broad textual similarity.", + "confidence": 0.68, + "canonical_issue_reason": "issue:45920 is the best canonical issue because it is open, broader in scope, and explicitly aggregates the v5 AutoTokenizer wrong-token-ID regression across multiple model families.", "canonical_pr_reason": null, - "best_issue_reason": "#45812 is the best overall issue representative because it is the widest-scope regression in the set and most plausibly central to the cluster theme, even though it should not absorb the others as duplicates.", + "best_issue_reason": "issue:45920 is the best global issue target for the tokenizer duplicate cluster because it is the most current umbrella report for the same concrete regression.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Different failures: float8 storage lookup vs distributed-pytorch support. Same broad v5 timing, but not the same bug." + "reason": "Different failure classes: float8 storage lookup vs USE_DISTRIBUTED=0 compatibility." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Both involve Qwen3.5/Gemma4-era multimodal behavior, but one is a PLE device-mismatch forward bug and the other is a video chat-template/token mismatch." + "reason": "Both are model-input/runtime bugs, but the symptoms and affected paths differ." }, { "left": "issue:43901", "right": "issue:44512", "accept": false, - "reason": "Both are docs-related, but they concern different API surfaces and different kinds of stale documentation." + "reason": "Both mention docs, but they concern different APIs and different missing/changed behavior." }, { "left": "issue:45237", "right": "issue:45901", "accept": false, - "reason": "Both crashy, but on different tasks/models and with different root symptoms; too generic to be the same bug." + "reason": "Different model/backend failure vs table-question-answering crash." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Different model families and failures: Voxtral processor issue versus Gemma4 PLE device mismatch." + "reason": "Different model-specific loading/processor problems." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger cross-entropy invocation and Qwen3.5 video StopIteration are unrelated beyond both being runtime errors." + "reason": "Unrelated kernel/loss-path issue vs video StopIteration error." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "DeepSeek detokenization regression is a model-specific bug; the other is generic outdated RAG examples docs work." + "reason": "Tokenizer detokenization bug vs RAG docs/example drift." }, { "left": "issue:45812", "right": "issue:45920", - "accept": false, - "reason": "Both are tokenizer regressions, but they affect different model sets and likely different code paths; not enough to unify." + "accept": true, + "reason": "Same underlying v5 tokenizer wrong-token-ID regression across models; 45920 is the broader umbrella report." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint loading and RAG docs are entirely unrelated." + "reason": "Checkpoint-loading/model implementation failure vs documentation examples issue." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Remote-code breakage and distributed build/runtime support are distinct compatibility issues." + "reason": "Remote-code regressions are not the same as PyTorch USE_DISTRIBUTED=0 support breakage." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Corrupted Ovis2 checkpoint versus NemotronH load failure are different model/loading problems." + "reason": "Corrupted checkpoint vs unsupported checkpoint-loading implementation; not the same bug." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "bitsandbytes initializer incompatibility is unrelated to USE_DISTRIBUTED=0 PyTorch support." + "reason": "bitsandbytes constructor incompatibility is unrelated to distributed build/runtime support." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime failure and RAG docs are unrelated." + "reason": "Model loading/type error vs docs/examples issue." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding_window, but one is a missing config attribute under vLLM and the other is a Qwen3-MoE sliding_window behavior issue." + "reason": "Both mention sliding_window, but they involve different models and error modes." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are mask/index errors with ZeRO-3 context, but one is a shape mismatch and the other is out-of-bounds during BertModel init; likely different causes." + "reason": "Related ZeRO-3/index-error family, but the concrete failures differ: mask/tensor shape mismatch versus out-of-bounds loading a non-pretrained Bert model; too weak for a clear duplicate." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper-large-v2 runtime failure and table-question-answering crash do not point to the same code path." + "reason": "Whisper runtime failure is unrelated to table-question-answering." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "Model loading in unit tests versus a corrupted checkpoint are different failure modes." + "reason": "Different model-loading issues with no shared failure mode." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch 2.10 incompatibility and a v5.4.0 pydantic PretrainedConfig regression are distinct versioning issues." + "reason": "Torch version incompatibility is not the same as a pydantic PretrainedConfig regression." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "Both mention ZeRO-3-style init/version problems, but one is BertModel index failure and the other is torch 2.4.1 vs transformers 5.8.0 incompatibility." + "reason": "ZeRO-3/Bert loading index error is unrelated to torch/transformers compatibility." }, { "left": "issue:43975", "right": "issue:45901", "accept": false, - "reason": "Different model/task contexts and different symptom classes; no shared bug evidence." + "reason": "Different feature areas: detokenization vs table QA crash." }, { "left": "issue:43792", "right": "issue:45070", "accept": false, - "reason": "Whisper loading/runtime issue versus pydantic PretrainedConfig regression are not the same problem." + "reason": "Whisper runtime failure is unrelated to PretrainedConfig pydantic breakage." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both involve Qwen3.5, but one is a chat crash and the other is a streaming inference API failure; not enough to merge." + "reason": "Same general model family, but different APIs and concrete failures." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "Different model/application areas and no concrete shared failure mechanism." + "reason": "Model load/type error vs table QA crash." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Shared version era only; Gemma3 sliding_window_pattern attribute error is unrelated to Gemma4 PLE device placement." + "reason": "Unrelated sliding-window config issue vs Gemma4 device-mismatch forward error." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Streaming chat/completions failure and MPS histogram int error are unrelated." + "reason": "Streaming chat API failure is unrelated to MPS histogram/int issues." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Torch dynamo grouped_mm tracing issue versus MPS int histogram error are different backends and code paths." + "reason": "Different model/backend problems; no shared code path." }, { "left": "issue:43404", "right": "issue:43866", "accept": false, - "reason": "Tied lm_head weights in Mistral3 is a model implementation bug; Ovis2 corrupted checkpoint is a load/data issue." + "reason": "Weight-tying bug is unrelated to checkpoint corruption." }, { "left": "issue:43541", "right": "issue:43828", "accept": false, - "reason": "Both are MoE-related, but one is a grouped_mm tracing failure and the other is an autocast dtype mismatch; different underlying bugs." + "reason": "Different runtime failures and model families; not mergeable as one fix." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor failure for Voxtral and template-node compilation errors are unrelated." + "reason": "Processor/model bug vs template compilation error." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "ImportError for GenerationMixin and Gemma3 sliding_window_pattern error are separate API regressions." + "reason": "ImportError for GenerationMixin is unrelated to sliding_window config errors." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/transformers version incompatibility and Gemma4 device mismatch are distinct bugs." + "reason": "Torch/transformers compatibility issue is unrelated to Gemma4 device placement." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are index errors with ZeRO-3, but one is a mask/tensor length mismatch and the other is zero-sized tensor access while loading Gemma3ForConditionalGeneration." + "reason": "Both are ZeRO-3/index errors, but the concrete symptoms and triggers differ enough to avoid merging." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and GGUF gpt-oss support are different model-compatibility requests." + "reason": "PaddleOCR-VL load failure is unrelated to GGUF gpt-oss support." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is a model config regression; pydantic PretrainedConfig breakage is a framework-level API issue." + "reason": "StableLm config field regression is unrelated to the pydantic PretrainedConfig issue." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Both are tensor/embedding shape problems, but one is a mask indexing mismatch and the other is unpacking Qwen3.5 input embeddings with TRL SFT trainer." + "reason": "Both are index/mask-related, but they hit different training paths and likely different bugs." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Int4 OOM regression and tied-weights translation-model issues are different change sets and failure modes." + "reason": "Int4 OOM regression is unrelated to tied-weights translation-model issues." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Grouped_mm tracing crash and GPT-OSS only-TP-not-working are separate MoE/tracing versus parallelism issues." + "reason": "Different MoE/tracing versus TP-only GPT-OSS issues." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch for Phi-tiny-MoE and Qwen3.5 streaming API failure are unrelated." + "reason": "Autocast dtype mismatch is unrelated to streaming chat API failure." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized TypeError and pydantic PretrainedConfig regression are related only by broad v5 breakage, not by code path." + "reason": "Unexpected _is_hf_initialized argument is a different loading-path bug than the pydantic PretrainedConfig regression." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF gpt-oss support and Qwen3.5 streaming failure are different features and bugs." + "reason": "GGUF architecture support and Qwen streaming failure are unrelated." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and pydantic model-field regression are not the same issue." + "reason": "Sliding-window model issue is unrelated to the pydantic config-field regression." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, + "evaluator_result": { + "accept": false, + "feedback": "The tokenizer duplicate call (45812 \u2194 45920) is well grounded, but the ZeRO-3 pair (41093 \u2194 43638) is overstated. The packet only shows broad textual similarity plus different concrete symptoms: one is a mask/tensor shape mismatch, the other is an out-of-bounds error when loading a non-pretrained Bert model under ZeRO-3. That is related, but not enough to call it a clear duplicate under a conservative standard. The summary also overstates this by saying these are the only clear duplicates. Please downgrade or mark the ZeRO-3 link as non-duplicate/uncertain." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -868,7 +955,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36331", "issue:38175", @@ -930,8 +1017,8 @@ "issue:39186|issue:44918", "issue:43716|issue:44918", "issue:40990|issue:44960", - "issue:45632|issue:45698", "issue:43645|issue:45399", + "issue:45632|issue:45698", "issue:43404|issue:45005", "issue:42915|issue:45399", "issue:44530|issue:45005", @@ -953,241 +1040,235 @@ "issue:43866|issue:44403", "issue:38175|issue:43054" ], - "prepared_review_unit_hash": "d63fea1f3dee2113b0e6c092a44ac5c63fd28c3373d7891f3f00f12e64d6b20f", + "prepared_review_unit_hash": "8846a2b1cf07cc9c99b4d712033ebfc09c60e5745fd087512abca6e621de1c60", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 33, - "serialized_chars": 36897, - "estimated_input_tokens": 9225, - "estimated_eval_tokens": 18706 + "serialized_chars": 36896, + "estimated_input_tokens": 9224, + "estimated_eval_tokens": 18704 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only the non-persistent-buffer corruption chain looks like a clear duplicate set. The SigLIP2, Qwen video, custom-module caching, and other same-subsystem pairs are too weakly evidenced from titles alone to merge conservatively.", - "confidence": 0.84, - "canonical_issue_reason": "issue:43950 is the clearest duplicate hub: it states the non-persistent-buffer corruption regression directly and sits between the exact/near-exact duplicate phrasing in issue:43644 and issue:44534.", + "summary": "Mostly a collection of unrelated issue threads, with a few small duplicate clusters around non-persistent buffer loading, remote-code/custom-module loading, SigLIP2 regressions, and Qwen video-input regressions.", + "confidence": 0.72, + "canonical_issue_reason": "Issue 43950 is the clearest representative of the shared non-persistent-buffer regression in v5, and it has the strongest duplicate linkage in this set (notably 44534, and likely 45902).", "canonical_pr_reason": null, - "best_issue_reason": "issue:43950 is the best global issue candidate because it most precisely captures a concrete regression with the strongest duplicate evidence and a clear underlying fix target.", + "best_issue_reason": "43950 is the best single issue to anchor the cluster because it states the regression clearly, is broad enough to cover closely related reports, and is more canonical than the more model-specific variants.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Different model families and failure modes; only a broad multimodal theme overlaps." + "reason": "Different bugs: one is an image preprocessor/model dtype mismatch, the other is a 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 incompatibility." + "reason": "Both are compatibility complaints, but the reported failures and root causes differ: remote-code breakage vs torch/transformers version incompatibility." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Different runtime failures in unrelated code paths." + "reason": "Unrelated failures: grouped_mm tracing error vs missing Float8 storage type." }, { "left": "issue:44479", "right": "issue:46294", - "accept": false, - "reason": "Both involve Qwen video inputs, but the packet does not show the same concrete bug or a duplicate chain." + "accept": true, + "reason": "Same Qwen video-input regression family; both point to a video template / token-feature mismatch affecting Qwen3.5-class models." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are index errors, but the shapes, triggers, and model contexts differ." + "reason": "Both are IndexErrors, but the symptoms and code paths differ (mask shape mismatch vs zero-size tensor during ZeRO-3 init)." }, { "left": "issue:43054", "right": "issue:43994", "accept": true, - "reason": "Both report the same SigLIP2 bad-output regression: degraded/nonsensical text-embedding behavior for google/siglip2-base-patch16-224." + "reason": "Same SigLIP2 regression area; both describe nonsensical/poor outputs from siglip2-base-patch16-224 and likely stem from the same loading/processor issue." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Different composite-config loading bug vs buffer-materialization bug." + "reason": "Different problems: multimodal config unwrapping vs uninitialized rotary-buffer memory after materialization." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "TensorBoard/TensorFlow backend issue is unrelated to version compatibility." + "reason": "No meaningful overlap beyond general environment/setup incompatibility language." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "Different model setups and different concrete failures." + "reason": "Different failures and models: FSDP 2-D weight error vs TRL embedding unpacking failure." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Different model families and unrelated dtype vs embedding-unpack failures." + "reason": "Different model families and different failure modes." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "Completely different models and symptoms; only the general training/eval context overlaps." - }, - { - "left": "issue:45632", - "right": "issue:45698", - "accept": false, - "reason": "Cache-path collision and wrong custom module after save/load are related, but not the same bug from the packet alone." + "reason": "Unrelated issues: perplexity evaluation on gpt-oss-20b vs a GLM5-specific report." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Notebook model initialization regression is unrelated to flash-attn fallback gating." + "reason": "Different issues: notebook custom-model initialization vs flash-attn fallback gating." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": true, + "reason": "Both concern trust_remote_code/custom-module resolution and cache/path collisions leading to loading the wrong code module." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both mention tied weights, but the concrete model classes and bugs differ." + "reason": "Related only at a high level ('tied weights'); the model families and concrete bug paths are different." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "FP8 config failure is unrelated to flash-attn fallback checks." + "reason": "Unrelated: FineGrainedFP8Config failure vs flash-attn fallback checks." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttention cache crash is not the same as a tied-weights loading regression." + "reason": "Different code paths: paged attention cache/group type error vs tied-weights handling." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior is a different problem from tied-weight loading." + "reason": "Quantization/offloading behavior is not the same bug as tied-weights regression." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading is a separate model-specific bug." + "reason": "Checkpoint loading failure for NemotronH is unrelated to tied-weights translation-model issues." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "Indexer discussion is unrelated to a docs wording issue." + "reason": "A DSA/indexing discussion is unrelated to a docs typo about the removed run command." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Different models and failures; no shared code-path bug is evident." + "reason": "Different models and failures: TRL embedding unpacking vs invalid method call in GraniteMoEHybrid." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "Separate documentation/wording issues." + "reason": "Unrelated documentation nits." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Different models and symptoms; no concrete shared bug is shown." + "reason": "Different models and unrelated inference failures." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Different model-loading/runtime issues." + "reason": "Different root causes and subsystems." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same non-persistent-buffer corruption regression, described with different wording." + "reason": "Same underlying regression: non-persistent buffers are being corrupted/junk-filled during v5 loading." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "Docs-example errors and SigLIP2 embedding-quality regressions are related only thematically." - }, - { - "left": "issue:43644", - "right": "issue:43950", - "accept": true, - "reason": "Exact same non-persistent-buffer junk/corruption regression." + "reason": "Both mention SigLIP2, but one is a documentation example error and the other is a model-quality regression." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Both involve numeric instability, but the trigger and failure mechanism differ." + "reason": "Different numerical issues in different training/inference paths; not the same bug." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Trainer API signature mismatch and ZeRO-3 loading error are unrelated." + "reason": "Different errors and code paths: custom trainer API mismatch vs DeepSpeed ZeRO-3 loading failure." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Different quantization/storage incompatibilities." + "reason": "Different storage/type incompatibility reports with no shared concrete bug." }, { "left": "issue:43950", "right": "issue:45902", - "accept": false, - "reason": "Both touch non-persistent buffers, but the triggering paths differ enough to keep separate." + "accept": true, + "reason": "Same non-persistent-buffer materialization/loading regression; 45902 looks like a model-specific manifestation of the broader 43950 bug." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "SigLIP2 output quality and tokenization-version changes are distinct issues." + "reason": "SigLIP2 embedding quality is unrelated to a tokenizer-version change report." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Checkpoint corruption is unrelated to loading noise/warnings." + "reason": "Checkpoint corruption is unrelated to noise/logging when loading a transformer." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Same model family, but the packet does not demonstrate the same underlying bug." + "reason": "Both are SigLIP2-related, but one is a probabilities anomaly and the other is a text-embedding quality complaint; not clearly the same underlying bug." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The summary is mostly grounded, but a few soft-edge accepts overstate the evidence. In particular, `issue:45632` vs `issue:45698` is not clearly the same bug from the packet alone (cache-path collision vs wrong custom module after save/load), and `issue:44479` vs `issue:46294` is also only title-level similarity rather than a demonstrated duplicate chain. `issue:38175` vs `issue:43054` is plausible, but still weakly supported. Because the instructions call for conservative soft-edge verdicts, I can\u2019t accept this as-is." + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The canonical choice of 43950 is well supported by the explicit duplicate-style match to 44534, and the other accepted soft pairs are phrased cautiously enough to fit the issue titles. The rejected pairs also avoid overclaiming beyond the evidence." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1203,7 +1284,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:41553", "issue:41762", @@ -1286,20 +1367,20 @@ "issue:43792|issue:44220", "issue:43257|issue:45305" ], - "prepared_review_unit_hash": "d84ed3afc9f7cb676581f522bf8a3bc3d04106e9783cc33999f079ab2edc2605", + "prepared_review_unit_hash": "dd0b94d6366d79d893300215d9388062470d33f14c0ee9dfcc2150a4ab20febd", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36795, + "serialized_chars": 36796, "estimated_input_tokens": 9199, "estimated_eval_tokens": 18654 }, @@ -1308,204 +1389,204 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are separate model-specific loading or training regressions, not duplicates. Only two pairs look like clear duplicates: the v5 pipeline-docs cleanup and the `_is_hf_initialized` constructor incompatibility.", - "confidence": 0.76, - "canonical_issue_reason": "Issue 45020 is the broadest representative of the cluster\u2019s main theme: recent Transformers v5 loading regressions affecting remote-code/model loading across multiple models.", + "summary": "This cluster is mostly a set of false-positive similarity matches across unrelated issues. Only two pairs look like true duplicates or near-duplicates: the v5 docs/pipeline-removal reports and the `_is_hf_initialized` loading regression.", + "confidence": 0.93, + "canonical_issue_reason": "No single cluster-wide canonical issue: the items span many unrelated bugs. If one duplicate target is needed, 43827 is the cleanest canonical issue because 44509 restates the same v5 docs/pipeline-removal regression.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the strongest single anchor for triage because it is cross-cutting, clearly describes a user-facing regression, and is broad enough to represent the surrounding loading-breakage reports.", + "best_issue_reason": "43827 is the best issue to keep: it is specific, actionable, and has a near-identical duplicate in 44509.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Different models and different failure modes: tokenizer error messaging vs missing tied-weights attribute on Molmo." + "reason": "Different models and different failure modes: tokenizer error messaging vs missing `all_tied_weights_keys` attribute." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Different models and code paths: Gemma4 FSDP2 shared_kv_states bug vs Qwen3_5MoE rotary embedding materialization bug." + "reason": "Both involve Gemma/Qwen3.5 MoE memory behavior, but the concrete bugs differ: FSDP2 `shared_kv_states` rebuild vs uninitialized `inv_freq` after `to_empty`." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Both are Qwen3-family issues, but one is Flash Attention position_ids handling and the other is CPU-offload compatibility in rotary embedding." + "reason": "Different code paths: Flash Attention packed-sequence misinterpretation vs CPU-offload incompatibility in rotary embedding." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Different regressions: PEFT LoRA gradient checkpointing vs int4 quantization memory growth." + "reason": "Unrelated regressions: LoRA gradient checkpointing vs int4 quantization memory reservation." }, { "left": "issue:43054", "right": "issue:43493", "accept": false, - "reason": "Related to SigLIP2, but one is a quality complaint and the other is an HF-vs-JAX implementation discrepancy; not clearly the same bug." + "reason": "Both concern SigLIP2, but one is a quality comparison and the other is an HF-vs-JAX implementation discrepancy; not the same bug." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Different model-specific loading problems with no shared code-path evidence." + "reason": "Voxtral AutoTokenizer error handling is unrelated to glm-4v-9b loading failure." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "One is a noisy-warning complaint, the other is a concrete model load failure." + "reason": "Generic loading-noise complaint vs a specific model load failure; not the same underlying issue." }, { "left": "issue:43975", "right": "issue:45520", "accept": false, - "reason": "Different subsystems: tokenizer detokenization regression vs import_utils flash-attn KeyError." + "reason": "Tokenizer detokenization bug and Python 3.13 flash_attn import issue are unrelated." }, { "left": "issue:44403", "right": "issue:44488", "accept": false, - "reason": "Same general loading area, but one is noise and the other is a model-specific load failure." + "reason": "Both mention loading problems, but they are different models/issues with no shared concrete bug." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Both involve remote-code loading, but 43643 is a specific missing-fields bug in AutoConfig while 45020 is a broader compatibility report across models." + "reason": "`AutoConfig` missing fields with `trust_remote_code` is not the same as broad remote_code loading breakage." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Different problems: LoRA/gradient-checkpointing behavior vs Qwen3 MoE training memory efficiency." + "reason": "Different issues: PEFT gradient checkpointing vs MoE training memory inefficiency." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Both are model-loading issues, but Gemma3 ZeRO-3 index error and Qwen3 MoE FP8 failure are unrelated code paths." + "reason": "Gemma3 ZeRO-3 load failure and Qwen3 MoE FP8 failure are unrelated model-loading bugs." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "Different device/runtime issues: MPS histogram op limitation vs Gemma4 PLE device mismatch." + "reason": "MPS histogram/int handling in MoE is unrelated to Gemma4 device-map forward mismatch." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Different models and symptoms; no evidence they share the same underlying bug." + "reason": "Qwen3VL weight_only loading error and GLM5 issue are different model-specific problems." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Different incompatibilities: missing Qwen2_5_VL import vs bitsandbytes `_is_hf_initialized` constructor error." + "reason": "Importing Qwen2.5-VL and bitsandbytes `_is_hf_initialized` are separate failures." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer load failure and flash-attn fallback logic are unrelated." + "reason": "Tokenizer loading failure and flash-attn2 fallback gating are unrelated." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "Different model failures: GLM-4V loading vs Molmo tied-weights attribute error." + "reason": "Different models and root causes: glm-4v-9b loading vs missing `all_tied_weights_keys` on Molmo." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote-code loading regressions and Gemma4 PLE device placement are unrelated." + "reason": "Remote_code regression and Gemma4 device-map mismatch are not the same bug." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "Both mention Qwen MoE/LoRA, but one is a version-compatibility failure and the other is a tied-embeddings warning." + "reason": "Qwen3 MoE LoRA compatibility across versions is unrelated to a tie-word-embeddings warning in Qwen3.5 fine-tuning." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "A tokenizer-mapping edge case is not the same bug as remote-code breakage." + "reason": "`add-new-model-like` token-mapping failure is a different code path than remote_code loading regressions." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Different runtime failures: autocast dtype mismatch vs device_map placement mismatch." + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Gemma4 device placement." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Tokenizer regex patch crash and flash-attn fallback checks are separate issues." + "reason": "Mistral tokenizer regex patch crash and flash-attn2 fallback blocking are unrelated." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer performance regressions, but they affect opposite conversion directions and different hot paths." + "reason": "Both are tokenizer performance regressions, but they hit different functions (`convert_ids_to_tokens` vs `convert_tokens_to_ids`) and are not the same bug." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "General remote-code breakage is not the same as the Kimi-K2.5 tokenizer codec regression." + "reason": "Broad remote_code breakage and Kimi-K2.5 tokenizer codec regression are different concrete issues." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "KeyError loading MobileLLM and flash-attn fallback gating are unrelated." + "reason": "MobileLLM loading key error is unrelated to flash-attn2 fallback gating." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "lm_head tying in Mistral3 is unrelated to silent corruption of non-persistent buffers." + "reason": "lm_head weight tying in Mistral3 and non-persistent buffer corruption are distinct loading/state bugs." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Same docs cleanup: both note pipeline tasks removed in v5 and still referenced in documentation." + "reason": "Same underlying docs regression: v5 still references removed pipeline tasks in summarization/translation docs." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Same underlying `_is_hf_initialized` incompatibility during loading/init_empty_weights, described from two angles." + "reason": "Same underlying `_is_hf_initialized` regression during model loading/init-empty-weights; both report the same TypeError from the same code path." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Different Qwen/DeepSpeed and bitsandbytes incompatibility bugs." + "reason": "Qwen3 MoE FP8 failure is unrelated to the `_is_hf_initialized` loading regression." }, { "left": "issue:43792", "right": "issue:44220", "accept": false, - "reason": "The Whisper load failure is too generic to confirm it is the same bug as the fbank-feature helper issue." + "reason": "Whisper loading/run failure and `_torch_extract_fbank_features()` issue are not clearly the same underlying bug." }, { "left": "issue:43257", "right": "issue:45305", "accept": false, - "reason": "Both involve DeepSpeed/Qwen3-ish training, but one is weight conversion and the other is gradient averaging with GAS." + "reason": "Different MoE/deepspeed training problems: weight conversion vs gradient averaging." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well-grounded in the packet and stays conservative on the soft-edge calls. The two accepted pairs are plausibly the same bugfix/doc cleanup, and the rejected pairs are mostly justified by model- or code-path-specific differences. No clear overstatement of evidence." + "feedback": "The summary is well grounded in the packet and stays conservative on the soft-edge pairs. The two accepted pairs are plausibly the same underlying regressions by title/code path, and the rest are reasonably rejected as distinct bugs. No evidence of overstatement." }, "evaluator_used": true, "retried": false, @@ -1524,7 +1605,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:24643", "issue:30064", @@ -1596,10 +1677,10 @@ "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:42915|issue:43957", "issue:43645|issue:43646", + "issue:43550|issue:44910", "issue:44560|issue:44918", "issue:42617|issue:43366", "issue:36010|issue:42915", @@ -1609,232 +1690,226 @@ "issue:38175|issue:43994", "issue:43638|issue:44661" ], - "prepared_review_unit_hash": "5d256cfc11c36bff7172663c9ddf9a61ab39c79142c828b71679e26358c3b64e", + "prepared_review_unit_hash": "b65625dd76676f389c7a50ef81f36ba7ad7145bd99d817a56753ce81a6e1598e", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37486, - "estimated_input_tokens": 9372, - "estimated_eval_tokens": 19000 + "serialized_chars": 37484, + "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 grab bag of unrelated Transformers bug reports, with a few clear duplicate-like pairs around the same model or error path. I\u2019d keep the more general custom-model init issue as the canonical issue.", - "confidence": 0.81, - "canonical_issue_reason": "issue:43646 is the broad, general statement of the custom-model initialization regression; issue:43645 looks like a notebook-specific variant of the same problem.", + "summary": "This cluster is mostly a collection of unrelated Transformers bug reports across different models and subsystems. The only clear duplicate-like pair is the two custom model initialization regressions in Transformers 5.0.0; the rest are too model- or path-specific to merge safely.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43646 is the broader report of the custom model initialization regression, while issue:43645 is the notebook-specific variant of the same underlying bug.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the cleanest representative of the duplicate pair because it describes the underlying regression without the extra notebook-specific constraint.", + "best_issue_reason": "issue:43646 is the best representative issue because it states the core regression cleanly and covers the underlying problem without narrowing to a specific notebook scenario.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Both are training-time errors, but the compute_loss kwarg mismatch and the mask/tensor shape mismatch are different failures and different code paths." + "reason": "Different failure modes: a Trainer signature mismatch versus a tensor/mask shape mismatch." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Both mention gradient checkpointing, but one is about PEFT LoRA behavior 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 an effectiveness issue and the other is a missing flag on a trust_remote_code variant." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "Both concern tokenizer loading, but they affect different models and one is an error-message issue while the other is a hard load failure." + "reason": "Different tokenizer-loading regressions affecting different models; same broad area but not the same bug." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both are ZeRO-3 weight-loading problems, but they target different Qwen model families and appear to be separate model-specific failures." + "reason": "Both are ZeRO-3 weight-loading failures, but they involve different Qwen models and different symptoms." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "One is a Qwen3.5 RLHF gradient explosion from 3D position_ids; the other is a flash-attention AttributeError on sink-less models. Different issues." + "reason": "One is an RLHF numerical stability issue; the other is a flash-attention crash on a null auxiliary state." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different models and different loading failures; too broad to treat as the same underlying bug." + "reason": "Different model-loading problems for different architectures; no clear shared root cause." }, { "left": "issue:41093", "right": "issue:44805", - "accept": true, - "reason": "These describe the same mask-vs-indexed-tensor shape mismatch IndexError, with essentially the same error signature." + "accept": false, + "reason": "The error text is similar, but there is not enough evidence that these reports share the same concrete code path." }, { "left": "issue:44479", "right": "issue:44560", - "accept": true, - "reason": "Both are Qwen3-VL family video-input regressions in the same release line, and the StopIteration looks like the same broken video path." + "accept": false, + "reason": "Both concern Qwen video regressions, but the described failures are different enough that they likely need separate fixes." }, { "left": "issue:44863", "right": "issue:46402", "accept": false, - "reason": "Different architectures and different loading failures; no strong evidence of a shared root cause." + "reason": "One is a checkpoint-compatibility problem for NemotronH; the other is missing text-encoder weight loading for Sam3Model." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both are loading-related, but one is Voxtral AutoTokenizer messaging and the other is a separate tokenizer load failure for a different model." + "reason": "Both are loading errors, but they concern different models and different error behavior." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Completely different subsystems: DeepSpeed training vs image processor segmentation maps." + "reason": "Unrelated errors: DeepSpeed training vs image processor segmentation-map handling." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Qwen3.5 PagedAttentionCache/linear_attention crash is unrelated to NemotronH checkpoint loading." + "reason": "PagedAttention cache parsing and checkpoint loading are unrelated code paths." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Both touch rotary/position handling, but one is Qwen2.5-VL video still-image scaling and the other is a meta/to_empty initialization memory bug." + "reason": "Different Qwen3.5 bugs: rope-index scaling versus uninitialized memory after meta-to-empty materialization." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Different models and different failure modes: GGUF architecture support vs autocast dtype mismatch." + "reason": "GGUF architecture support and autocast dtype mismatch are different problems." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Same model family, but one is an FSDP2 KeyError in attention and the other is a device-map forward mismatch; different code paths." + "reason": "Different Gemma4 failures: FSDP2 shared_kv_states handling versus device_map placement mismatch." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "No overlap beyond broad model-family similarity; gpt-oss GGUF support is unrelated to Qwen3-MoE sliding_window behavior." + "reason": "Support for a new GGUF architecture and a sliding_window issue in Qwen3-MoE are not the same bug." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "AudioFlamingo3 batched inference leakage and Whisper empty-transcription generation are unrelated bugs." + "reason": "Audio batching leakage and Whisper empty-transcription handling are separate defects." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Non-persistent buffer corruption during from_pretrained is unrelated to Kosmos2.5 OCR index errors." + "reason": "Non-persistent buffer corruption versus long-OCR index errors; different subsystems and symptoms." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Video-input regression and an MPS histogram-not-implemented MoE issue are unrelated." + "reason": "Qwen3-VL video input regression and MPS histogram support are unrelated." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Import/export of a specific class and meta-device loading breakage are different issues." - }, - { - "left": "issue:42915", - "right": "issue:43957", - "accept": false, - "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to meta-device model loading." + "reason": "Importing a missing class and meta-device loading breakage are different loading issues." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "GGUF support for gpt-oss and Float8 storage lookup failure are different problems." + "reason": "GGUF model architecture support is unrelated to the float8 storage import/storage error." }, { - "left": "issue:43550", - "right": "issue:44910", + "left": "issue:42915", + "right": "issue:43957", "accept": false, - "reason": "Different code paths: torch.compile/SDPA on Bamba versus flash-attention with 3D position_ids on Qwen3.5." + "reason": "Fine-grained FP8 config failure and meta-device loading regressions are distinct bugs." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "Same custom model initialization regression; one report is just notebook-specific while the other is the general failure." + "reason": "These look like the same underlying Transformers 5.0.0 custom-model-initialization regression; one report is just notebook-specific." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Torch.compile SDPA failure on Bamba and flash-attention illegal memory access on Qwen3.5 are different code-path problems." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Both involve Qwen3.5/VL embeddings, but one is a video StopIteration regression and the other is TRL SFT unpacking failure." + "reason": "Both mention Qwen3-related video/embedding flows, but one is a StopIteration video regression and the other is a TRL SFT unpacking failure." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "A script execution problem and GGUF architecture support are not the same bug." + "reason": "A script execution failure and GGUF architecture support are not the same issue." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "GenerationMixin import failure and Qwen3Moe FP8 config failure are unrelated." + "reason": "An import error for GenerationMixin and a Qwen3Moe FP8 failure are unrelated." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Tokenizer-mapping add-new-model-like failure is unrelated to Gemma4 PLE device mismatch." + "reason": "Tokenizer mapping-name handling and Gemma4 device placement are different bugs." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Both are IndexError-shaped titles, but one is a Qwen3-VL video issue and the other is a generic mask/tensor mismatch; not enough to merge." - }, - { - "left": "issue:43632", - "right": "issue:45834", - "accept": false, - "reason": "_is_hf_initialized regression and Kosmos2.5 OCR long-input indexing are unrelated." + "reason": "Different generic runtime failures with no evidence of a shared root cause." }, { "left": "issue:38175", "right": "issue:43994", - "accept": true, - "reason": "Same model and same symptom class: incorrect/zeroed outputs from siglip2-base-patch16-224, likely the same underlying regression." + "accept": false, + "reason": "Both involve SigLIP2, but one is zero probabilities and the other is nonsensical results under AutoModel/pipeline; too little evidence for a single bug." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "DeepSpeed zero3 with a non-pretrained Bert model is unrelated to tokenizer mapping support failures." + "reason": "ZeRO-3 training on a non-pretrained Bert and tokenizer-mapping-name handling are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays mostly conservative. The accepted soft-edge pairs are plausibly supported by the titles alone, and the rejected pairs are not overclaimed. The canonical custom-model init choice also looks well supported." + "feedback": "The summary is well grounded in the packet and stays conservative. The only accepted duplicate-like pair, issue:43645/issue:43646, is plausibly the same Transformers 5.0.0 custom-model initialization regression, and the other soft-edge rejections do not overclaim shared root cause." }, "evaluator_used": true, "retried": false, @@ -1853,7 +1928,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:38175", "issue:39692", @@ -1862,7 +1937,9 @@ "issue:42915", "issue:43257", "issue:43317", + "issue:43404", "issue:43425", + "issue:43454", "issue:43596", "issue:43644", "issue:43645", @@ -1875,7 +1952,6 @@ "issue:43957", "issue:43994", "issue:44075", - "issue:44155", "issue:44164", "issue:44291", "issue:44387", @@ -1908,8 +1984,8 @@ "issue:43596|issue:44661", "issue:44661|issue:44805", "issue:43317|issue:44756", - "issue:38175|issue:39692", "issue:43644|issue:44534", + "issue:38175|issue:39692", "issue:42491|issue:43720", "issue:41720|issue:42915", "issue:44960|issue:44977", @@ -1923,8 +1999,8 @@ "issue:39692|issue:43994", "issue:45488|issue:45812", "issue:43856|issue:43873", - "issue:43257|issue:43854", "issue:43854|issue:43931", + "issue:43257|issue:43854", "issue:43856|issue:44387", "issue:44960|issue:46050", "issue:45362|issue:45406", @@ -1933,229 +2009,229 @@ "issue:44164|issue:44291", "issue:44910|issue:45588", "issue:45276|issue:45335", - "issue:41720|issue:44155", - "issue:43425|issue:43957" + "issue:43425|issue:43957", + "issue:43404|issue:43454" ], - "prepared_review_unit_hash": "b8a6bf62ad043f7ef0acf1ff79816fba6d788c8e8d9db5de4b2b14fe060856a0", + "prepared_review_unit_hash": "7beb130493026c84956e6ecc2026a701760085860d5c6f139d70e3d1a5a62643", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36396, - "estimated_input_tokens": 9099, - "estimated_eval_tokens": 18454 + "serialized_chars": 36977, + "estimated_input_tokens": 9245, + "estimated_eval_tokens": 18746 }, "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-era regression reports around loading, quantization, offload, and model-specific configuration, with only a few clear duplicate pairs. The strongest duplicate signal is the non-persistent buffer corruption report, while most other pairs are only thematically similar and should stay separate.", - "confidence": 0.61, - "canonical_issue_reason": "issue:43644 is the cleanest anchor: it describes a concrete Transformers 5 regression about non-persistent buffers being filled with junk, and issue:44534 is a near-exact duplicate with the same symptom.", + "summary": "Mostly a heterogeneous collection of Transformers v5/model-loading regressions across different model families. The only clear duplicate groups are the non-persistent-buffer junk reports and the resize_token_embeddings propagation reports; the two generic BERT load failures also look like the same underlying load-regression.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43644 is the cleanest canonical anchor: it is the earlier, more descriptive version of the exact duplicate pair with issue:44534 about non-persistent buffers being filled with junk in Transformers 5.0.0.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is the most representative and actionable issue in the set, with the clearest duplicate relationship and a concrete library-wide v5 regression rather than a model-specific one.", + "best_issue_reason": "issue:43644 is the best overall issue representative among the clear overlaps because it states the bug broadly and cleanly, making it the strongest duplicate-triage anchor in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Both are config-loading breakages, but they affect different models and appear to stem from different validation paths." + "reason": "Different models and different symptoms; both mention config/loading failures, but there is no evidence they share the same concrete bug." }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "One is a DeepSpeed/BertModel init failure; the other is an add-new-model-like tokenizer mapping issue." + "reason": "One is a DeepSpeed/zero3 BertModel index error, the other is an add-new-model-like tokenizer mapping failure; unrelated code paths." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer mapping failure and mask-shape IndexError are unrelated code paths." + "reason": "Tokenizer-mapping validation vs a tensor mask shape mismatch; not the same underlying issue." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Both mention loading/offload behavior, but one is dequantized gpu+cpu offload and the other is mmap/OOM on Strix Halo." - }, - { - "left": "issue:38175", - "right": "issue:39692", - "accept": false, - "reason": "Same model family, but one is zero probabilities and the other is a docs-example mix of model/processor mismatch and quantization errors." + "reason": "Both are loading/offload problems, but one is dequantized model offload failure and the other is mmap/OOM on Strix Halo; different bugs." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "They describe the same Transformers 5 bug: non-persistent buffers getting corrupted/junk values." + "reason": "Near-verbatim duplicate about Transformers v5 filling non-persistent buffers with junk; same bug and same fix target." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "SigLIP2 zero-probabilities bug vs a documentation example error; different root causes." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "LoRA/Qwen3 MoE version incompatibility and BitNet packed-weight loading are different failures." + "reason": "LoRA/qwen3_moe compatibility issue vs BitNet packed-weight loading; unrelated failures." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Both are Qwen3-related, but one is a device-map CUDA assert and the other is a FineGrainedFP8Config failure." + "reason": "Both involve Qwen3, but one is auto device mapping/cuda assert and the other is FineGrainedFP8Config; not the same concrete problem." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Different models and different failure modes; too broad to be the same bug." + "reason": "Different model families and different flash-attention failures; not mergeable into one fix." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Qwen3Moe FP8 loading and Strix Halo mmap/OOM are unrelated." + "reason": "Qwen3Moe FP8 loading vs Strix Halo mmap/OOM; unrelated issues." }, { "left": "issue:44451", "right": "issue:44488", - "accept": false, - "reason": "Both are load failures for different models, but there is no concrete evidence they share the same underlying bug." + "accept": true, + "reason": "Both report the current/latest version failing to load BERT-family models; likely the same model-loading regression affecting multiple checkpoints." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "DeepSpeed ZeRO-3 weight loading and incorrect visual encoder save keys are different issues." + "reason": "DeepSpeed ZeRO-3 language_model weight loading vs save_pretrained visual encoder key regression; different code paths." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "Qwen3 MoE LoRA incompatibility is not the same as resize_token_embeddings propagation in Gemma4." + "reason": "LoRA/qwen3_moe compatibility bug is unrelated to resize_token_embeddings behavior." }, { "left": "issue:43708", "right": "issue:44075", "accept": false, - "reason": "Trainer checkpoint-resume math and SGD args not being used are unrelated." + "reason": "Trainer resume/max_steps calculation vs optimizer argument handling; separate trainer/optimizer issues." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer/model load complaints, but for different models and without a shared concrete failure path." + "reason": "Both are tokenizer/load regressions, but they affect different models and likely different tokenizer paths." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Non-persistent buffer corruption and Jupyter custom-model initialization are distinct regressions." + "reason": "Non-persistent buffer junk vs notebook custom-model initialization; both are v5 regressions but not the same bug." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both involve SigLIP2, but one is a docs-example/mismatch report and the other is a nonsensical-results inference report; not clearly the same bug." + "reason": "Docs example/model-processor mismatch vs incorrect inference results in a model; distinct problems." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Tokenizer behavior regressions, but they affect different tokenizer stacks and model families." + "reason": "Different tokenizer regressions: one is LlamaTokenizer pre-tokenizer override, the other is wrong token IDs for Granite models." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Both mention memory/offloading around Qwen3 MoE, but one is training memory inefficiency and the other is quantization offload behavior." + "reason": "Both mention memory/offloading, but one is Qwen3 MoE training efficiency and the other is quantization offloading behavior; not the same bug." }, { - "left": "issue:43257", - "right": "issue:43854", + "left": "issue:43854", + "right": "issue:43931", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to GLM-4.7-Flash loading in tests." + "reason": "Different model loading failures: unit-test load issue for GLM-4.7 vs shape mismatch for Qwen3-VL-30B." }, { - "left": "issue:43854", - "right": "issue:43931", + "left": "issue:43257", + "right": "issue:43854", "accept": false, - "reason": "Different model-loading failures with different symptoms and code paths." + "reason": "Accelerate+deepspeed Qwen3 MoE conversion issue vs GLM-4.7 unit-test loading; unrelated." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Both are memory-related, but one is training inefficiency and the other is int4 quantization reserved-memory/OOM regression." + "reason": "One is training-memory inefficiency, the other is int4 quantization reserved-memory/OOM regression; same broad theme but different code paths." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "GLM5 issue and quantized save/load breakage are separate problems." + "reason": "GLM5 issue vs sinq save/load regression; no shared underlying bug." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Qwen3.5 chat crash and Gemma4Processor serve crash are unrelated." + "reason": "Qwen3.5 chat crash vs Gemma4Processor missing _tokenizer in transformers serve; different failures." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "LoRA/hf4.x versus resize_token_embeddings propagation are different bugs." + "reason": "Qwen3 MoE hf4.x/hf5.x compatibility issue is unrelated to resize_token_embeddings propagation." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Qwen3 MoE version incompatibility is not the same as Qwen3-VL shape mismatch." + "reason": "Qwen3 MoE LoRA compatibility issue vs Qwen3-VL weight-shape mismatch; different bugs." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "save/from_pretrained extra_state handling and init_empty_weights argument validation are different loading issues." + "reason": "save/from_pretrained extra_state handling vs init_empty_weights unexpected argument; distinct loading/saving regressions." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both touch flash_attention, but one is a packed-sequence position_ids crash and the other is an s_aux=None AttributeError." + "reason": "Both crash in attention-related code, but one is packed sequence misinterpretation and the other is s_aux=None AttributeError; separate issues." }, { "left": "issue:45276", "right": "issue:45335", "accept": true, - "reason": "Both point to the same resize_token_embeddings propagation bug across model-specific embedding/output modules." + "reason": "Same resize_token_embeddings regression: embeddings are not propagated to the output/decoder embeddings across model variants." }, { - "left": "issue:41720", - "right": "issue:44155", + "left": "issue:43425", + "right": "issue:43957", "accept": false, - "reason": "CUDA assert during Qwen3 auto device map is unrelated to AudioFlamingo3 batched inference token/embedding leakage." + "reason": "Torch 2.10 compatibility vs meta-device loading breakage; not the same bug." }, { - "left": "issue:43425", - "right": "issue:43957", + "left": "issue:43404", + "right": "issue:43454", "accept": false, - "reason": "Torch 2.10 incompatibility and meta-device loading regressions are different problems." + "reason": "Both are weight-tying bugs, but they affect different model classes and are not clearly the same code-path problem." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative on the soft edges. The two accepted duplicate pairs are well supported by the issue titles/body themes, and the rejected pairs do not overclaim similarity. The summary is appropriately cautious overall." + "feedback": "The duplicate triage is generally grounded in the packet and the soft-edge calls are mostly conservative. The two confirmed duplicate-like groups (43644/44534 and 45276/45335) are well supported by the issue titles, and the BERT loading pair is phrased cautiously enough as a likely shared regression. No mergeability problems here." }, "evaluator_used": true, "retried": false, @@ -2174,21 +2250,21 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36010", "issue:39404", "issue:39692", + "issue:41720", "issue:42947", "issue:43335", "issue:43381", - "issue:43404", "issue:43425", - "issue:43454", "issue:43611", "issue:43645", "issue:43646", "issue:43708", + "issue:43716", "issue:43828", "issue:43854", "issue:43873", @@ -2226,7 +2302,7 @@ "issue:46315" ], "soft_pairs": [ - "issue:43404|issue:43454", + "issue:41720|issue:44155", "issue:39692|issue:44960", "issue:44402|issue:44488", "issue:42947|issue:45704", @@ -2236,8 +2312,8 @@ "issue:44291|issue:44756", "issue:45663|issue:45859", "issue:44155|issue:45823", - "issue:44292|issue:44912", "issue:44387|issue:44756", + "issue:44292|issue:44912", "issue:43335|issue:45663", "issue:43957|issue:44756", "issue:42947|issue:43381", @@ -2259,246 +2335,253 @@ "issue:39404|issue:45701", "issue:44683|issue:45823", "issue:43611|issue:43645", - "issue:43425|issue:45823" + "issue:43425|issue:45823", + "issue:43425|issue:43716" ], - "prepared_review_unit_hash": "2d2fd9a038855907456d2c3c9611352f94442864a49190ca8f782db10c7aba68", + "prepared_review_unit_hash": "db9e4ece86bb3656228be1bd28d85c6ce338ef07e4a0535471306dc39bc240b7", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37967, - "estimated_input_tokens": 9492, - "estimated_eval_tokens": 19240 + "soft_pair_count": 35, + "serialized_chars": 38213, + "estimated_input_tokens": 9554, + "estimated_eval_tokens": 19364 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers bug reports. The only clear duplicate family here is the resize_token_embeddings regression around downstream embeddings; most other pairs are only loosely related by subsystem and should stay separate.", - "confidence": 0.64, - "canonical_issue_reason": "issue:45292 is the most general, model-agnostic report of the resize_token_embeddings/output_embeddings regression, so it works best as the cluster anchor.", + "summary": "This cluster is mostly noisy and heterogeneous, spanning several unrelated Transformers 5.x regressions (model loading, quantization/offload, gradient checkpointing, multimodal configs, and tokenizer behavior). The only clear duplicate family here is the resize_token_embeddings regression affecting output/decoder embeddings.", + "confidence": 0.57, + "canonical_issue_reason": "Issue 45292 is the cleanest, most general anchor for the resize_token_embeddings bug family; the Gemma4 and t5gemma reports read like model-specific manifestations of the same missing embedding propagation.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 is the best duplicate target because its title captures the shared API-level bug without narrowing to one model family.", + "best_issue_reason": "45292 is the strongest duplicate target in the cluster because it states the shared failure mode without model-specific noise and can subsume the closely related embedding-update reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43404", - "right": "issue:43454", + "left": "issue:41720", + "right": "issue:44155", "accept": false, - "reason": "Both mention missing weight tying/generation quality, but they are different model-specific failures and not the same concrete bug." + "reason": "Different failures: Qwen3 device-map CUDA assert vs AudioFlamingo3 batched inference token/embedding leakage." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "Unrelated topics: SigLIP2 documentation example errors vs a generic GLM5 issue." + "reason": "Doc example errors for SigLIP2 are unrelated to the GLM5 loading issue." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Both are model-loading complaints, but the underlying failures appear different (vocab mismatch vs a specific model load regression)." + "reason": "Both are model-loading complaints, but they affect different models and likely different root causes." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Different bugs: LoRA gradient checkpointing behavior vs a T5 RMSNorm memory leak." + "reason": "Gradient checkpointing/LoRA behavior is unrelated to the T5 apex RMSNorm memory leak." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for GenerationMixin is unrelated to Whisper pipeline return_language behavior." + "reason": "Import path breakage for GenerationMixin is not the same as Whisper pipeline return_language regression." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both involve quantization, but one is offloading behavior and the other is CUDA memory/OOM regression." + "reason": "Both involve quantization/memory, but one is offloading behavior and the other is increased reserved memory/OOM under int4." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Different code paths: Gemma4 FSDP attention state vs Qwen multimodal config unwrapping." + "reason": "Gemma4 FSDP2 shared_kv_states rebuilding is unrelated to Qwen multimodal text_config unwrapping." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized is unrelated to mmap-related OOM on Strix Halo." + "reason": "Unexpected init_empty_weights argument handling and Strix Halo mmap OOM are different code paths." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Different model families and failures: Gemma4 FSDP attention vs Qwen3_5Moe CPU offload compatibility." + "reason": "Gemma4 attention/FSDP state handling is unrelated to Qwen3_5Moe rotary embedding CPU offload." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "AudioFlamingo3 batch leakage and Gemma4 device mismatch are unrelated bugs." + "reason": "AudioFlamingo3 track leakage is unrelated to Gemma4 PLE device-mapping mismatch." }, { - "left": "issue:44292", - "right": "issue:44912", + "left": "issue:44387", + "right": "issue:44756", "accept": false, - "reason": "Both are quantization load issues, but the concrete models and failure modes differ." + "reason": "int4 quantization OOM and Strix Halo mmap OOM are both memory issues but not the same bug." }, { - "left": "issue:44387", - "right": "issue:44756", + "left": "issue:44292", + "right": "issue:44912", "accept": false, - "reason": "OOM under int4 quantization is not the same as disabling mmap to avoid OOM on Strix Halo." + "reason": "Different quantization/model-loading failures for different models and backends." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "Sparse-layer construction in SwitchTransformers is unrelated to Gemma4 FSDP attention state rebuilding." + "reason": "SwitchTransformers sparse-layer config bug is unrelated to Gemma4 FSDP attention state handling." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "meta-device model loading breakage is a different issue from Strix Halo mmap OOM." + "reason": "Meta-device model loading breakage is a different problem from disabling mmap to avoid OOM." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but one is ineffective under LoRA and the other is an eval-mode restriction." + "reason": "Both mention gradient checkpointing, but one is ineffective under PEFT LoRA and the other is blocked in eval mode." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same underlying regression: resize_token_embeddings does not propagate to downstream output embeddings; 45276 is the Gemma4-specific instance." + "reason": "Same resize_token_embeddings regression; 45276 is a Gemma4-specific manifestation and 45292 states the general output_embeddings failure." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "save/from_pretrained extra_state handling is unrelated to mmap/OOM behavior." + "reason": "save/from_pretrained extra_state handling is unrelated to mmap-related OOM on Strix Halo." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Gradient checkpointing in eval mode is unrelated to RLHF NaN explosions from 3D position_ids and SDPA fallback." + "reason": "Gradient checkpointing in eval mode and RLHF NaN from 3D position_ids are unrelated." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Different save/load regressions affecting different models and checkpoints." + "reason": "Unit-test model loading for GLM-4.7 is unrelated to Qwen3.5 save_pretrained regression." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video input regression in Qwen variants is unrelated to granite_speech config loading." + "reason": "Video-input regression for Qwen VL models is unrelated to strict config blocking granite_speech loading." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both involve save_pretrained/custom modules, but one is permissions propagation and the other is loading the wrong module." + "reason": "Both concern custom-model save/load flows, but one is a permissions side effect and the other is wrong-module loading." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Flash-attention sink-less-model crash is unrelated to Gemma4 FSDP shared_kv_states." + "reason": "Flash-attention sink handling crash is unrelated to Gemma4 FSDP shared_kv_states rebuilding." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization failure and checkpoint-resume max_steps miscalculation are different bugs." + "reason": "Custom model initialization breakage and Trainer resume_from_checkpoint max_steps miscalculation are different bugs." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook custom-model initialization and extra_state save/load handling are unrelated." + "reason": "Jupyter custom-model initialization and extra_state save/load handling do not describe the same issue." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Same API regression class: resize_token_embeddings fails to update the relevant downstream embeddings, with different model-specific manifestations." + "reason": "Same underlying resize_token_embeddings propagation bug; both report embeddings not updating after resize, with different affected model classes." }, { "left": "issue:45812", "right": "issue:46315", "accept": false, - "reason": "Token-ID correctness regression and tokenization performance regression are different bugs." + "reason": "Tokenizer wrong IDs regression and convert_tokens_to_ids slowdown are separate tokenizer issues." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to meta-device model loading breakage." + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is not the same as meta-device loading failures." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch 2.10 compatibility and autocast dtype mismatch are different failure classes." + "reason": "Torch version incompatibility and autocast dtype mismatch are different failures." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Tokenizer vocab-size mismatch and a specific model loading failure are not the same bug." + "reason": "Different model-loading failures affecting different checkpoints." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Missing Float8 storage type and mmap-related OOM are unrelated." + "reason": "Float8 storage lookup failure is unrelated to disabling mmap on Strix Halo." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to tokenizer-version changes." + "reason": "Whisper return_language pipeline regression is unrelated to tokenizer-version behavior changes." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device mismatch are unrelated." + "reason": "Compiled flex_attention on torch >= 2.9 and Gemma4 device mismatch are separate failures." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both concern Transformers 5 custom-model behavior, but they describe different regressions and code paths." + "reason": "Both are Transformers 5.0 custom-model loading regressions, but the reported breakage and code paths are not the same." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "Torch version incompatibility is unrelated to Gemma4 PLE device placement." + "reason": "Torch 2.10 compatibility and Gemma4 device-map forward mismatch are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch incompatibility and Mistral-3 dtype mismatch are different issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet: the cluster is mostly heterogeneous, and the only clearly supported duplicate family is the resize_token_embeddings/output_embeddings regression centered on issue:45292 with the Gemma4 and t5gemma variants. The rejected soft-edge pairs are phrased conservatively and do not overclaim sameness; the accepted pairs within the resize_token_embeddings family look like legitimate single-bug variants." + "feedback": "The assessment is grounded in the packet and stays conservative overall. The only clearly supported duplicate family is the resize_token_embeddings regression around 45292/45276/45335, and the remaining soft-edge rejections are appropriately cautious given the titles available." }, "evaluator_used": true, "retried": false, @@ -2517,7 +2600,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:32090", "issue:36032", @@ -2569,17 +2652,16 @@ "issue:46402" ], "soft_pairs": [ - "issue:43425|issue:43716", "issue:44530|issue:45084", "issue:43906|issue:44530", "issue:39404|issue:44912", "issue:44265|issue:44898", "issue:44479|issue:44805", - "issue:43906|issue:46326", "issue:44589|issue:45084", - "issue:46050|issue:46326", - "issue:43901|issue:44509", + "issue:43906|issue:46326", "issue:43425|issue:44918", + "issue:43901|issue:44509", + "issue:46050|issue:46326", "issue:43299|issue:43931", "issue:43994|issue:44079", "issue:36032|issue:43452", @@ -2591,244 +2673,250 @@ "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:43716|issue:45084", "issue:46285|issue:46402", "issue:40990|issue:43421", "issue:36032|issue:42915", "issue:43643|issue:44530" ], - "prepared_review_unit_hash": "358b24bdc4efbbfae0ea3c8f533959461009987a0cf97b7b3b5eb7edf929b93e", + "prepared_review_unit_hash": "5a0321d764168ee413e0582e4cd6a15f958791cd7dfdd564ca5f122665207511", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37710, - "estimated_input_tokens": 9428, - "estimated_eval_tokens": 19112 + "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": "These items span many unrelated Transformers regressions; none of the soft pairs look like the same underlying bug/change, so all soft edges are rejected. The most general issue in the set is the v5 model-loading regression around `base_model_prefix`, but there is no strong single duplicate cluster overall.", - "confidence": 0.81, - "canonical_issue_reason": "Issue 43611 is the broadest, most user-facing loading regression in the set and is the closest thing to a canonical issue, but the overall cluster is heterogeneous.", + "summary": "This cluster is mostly a grab-bag of distinct Transformers regressions and model-specific loading/runtime bugs, many around v5 changes, tokenizers, or quantization. None of the soft pairs look like the same underlying bug/change closely enough to merge.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43611 is the broadest and most representative report here: a v5 model-loading regression tied to `base_model_prefix`, which matches the cluster\u2019s dominant theme of loading breakages.", "canonical_pr_reason": null, - "best_issue_reason": "43611 best represents the recurring v5 model-loading breakages because it is broad, concrete, and likely to subsume adjacent loading regressions better than the more model-specific reports.", + "best_issue_reason": "issue:43611 is the best single issue to anchor this cluster because it is general, high-impact, and clearly about a core regression rather than a model-specific edge case.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43425", - "right": "issue:43716", - "accept": false, - "reason": "Torch version incompatibility vs. Qwen3-Mistral image preprocessor/model dtype mismatch; different failure modes and code paths." - }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "PagedAttentionCache runtime crash on Qwen3.5 vs tokenizer template compilation error; unrelated subsystems." + "reason": "Both are errors, but one is Qwen3.5 PagedAttentionCache/linear_attention and the other is a template compilation TypeError; different code paths." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "Isolated reproduction of another issue vs Qwen3.5 cache crash; no shared bug surface." + "reason": "A reproduction of issue 38071 is unrelated to Qwen3.5 cache crashes." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline `return_language` regression vs MXFP4 quantization fallback; different model/features." + "reason": "Whisper `return_language` pipeline regression is unrelated to gpt-oss MXFP4 quantization fallback." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "`torch.export` failure around `torch_compilable_check` vs Perceiver interpolation/image-size issue; distinct problems." + "reason": "`torch.export` with `torch_compilable_check` is a different path than Perceiver image interpolation failures." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression in VL models vs mask-shape index error; not the same underlying defect." - }, - { - "left": "issue:43906", - "right": "issue:46326", - "accept": false, - "reason": "Reproduction of issue 38071 vs mistral_common tokenizer `continue_final_message` bug; different tokenizer behavior." + "reason": "Both are regressions, but one is video input on Qwen VL variants and the other is a mask shape `IndexError`; no shared bug." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage lookup error vs template compilation error; unrelated runtime failures." + "reason": "Float8 storage lookup failure is unrelated to non-template node compilation errors." }, { - "left": "issue:46050", + "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "Quantization save/load corruption vs tokenizer final-message handling; different code paths." + "reason": "A reproduction of issue 38071 is not the same as the mistral_common `continue_final_message` bug." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch version incompatibility is unrelated to Qwen3.5 embedding unpacking with TRL SFT trainer." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Docs mismatch for `return_all_scores` vs docs mentioning removed pipeline tasks; documentation-only but not the same issue." + "reason": "Both touch docs, but one is about `return_all_scores` in text classification and the other about removed pipeline tasks; different fixes." }, { - "left": "issue:43425", - "right": "issue:44918", + "left": "issue:46050", + "right": "issue:46326", "accept": false, - "reason": "Torch incompatibility vs TRL SFT unpacking embeddings; different components and failure signatures." + "reason": "Quantized model save/load corruption is unrelated to tokenizer `continue_final_message` behavior." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Qwen3VL MoE loading regression vs Qwen3-VL weight shape mismatch; related model family but different concrete bug." + "reason": "Both mention Qwen3-VL, but one is a v5 loading regression and the other is a specific weight-shape mismatch; not the same bug." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "SigLIP2 nonsensical outputs vs `ModelOutput` key assignment bug; one is inference quality, the other data-structure behavior." + "reason": "Nonsensical SigLIP2 outputs are not the same as `ModelOutput` key assignment when a previous key was `None`." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer special-token method conflict vs gguf file loading breakage; different tokenizer/load paths." + "reason": "T5Tokenizer special-tokens conflict is unrelated to gguf loading failures for AutoTokenizer/AutoModel." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "`extra_state` save/load handling vs composite Qwen config `text_config` unwrapping; both loading-related but not the same defect." + "reason": "Extra state handling in save/from_pretrained is a different problem than unwrapping `text_config` in composite Qwen configs." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT masked-LM prediction regression vs tokenizer load failure for a specific model; unrelated model-specific issues." + "reason": "CamemBERT masked-LM prediction regression and tokenizer load failure for EMBEDDIA/est-roberta are different failures." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "DeepSpeed ZeRO3/BertModel indexing error vs compiled flex_attention on newer torch; different stack and root cause." + "reason": "DeepSpeed ZeRO-3 init `IndexError` is unrelated to compiled flex_attention incompatibility." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Tokenizer load failure for EMBEDDIA/est-roberta vs Mistral regex patch crash; same broad area, different bugs." + "reason": "Both are tokenizer load issues, but one is a generic tokenizer failure and the other is a Mistral regex patch crash; different root causes." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Qwen3 MoE training memory inefficiency vs mmap OOM on Strix Halo; not the same underlying issue." + "reason": "Training memory inefficiency on Qwen3 MoE is not the same as disabling mmap on Strix Halo to avoid OOM." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "Save/from_pretrained extra_state handling vs VL video-input regression; no shared code path." + "reason": "Save/load `extra_state` handling is unrelated to video input regressions in Qwen VL models." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline language return regression vs CamemBERT masked LM prediction issue; unrelated models and symptoms." + "reason": "Whisper pipeline language return issue and CamemBERT MLM prediction regression are unrelated." }, { "left": "issue:44737", "right": "issue:45859", "accept": false, - "reason": "XLNet CPU arange/device bug vs Qwen3_5Moe CPU offload incompatibility; both device-related but different code paths." + "reason": "Both mention device/offload, but XLNet CPU arange placement and Qwen3_5MoE rotary embedding offload are separate code paths." }, { - "left": "issue:43582", - "right": "issue:43632", + "left": "issue:44488", + "right": "issue:45081", "accept": false, - "reason": "Apple Silicon allocator warmup TypeError vs `_is_hf_initialized` flag regression; different internal failures." + "reason": "Different tokenizer loading failures for different models; no evidence they share the same underlying patch." }, { - "left": "issue:44488", - "right": "issue:45081", + "left": "issue:43582", + "right": "issue:43632", "accept": false, - "reason": "Loading `cjvt/sleng-bert` failure vs Mistral regex patch crash; distinct tokenizer-loading problems." + "reason": "Apple Silicon warmup TypeError is unrelated to the `_is_hf_initialized` v5 flag regression." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "`add-new-model-like` generator logic vs compiled flex_attention; unrelated feature/tooling areas." + "reason": "Tokenizer mapping/model-addition logic is unrelated to flex_attention compilation failures." }, { "left": "issue:44488", "right": "issue:44987", "accept": false, - "reason": "Model load failure for `cjvt/sleng-bert` vs `physical-intelligence/fast` load regression; different affected models and likely causes." + "reason": "Both are load failures, but for different models and likely different causes; not mergeable as one bug." }, { - "left": "issue:32090", - "right": "issue:39404", + "left": "issue:43611", + "right": "issue:43644", "accept": false, - "reason": "Trainer `_gpu_broadcast_one` NoneType TypeError vs Whisper `return_language` pipeline regression; unrelated behavior." + "reason": "Both are v5 regressions, but one is model loading with `base_model_prefix` and the other is non-persistent buffer initialization; distinct bugs." }, { - "left": "issue:43611", - "right": "issue:43644", + "left": "issue:32090", + "right": "issue:39404", "accept": false, - "reason": "`base_model_prefix` loading regression vs non-persistent buffer junk initialization; both v5-related but different bugs." + "reason": "Trainer `_gpu_broadcast_one` NoneType error is unrelated to Whisper pipeline `return_language` behavior." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Image preprocessor/model dtype mismatch vs tokenizer template compilation error; no shared underlying change." + "reason": "Mistral-3 dtype mismatch and a template compilation TypeError do not share an underlying fix." }, { "left": "issue:46285", "right": "issue:46402", "accept": false, - "reason": "CLIP text-model structure inconsistency vs Sam3 text encoder weights not loading; separate model architecture/loading issues." + "reason": "CLIP structure inconsistency and Sam3Model text encoder weight loading are different model-specific issues." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "High perplexity on gpt-oss-20b vs runtime special-token/post-processor update behavior; unrelated problems." + "reason": "High perplexity on gpt-oss-20b is unrelated to tokenizer backend post-processor updates." }, { - "left": "issue:36032", - "right": "issue:42915", + "left": "issue:44589", + "right": "issue:45084", "accept": false, - "reason": "Tokenizer method name conflict vs Qwen3Moe FineGrainedFP8Config failure; different tokenizer/model-loading paths." + "reason": "Float8 storage lookup failure is unrelated to compile-time non-template node errors." }, { - "left": "issue:43643", - "right": "issue:44530", + "left": "issue:44683", + "right": "issue:44661", + "accept": false, + "reason": "Flex attention compilation on torch>=2.9 and `add-new-model-like` tokenizer mapping failure are unrelated." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "These are different model-loading failures; shared version context is not enough to merge." + }, + { + "left": "issue:43611", + "right": "issue:43644", "accept": false, - "reason": "Missing fields from `trust_remote_code=True` config objects vs Qwen3.5 cache crash; different APIs and failure modes." + "reason": "Shared v5 context only; the failure modes and code paths differ." } ] }, @@ -2850,7 +2938,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:34567", "issue:36032", @@ -2903,12 +2991,12 @@ ], "soft_pairs": [ "issue:34567|issue:40444", - "issue:40444|issue:46326", "issue:44451|issue:44987", "issue:43541|issue:45084", - "issue:44811|issue:45584", + "issue:40444|issue:46326", "issue:42175|issue:46178", "issue:42915|issue:44530", + "issue:44811|issue:45584", "issue:43644|issue:43646", "issue:44206|issue:44220", "issue:43452|issue:43482", @@ -2925,8 +3013,8 @@ "issue:43646|issue:44534", "issue:44403|issue:45081", "issue:43404|issue:44466", - "issue:42915|issue:45084", "issue:43883|issue:44991", + "issue:42915|issue:45084", "issue:43089|issue:45987", "issue:43994|issue:44610", "issue:43322|issue:43540", @@ -2938,244 +3026,244 @@ "issue:42915|issue:44291", "issue:44220|issue:44683" ], - "prepared_review_unit_hash": "dd254cdcf3bf9ad46bb27d4dda0415656d71964f39f0bbe813097e96d1f6c69d", + "prepared_review_unit_hash": "243551020d5c6ff16764e75c7ba540c4340c8f1b8e7f5eb02b0a44d3763f3205", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 38320, - "estimated_input_tokens": 9580, - "estimated_eval_tokens": 19416 + "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 a loose set of mostly unrelated Transformers regressions and model-loading issues. The soft pairs generally share only broad subsystem similarity, not the same concrete bug or fixable code path, so they should be rejected as duplicates.", - "confidence": 0.68, - "canonical_issue_reason": "issue:43646 is the broadest representative in the set: it describes a general Transformers 5.0.0 custom model initialization regression, which is closer to the recurring v5 initialization/load failures than the more specific or unrelated issues nearby.", + "summary": "This cluster is heterogeneous: most items are unrelated v5 regressions across loading, multimodal, audio, generation, and custom-model paths. I did not find a clear duplicate pair with enough evidence to merge.", + "confidence": 0.84, + "canonical_issue_reason": "No single canonical issue cleanly represents the whole cluster; if one must be chosen, issue 43646 is the broadest v5 regression around custom model initialization.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best issue candidate because it is the most general and reusable description of the recurring v5 initialization breakage, while the other candidates are narrower symptoms (Jupyter-only, non-persistent buffers, tokenizer/model-specific loading failures).", + "best_issue_reason": "Issue 43646 is the most general and representative of the recurring v5 custom-model/init regressions in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Unrelated bugs: token-count tracking in TrainerState vs multimodal IterableDataset loading." - }, - { - "left": "issue:40444", - "right": "issue:46326", - "accept": false, - "reason": "Both mention model-processing failures, but one is Qwen2.5-VL multi-image dataset handling and the other is mistral_common continue_final_message behavior." + "reason": "Unrelated bugs: trainer token-count tracking vs multimodal fine-tuning failure." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Different model-loading failures for different checkpoints; no evidence of the same underlying bug." + "reason": "Both are load failures, but for different models and likely different underlying causes." }, { "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "One is a grouped_mm tracing failure in Mixtral; the other is a template compilation error. Different code paths." + "reason": "Different code paths: torch dynamo grouped_mm tracing vs template compilation error." }, { - "left": "issue:44811", - "right": "issue:45584", + "left": "issue:40444", + "right": "issue:46326", "accept": false, - "reason": "Both involve Whisper generation/token handling, but one is batch_decode skip_special_tokens and the other is empty transcription after align_special_tokens." + "reason": "Different areas: iterable-dataset multimodal finetuning vs tokenizer continue_final_message behavior." }, { "left": "issue:42175", "right": "issue:46178", "accept": false, - "reason": "Backend packaging vs backend registry validation are related at a high level, but not the same defect." + "reason": "Both mention backend packaging, but one is torch extras and the other is tensorflow_text backend registration." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Different model families and failures: FineGrainedFP8Config on Qwen3Moe vs PagedAttentionCache linear_attention crash on Qwen3.5." + "reason": "Different failures in different Qwen model families and different runtime components." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Whisper-related, but one is batch_decode skipping special tokens and the other is empty-transcription generation after alignment." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are Transformers 5 regressions, but one is non-persistent buffer initialization and the other is custom model initialization." + "reason": "Related v5 init regressions, but one is non-persistent buffer corruption and the other is custom model initialization; not clearly the same bug." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Similar audio preprocessing area, but the failures differ: unsupported center arg vs _torch_extract_fbank_features behavior." + "reason": "Both audio-feature extraction related, but the titles point to different concrete failures." }, { "left": "issue:43452", "right": "issue:43482", "accept": false, - "reason": "Both are gguf-related loading problems, but one targets AutoTokenizer/AutoModelForCausalLM with gguf_file and the other is Qwen2.5-GGUF under v5." + "reason": "Both involve GGUF loading, but one is a from_pretrained/gguf_file API break and the other is a specific Qwen2.5-GGUF regression." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Same broad loader regression theme, but one is missing all_tied_weights_keys and the other is a specific model load failure." + "reason": "Completely different problems: missing tied-weights metadata vs model loading failure for a specific checkpoint." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer class/method conflict vs notebook custom model initialization; unrelated." + "reason": "Tokenizer method conflict vs custom model initialization in notebooks." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Device-map offload/dequantized model loading vs GGUF model loading; different failure surfaces." + "reason": "Different loading problems: device_map/offload dequantization vs GGUF/v5 model loading." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation inconsistencies vs StaticCache.get_seq_length return type; not the same bug." + "reason": "Both generation/cache-adjacent, but one is compiled-forward chunked generation inconsistency and the other is a return-type contract bug." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Template compilation error vs GraniteMoEHybrid invalid method call; different underlying issues." + "reason": "Different issues entirely: template-node compilation vs GraniteMoEHybrid invalid method call." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Both are v5 loading regressions, but they affect different models and likely different loader/tokenizer paths." + "reason": "Both are version-gated loading failures, but they affect different models and likely different loaders." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Different Qwen3.5 bugs: hidden-state output vs rope index scaling for Qwen2.5-VL." + "reason": "Different Qwen variants and different bugs: hidden-states output vs rope-index scaling." }, { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "PEFT/local-checkpoint loading for GraniteSpeech is unrelated to mistral_common continue_final_message handling." + "reason": "PEFT checkpoint loading vs tokenizer continue_final_message logic; no shared bug." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Audio sampling-rate defaulting vs MIMI encoder padding-mask correctness; different components and symptoms." + "reason": "Audio chat-template sampling-rate default vs MIMI encoder padding-mask mismatch." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Notebook/custom model initialization bug vs non-persistent buffer junk; related only by being v5 regressions." + "reason": "Both are v5 regressions around model initialization, but the symptoms are different and not clearly one code-path bug." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "General custom initialization regression vs buffer initialization corruption; not the same code path." + "reason": "Custom model initialization breakage vs non-persistent buffer junk; related theme, but not the same concrete defect." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Unnecessary loading noise vs Mistral regex patch backend_tokenizer AttributeError; unrelated." + "reason": "Generic loading noise vs Mistral regex patch crash; unrelated." }, { "left": "issue:43404", "right": "issue:44466", "accept": false, - "reason": "lm_head tying bug in Mistral3 vs device-dependent lm_head serialization inconsistency; related around weights, but not the same defect." + "reason": "Both mention tied lm_head weights, but one is an untied model class and the other is device-dependent serialization inconsistency." }, { - "left": "issue:42915", - "right": "issue:45084", + "left": "issue:43883", + "right": "issue:44991", "accept": false, - "reason": "Qwen3Moe FP8 config failure and template compilation failure are unrelated." + "reason": "Unrelated: missing all_tied_weights_keys vs tokenizer loading regression for a specific model." }, { - "left": "issue:43883", - "right": "issue:44991", + "left": "issue:42915", + "right": "issue:45084", "accept": false, - "reason": "Different loading failures for different models and causes; no shared concrete bug." + "reason": "Different model/runtime failures; no shared code path." }, { "left": "issue:43089", "right": "issue:45987", "accept": false, - "reason": "Generation performance overhead vs cache length return type contract; different problem classes." + "reason": "Generation performance overhead vs cache length return-type bug." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 incorrect outputs via AutoModel/pipeline vs OmDet-Turbo processor/model resolution mismatch; not the same issue." + "reason": "Different modalities and components: image/text model output quality vs processor image-size mismatch." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Segfault on Llava Next loading vs Qwen3OmniMoe video-input ValueError; both multimodal, but distinct failures." + "reason": "Both multimodal, but one is a segmentation fault on Llava Next load and the other is a video-input ValueError in Qwen3OmniMoe." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "MIMI batched-vs-single padding-mask mismatch vs DAC.from_latents missing STE mismatch; unrelated audio-model correctness bugs." + "reason": "Different audio model bugs: MIMI padding-mask mismatch vs DAC from_latents/forward mismatch." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "Different model-specific loading failure vs Mistral tokenizer regex patch crash." + "reason": "Different model-loading failures; no evidence they share the same loader bug." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Model loading regression for physical-intelligence/fast vs Mistral tokenizer regex backend issue; different subsystems." + "reason": "Different failures in different tokenizer/model paths; not the same underlying bug." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export/exportability with torch_compilable_check vs processor size mismatch; not the same defect." + "reason": "torch.export export failure vs processor output-size mismatch; unrelated." }, { "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "Tokenizer loading failure for a specific model vs slow convert_tokens_to_ids regression; different functions and symptoms." + "reason": "Tokenizer loading regression vs slow convert_tokens_to_ids performance regression; different defects." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "FineGrainedFP8Config on Qwen3Moe vs init_empty_weights/_is_hf_initialized argument mismatch; unrelated initialization paths." + "reason": "Different bugs and different triggers: FineGrainedFP8Config vs init_empty_weights/_is_hf_initialized handling." }, { "left": "issue:44220", "right": "issue:44683", "accept": false, - "reason": "Audio feature extraction issue vs compiled flex_attention on torch>=2.9; unrelated." + "reason": "Audio fbank feature extraction issue vs compiled flex_attention failure; unrelated." } ] }, @@ -3197,7 +3285,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:33666", "issue:35707", @@ -3251,15 +3339,15 @@ "issue:45216|issue:45357", "issue:36032|issue:43742", "issue:44062|issue:44987", - "issue:46209|issue:46326", - "issue:43632|issue:44756", "issue:42915|issue:44589", + "issue:43632|issue:44756", + "issue:46209|issue:46326", "issue:44479|issue:45325", "issue:36032|issue:41720", - "issue:43404|issue:45127", "issue:43531|issue:44877", - "issue:40444|issue:45325", + "issue:43404|issue:45127", "issue:43296|issue:43531", + "issue:40444|issue:45325", "issue:43611|issue:44534", "issue:43425|issue:44403", "issue:40990|issue:44811", @@ -3278,196 +3366,208 @@ "issue:44829|issue:45479", "issue:43482|issue:43525" ], - "prepared_review_unit_hash": "624248e23865daf1e07182c9a36a9e271ad65da8ec7db4a6240d8068ca30b4ca", + "prepared_review_unit_hash": "87b9bb80960ab254c0b4a4c09696543f80dbb9cdeea242217155a600efa7a50c", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35802, - "estimated_input_tokens": 8951, - "estimated_eval_tokens": 18158 + "serialized_chars": 35809, + "estimated_input_tokens": 8953, + "estimated_eval_tokens": 18162 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are distinct regressions or feature requests across tokenizers, Qwen/VL, FP8, checkpointing, and loading paths. There is no strong duplicate set, so the best single issue is only an umbrella representative, not a true canonical duplicate.", - "confidence": 0.31, - "canonical_issue_reason": "issue:45020 is the broadest umbrella report in the set, covering recent-version breakages in model loading/remote code, but it still does not subsume the many unrelated specific bugs here.", + "summary": "This cluster is heterogeneous: most items are distinct loading/regression bugs that only share broad transformer-version or model-family similarity. I would not merge any of the soft pairs as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "issue:45020 is the broadest cluster anchor: it describes recent-transformers breakage across models using remote_code, which is the closest high-level umbrella among these otherwise model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best representative because it captures the widest cross-cutting regression theme among these otherwise unrelated issues.", + "best_issue_reason": "issue:45020 is the best representative issue because it captures the widest symptom pattern in the set; however, the cluster is still too diverse for duplicate consolidation.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45216", "right": "issue:45357", "accept": false, - "reason": "Both are Qwen3.5 save_pretrained regressions, but they describe different failures: overall checkpoint corruption vs incorrect visual-encoder keys. Not the same concrete bug." + "reason": "Both are Qwen3.5 save_pretrained regressions, but one is about incorrect checkpoint contents generally and the other specifically about visual encoder key saving; related, not the same bug." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer constructor conflict in T5 vs a KeyError while loading MobileLLM-125M. Different subsystems and symptoms." + "reason": "Different tokenizer/model-load failures: T5 tokenizer method-name conflict versus a MobileLLM key error." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "AddedToken special-argument duplication is a tokenizer API bug; loading physical-intelligence/fast is a model loading regression. No shared concrete code path." + "reason": "Both mention loading problems, but the first is an AddedToken special-argument error while the second is a recent-version model load failure for a specific repo." }, { - "left": "issue:46209", - "right": "issue:46326", + "left": "issue:42915", + "right": "issue:44589", "accept": false, - "reason": "FP8 inference failure on B200 and continue_final_message tokenizer behavior are unrelated." + "reason": "Both involve fp8-related failures, but one is Qwen3MoE with FineGrainedFP8Config and the other is a missing Float8 storage object; different code paths." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "_is_hf_initialized init regression and Strix Halo mmap OOM are unrelated bugs." + "reason": "Transformers v5 internal flag breakage versus mmap OOM on Strix Halo; unrelated symptoms and fixes." }, { - "left": "issue:42915", - "right": "issue:44589", + "left": "issue:46209", + "right": "issue:46326", "accept": false, - "reason": "Both mention FP8, but one is a Qwen3MoE FineGrainedFP8Config failure and the other is a missing Float8 storage type error. Different failure modes." + "reason": "FP8 inference on B200 and continue_final_message behavior are unrelated problems." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "Both involve Qwen VL media handling, but one is a broad video-input regression while the other is a specific still-image temporal position_id scaling bug." + "reason": "Both are Qwen2.5-VL/video-related regressions, but one is a broad v5.3 video-input issue and the other is a rope-index scaling bug for still-image positions." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "T5 tokenizer load error and Qwen3 auto device mapping cuda assert are unrelated." - }, - { - "left": "issue:43404", - "right": "issue:45127", - "accept": false, - "reason": "Mistral3 lm_head tying bug and LoRA merge collapse with extended vocab are different issues, even if both touch tied weights." + "reason": "Tokenizer load conflict versus Qwen3 auto device-map cudaErrorAssert; no shared underlying bug." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and granite_speech strict-config loading are unrelated." + "reason": "Sliding_window behavior in Qwen3-MoE and strict config loading for granite_speech are different issues." }, { - "left": "issue:40444", - "right": "issue:45325", + "left": "issue:43404", + "right": "issue:45127", "accept": false, - "reason": "Both are Qwen2.5-VL-related, but one is training with multiple images per prompt and the other is a still-image position-id scaling regression." + "reason": "One is a missing lm_head weight tie in Mistral3 image-text generation, the other is LoRA merge collapse with extended vocab; different mechanisms." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and Qwen3-MoE sliding_window are different model/load-path bugs." + "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window bug are unrelated." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "IterableDataset multi-image fine-tuning failure is not the same as the Qwen2.5-VL rope-index scaling regression." }, { "left": "issue:43611", "right": "issue:44534", "accept": false, - "reason": "base_model_prefix loading breakage and non-persistent buffer junk are unrelated implementation issues." + "reason": "Both are Transformers v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffers getting junk values." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch 2.10 incompatibility is a compatibility issue; unnecessary load noise is a separate UX/logging bug." + "reason": "Torch 2.10 incompatibility and loading noise are not the same bug." }, { "left": "issue:40990", "right": "issue:44811", "accept": false, - "reason": "High perplexity on gpt-oss-20b and Whisper batch_decode skip_special_tokens behavior are unrelated." + "reason": "Extremely high perplexity on GPT-OSS and Whisper batch_decode skip_special_tokens are unrelated." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper-large-v2 runtime failure and missing Float8 storage type are unrelated." + "reason": "Whisper model execution failure versus missing Float8 storage object; different failure modes." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "Different regressions: init flag handling vs buffer initialization." + "reason": "Both are v5 regressions, but one concerns _is_hf_initialized and the other buffer initialization; not duplicate." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment bug and Gemma-4 FSDP2 shared_kv_states KeyError are unrelated." + "reason": "ModelOutput key handling and Gemma4 FSDP2 shared_kv_states rebuilding are unrelated." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Loading noise and tokenizer loading regression for EMBEDDIA/est-roberta are different issues." + "reason": "Loading noise and tokenizer-loading failure for est-roberta are different bugs." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both are Qwen-VL training reports, but multi-GPU training and iterable-dataset multi-image failure are not the same bug." + "reason": "Both are Qwen2-VL/Qwen2.5-VL training-related, but one is multi-GPU training and the other is multi-image iterable dataset finetuning; not the same underlying issue." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Runtime special-token post-processor update bug and mdeberta add_special_tokens BOS/EOS regression are related in theme but not the same concrete failure." + "reason": "Runtime special-token post-processor updates and BOS/EOS not being added on add_special_tokens=True are tokenizer bugs, but they are distinct behaviors and fixes." }, { "left": "issue:44683", "right": "issue:45020", "accept": false, - "reason": "Torch flex_attention compilation failure and remote_code loading regressions are unrelated." + "reason": "Compiled flex_attention on torch>=2.9 is a torch/kernel compatibility issue, not the remote_code loading regression." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction and missing supports_gradient_checkpointing on a trust_remote_code model are different code paths." + "reason": "Gradient checkpointing in eval mode and missing supports_gradient_checkpointing on a trust_remote_code model are related only superficially." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "Unexpected _is_hf_initialized argument during init_empty_weights and missing Float8 storage type are separate loading/runtime issues." + "reason": "Unexpected _is_hf_initialized argument during init_empty_weights and missing Float8 storage object are separate transformer-v5 failures." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values and Qwen2.5-VL multi-image finetuning failure are unrelated." + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen2.5-VL multi-image iterable finetuning." }, { "left": "issue:44462", "right": "issue:45812", "accept": false, - "reason": "Both concern AutoTokenizer behavior, but one ignores tokenizer.json from the repo and the other produces wrong Granite token IDs. Too little evidence of the same underlying bug." + "reason": "Tokenizer.json being ignored and Granite models getting wrong token IDs are connected only at a high level; they are not the same bug." }, { "left": "issue:43482", "right": "issue:45538", "accept": false, - "reason": "Qwen2.5-GGUF loading with transformers v5 and CLIPTokenizer max_length semantics are unrelated." + "reason": "Qwen2.5-GGUF loading under v5 and CLIPTokenizer model_max_length are different tokenizer/loading problems." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Flash-attention-3 training degeneration and single_label_classification zero-loss are both training regressions, but they affect different code paths." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure and missing pad_token_id on Llama4Config are unrelated configuration/loading issues." } ] }, @@ -3489,7 +3589,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:40444", "issue:41553", @@ -3549,229 +3649,229 @@ "issue:43531|issue:44291", "issue:41720|issue:45081", "issue:43377|issue:45491", - "issue:40444|issue:45381", "issue:43644|issue:44079", + "issue:40444|issue:45381", "issue:44265|issue:45072", "issue:43749|issue:44079", - "issue:45325|issue:45381", "issue:43611|issue:43646", + "issue:45325|issue:45381", "issue:42175|issue:43296", - "issue:43632|issue:45072", "issue:43792|issue:43957", + "issue:43632|issue:45072", "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:44479|issue:45823", "issue:43873|issue:45701", - "issue:43606|issue:44265", "issue:43296|issue:45800", + "issue:43606|issue:44265", "issue:43262|issue:43746" ], - "prepared_review_unit_hash": "578ee650938b455b86ab8e7a7ee2321bd80b2f8aa788c28ca3afb72bbfded906", + "prepared_review_unit_hash": "1ab9eefd3714af22a32703d336c1b602266364855862cdd719fab187abb645fa", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35885, - "estimated_input_tokens": 8972, - "estimated_eval_tokens": 18200 + "serialized_chars": 35892, + "estimated_input_tokens": 8973, + "estimated_eval_tokens": 18202 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly separate bug reports, not true duplicates. The only clear common theme is a broad set of Transformers v5 loading/init regressions, but the proposed soft pairs each point to different code paths or different model families, so none should be merged.", - "confidence": 0.86, - "canonical_issue_reason": "Issue 43611 is the broadest, most representative loading regression in the set; it captures a core Transformers v5 breakage affecting model loading, whereas the others are narrower variants or unrelated bugs.", + "summary": "This cluster is mostly a semantic near-match set, not a true duplicate set: a few items overlap around Transformers v5 loading/initialization regressions and Qwen2.5-VL video/position-id issues, but the individual bugs are materially different. I would not merge any of the soft edges.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43611 is the broadest and most representative anchor in the cluster: a clear Transformers v5 model-loading regression. It sits near several other v5 init/load complaints, even though the cluster overall is too heterogeneous to treat as one duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "43611 is the best single issue to anchor a broader v5 initialization/loading regression cluster because it is general, concrete, and likely relevant to multiple downstream failures.", + "best_issue_reason": "Issue 43611 is the best single representative because it is concrete, broadly scoped, and central to the largest thematic slice of the cluster (v5 loading/init breakages).", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Both mention device/offload behavior, but one is a dequantized model load failure and the other is Qwen3 MoE training memory usage; different bug classes." + "reason": "Different problems: device-map/offload loading failure vs Qwen3 MoE training memory usage." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both involve TensorFlow/backend packaging, but one is a missing dependency in the torch extra and the other is a tensorboard-without-TF runtime backend error." + "reason": "Both involve TensorFlow/backend packaging, but one is missing TF in the torch extra while the other is an undefined tf backend from tensorboard-only installation." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Different models and failures: Video-LLaVA video_tower wiring vs Sam3Model text encoder weights not loading." + "reason": "Unrelated multimodal model bugs: Video-LLaVA visual tower vs Sam3Model text encoder weight loading." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "AutoTokenizer/Voxtral error messaging and Apple Silicon caching_allocator_warmup TypeError are unrelated." + "reason": "Completely different domains: tokenizer error messaging vs Apple Silicon allocator warmup TypeError." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "ModelOutput key assignment bug and non-persistent buffer junk are distinct internal behaviors." + "reason": "One is ModelOutput key assignment, the other is non-persistent buffer initialization in v5; not the same bug." }, { "left": "issue:43818", "right": "issue:45357", "accept": false, - "reason": "Video-LLaVA tower/weight-sharing bug is unrelated to Qwen3.5 save_pretrained key serialization." + "reason": "Both touch vision models, but one is Video-LLaVA tower/weights and the other is Qwen3.5 save_pretrained key serialization." }, { "left": "issue:43531", "right": "issue:44291", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and init_empty_weights/_is_hf_initialized TypeError are different regressions." + "reason": "Both are v5-related, but sliding-window Qwen3-MoE behavior and init_empty_weights argument handling are distinct bugs." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert and Mistral regex patching crash are unrelated code paths." + "reason": "No underlying overlap: Qwen3 auto device mapping CUDA assert vs Mistral tokenizer regex patch crash." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "MIMI padding-mask batching bug and Gemma3 NaN embeddings from sliding-window all-padding windows are different issues." + "reason": "Both mention batching/padding behavior, but they affect different models and different failure modes." }, { - "left": "issue:40444", - "right": "issue:45381", + "left": "issue:43644", + "right": "issue:44079", "accept": false, - "reason": "IterableDataset multi-image finetuning failure is not the same as Qwen2.5-VL video vision_position_ids regression." + "reason": "Different regressions: buffer serialization/initialization vs ModelOutput key bookkeeping." }, { - "left": "issue:43644", - "right": "issue:44079", + "left": "issue:40444", + "right": "issue:45381", "accept": false, - "reason": "Both touch model internals, but one is non-persistent buffer serialization and the other is ModelOutput key handling." + "reason": "Both are Qwen2.5-VL vision issues, but one is finetuning with multiple images and the other is video vision_position_ids." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "torch.export with torch_compilable_check is a different failure mode from bf16 dtype mismatches in inference." + "reason": "No common bug path: torch.export/exportability vs bfloat16 dtype mismatch in inference." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are unrelated." + "reason": "FSDP CPU-RAM-efficient loading is unrelated to ModelOutput key assignment." }, { - "left": "issue:45325", - "right": "issue:45381", + "left": "issue:43611", + "right": "issue:43646", "accept": false, - "reason": "Both are Qwen2.5-VL position-id issues, but one is still-image RoPE scaling and the other is video input vision_position_ids; not the same bug." + "reason": "Both are v5 loading/init regressions, but one targets base_model_prefix loading and the other custom model initialization; too broad to call duplicates." }, { - "left": "issue:43611", - "right": "issue:43646", + "left": "issue:45325", + "right": "issue:45381", "accept": false, - "reason": "Both are Transformers v5 regressions, but base_model_prefix loading failure and custom model initialization breakage are distinct concrete problems." + "reason": "Adjacent Qwen2.5-VL position-id regressions, but one is still-image temporal scaling and the other is video input vision_position_ids." }, { "left": "issue:42175", "right": "issue:43296", "accept": false, - "reason": "TensorFlow packaging/backend issue and PaddleOCR-VL loading failure in vLLM are unrelated." + "reason": "Packaging/backend issue vs a specific PaddleOCR-VL load failure in vLLM; not the same bug." }, { - "left": "issue:43632", - "right": "issue:45072", + "left": "issue:43792", + "right": "issue:43957", "accept": false, - "reason": "_is_hf_initialized flag breakage and bfloat16 dtype mismatch are different failure modes." + "reason": "Whisper runtime failure and meta-device model loading breakage are unrelated." }, { - "left": "issue:43792", - "right": "issue:43957", + "left": "issue:43632", + "right": "issue:45072", "accept": false, - "reason": "Whisper can't run and torch.device('meta') loading failures affect different model-loading paths." + "reason": "Different code paths: _is_hf_initialized flag breakage vs dtype mismatches during bfloat16 inference." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Siglip2 nonsensical outputs and Perceiver non-default resolution failure are different vision-model bugs." + "reason": "Both are vision-model inference issues, but they involve different architectures and different root causes." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Both are long-context/video-style positional issues, but one is Qwen2.5-VL RoPE scaling and the other is Kosmos2.5 OCR index overflow." + "reason": "Both are multimodal indexing/positioning style bugs, but on different models with different symptoms and likely different fixes." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "AyaVision lm_head tying and inconsistent lm_head serialization are related only loosely; they are not the same concrete bug." + "reason": "Both mention lm_head/tied weights, but one is a load/tie failure in AyaVision and the other is device-dependent serialization inconsistency." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3Moe FP8 failure and Qwen2.5-GGUF loading under v5 are unrelated." + "reason": "Different issues: Qwen3Moe FP8 config failure vs Qwen2.5-GGUF loading under transformers v5." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "resume_from_checkpoint key mismatch is unrelated to Qwen2.5-VL position-id scaling." + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL position-id scaling." }, { - "left": "issue:44479", - "right": "issue:45823", + "left": "issue:44291", + "right": "issue:44877", "accept": false, - "reason": "Qwen video-input regression and Gemma4 PLE device mismatch are different model-specific failures." + "reason": "Both are loading regressions, but one is init_empty_weights argument handling and the other is strict config validation for granite_speech." }, { - "left": "issue:44291", - "right": "issue:44877", + "left": "issue:44479", + "right": "issue:45823", "accept": false, - "reason": "init_empty_weights TypeError and strict config blocking granite_speech config are different issues." + "reason": "Qwen video-input regression and Gemma4 device_map offload mismatch are unrelated." }, { "left": "issue:43873", "right": "issue:45701", "accept": false, - "reason": "Quantization/offloading behavior and tokenization changes across versions are not duplicates." + "reason": "Quantization/offloading behavior is unrelated to tokenization changes across versions." }, { - "left": "issue:43606", - "right": "issue:44265", + "left": "issue:43296", + "right": "issue:45800", "accept": false, - "reason": "CPU offload device mismatch and torch.export/torch_compilable_check failure are unrelated." + "reason": "PaddleOCR-VL load failure in vLLM and torch/transformers version incompatibility are not the same underlying bug." }, { - "left": "issue:43296", - "right": "issue:45800", + "left": "issue:43606", + "right": "issue:44265", "accept": false, - "reason": "PaddleOCR-VL load failure in vLLM and torch 2.4.1/transformers 5.8 incompatibility are different problems." + "reason": "CPU offload device mismatch and torch.export exportability are different failure modes." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are unrelated." + "reason": "Audio processor sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are unrelated." } ] }, @@ -3793,7 +3893,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:34634", "issue:35707", @@ -3846,8 +3946,8 @@ "soft_pairs": [ "issue:40990|issue:45584", "issue:34634|issue:35707", - "issue:35707|issue:41720", "issue:44568|issue:45920", + "issue:35707|issue:41720", "issue:44186|issue:44265", "issue:41720|issue:45823", "issue:43606|issue:43819", @@ -3864,225 +3964,218 @@ "issue:43377|issue:44898", "issue:43322|issue:46050", "issue:43454|issue:43720", - "issue:41720|issue:43322", "issue:45325|issue:46315", + "issue:41720|issue:43322", "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" + "issue:40990|issue:45701" ], - "prepared_review_unit_hash": "a95972fe420ce92760dbf0d0f1010eb0b6a5d4ba370778af94ce55603fe55465", + "prepared_review_unit_hash": "527a6a29d1ee542da8bafe66e76dc3fbd33db5b67e53f14018ccbc6c08e8def1", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 36234, - "estimated_input_tokens": 9059, - "estimated_eval_tokens": 18374 + "soft_pair_count": 29, + "serialized_chars": 35932, + "estimated_input_tokens": 8983, + "estimated_eval_tokens": 18222 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No strong duplicate cluster emerges here. The soft links are mostly broad thematic overlaps (device_map/offloading, tokenizer regressions, serialization, generation) rather than the same underlying bug. No PRs are present.", - "confidence": 0.94, - "canonical_issue_reason": "issue:41720 is the clearest concrete runtime regression in the set and the most central anchor for the device-mapping/offloading subgroup, even though the neighboring reports are not true duplicates.", + "summary": "This set is a loose mix of mostly unrelated Transformers bugs and regressions. A few are superficially similar (tokenizer/version regressions, device/offload failures, multimodal loading issues), but none of the soft pairs look like the same underlying defect well enough to merge.", + "confidence": 0.19, + "canonical_issue_reason": "issue:45701 is the broadest umbrella report in the set: it frames a version-wide tokenization regression with multiple inbound references and can represent the general tokenizer breakage theme better than narrower model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:41720 is the best single issue to anchor triage because it is specific, reproducible, and has the strongest thematic overlap with several other candidates, but it should not absorb the others as duplicates.", + "best_issue_reason": "issue:45920 is the most actionable standalone bug report: it names concrete affected models, a clear regression, and has enough scope to be a useful reference for tokenization correctness issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "Different bugs: high perplexity on WikiText-2 vs Whisper empty-transcription generation after token alignment." + "reason": "Both mention generation quality, but one is perplexity on GPT-OSS and the other is empty-transcription Whisper failure; different models and code paths." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Processor voice_preset bug and progressive generation with inputs_embeds/past_key_values are unrelated code paths." + "reason": "Bark voice preset failure is audio-processing specific, while the other is a generation bug with inputs_embeds/past_key_values; no shared underlying defect." }, { - "left": "issue:35707", - "right": "issue:41720", + "left": "issue:44568", + "right": "issue:45920", "accept": false, - "reason": "Generation cache behavior and Qwen3 auto device-map CUDA assert are different failures." + "reason": "Both concern tokenizer behavior after v5, but one is mdeberta BOS/EOS insertion and the other is wrong token IDs across several models; likely separate tokenizer regressions." }, { - "left": "issue:44568", - "right": "issue:45920", + "left": "issue:35707", + "right": "issue:41720", "accept": false, - "reason": "One is a missing BOS/EOS insertion bug for a specific tokenizer; the other is a broad wrong-token-IDs regression across multiple models." + "reason": "Progressive generation with cached embeddings is unrelated to Qwen3 auto device-mapping cuda assert; different subsystems and symptoms." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer NER padding/truncation crash is not the same as torch.export.export failing on torch_compilable_check." + "reason": "LayoutLMv2 tokenizer NER/padding crash is unrelated to torch.export failures around torch_compilable_check." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Both involve device_map=auto, but they affect different models and fail differently (CUDA assert vs device mismatch in forward)." + "reason": "Both are device-map/forward failures, but one is a Qwen3 cuda assert and the other is a Gemma4 PLE mismatch; not the same bug." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "Bark CPU offload mismatch and DAC.from_latents missing STE are separate model/runtime bugs." + "reason": "Both are device/offload-related, but Bark CPU offload mismatch and DAC missing STE behavior are different model-specific issues." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "Weight tying/garbage text generation is unrelated to torch.compile + SDPA failures." + "reason": "AyaVision tied-weights garbage text and Bamba torch.compile/SDPA failure are distinct model implementation problems." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "Documentation example errors and multi-image IterableDataset finetuning failure are different problems." + "reason": "SigLIP2 doc example errors and Qwen2.5-VL iterable-dataset multi-image finetuning failure are unrelated examples of different workflows." }, { "left": "issue:43881", "right": "issue:44488", "accept": false, - "reason": "Different models with different load failures; no shared concrete code-path." + "reason": "glm-4v-9b loading failure and sleng-bert tokenizer loading failure are separate model/load paths with no clear common root cause." }, { "left": "issue:46305", "right": "issue:46315", "accept": false, - "reason": "Sliding-window off-by-one in masking is unrelated to tokenizer conversion performance regression." + "reason": "One is a MoonshineStreaming sliding-window off-by-one bug, the other is slow convert_tokens_to_ids performance; different failure modes." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "Missing padding-mask support in MIMI encoder and OmDet processor/model size mismatch are not the same bug." + "reason": "MIMI batch-vs-single padding-mask inconsistency is unrelated to OmDet-Turbo processor/model image-size mismatch." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "lm_head tying/garbage generation is unrelated to bfloat16 dtype mismatches in inference." + "reason": "AyaVision weight tying and SwitchTransformers/TimmWrapper dtype mismatches are separate model correctness issues." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "MOE weight conversion with accelerate+deepspeed and a missing supports_gradient_checkpointing flag are unrelated." + "reason": "Qwen3 MOE accelerate+deepspeed weight conversion and Nemotron gradient-checkpointing flag are unrelated configuration/load issues." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig is unrelated to unwrap-text_config behavior for composite Qwen configs." + "reason": "StableLm pad_token_idx missing and Qwen 3.5/3.6 from_config text_config unwrapping are different config-layer bugs." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Closely themed around offloading/quantization, but not the same concrete failure; 43873 is a broad umbrella report." + "reason": "Device-map dequantized-model offload failure and generic offloading-with-quantization report are too broad and lack a concrete shared code path." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Perplexity regression on GPT OSS is unrelated to gradient checkpointing not taking effect with PEFT LoRA." + "reason": "High perplexity on GPT-OSS and gradient-checkpointing ineffective with LoRA are unrelated training/inference problems." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Batched-padding discrepancy in MIMI and Perceiver non-default-resolution failure are different issues." + "reason": "MIMI encoder padding mask bug and Perceiver non-default-resolution failure are distinct vision/audio model issues." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Segmentation fault when loading Llava Next and sinq save/load corruption are distinct load-path bugs." + "reason": "Llava Next segfault on load and sinq quantized save/load breakage are not the same failure, even though both involve model loading." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "Weight tying problem and BitNet packed-weight unpacking during accelerate loading are unrelated." + "reason": "AyaVision tied weights and BitNet packed-weight unpacking during accelerate loading are different implementation defects." }, { - "left": "issue:41720", - "right": "issue:43322", + "left": "issue:45325", + "right": "issue:46315", "accept": false, - "reason": "Both are model-loading/runtime failures, but the failure modes and affected models are different." + "reason": "Qwen2.5-VL rope-index scaling and tokenizer conversion performance regression are unrelated." }, { - "left": "issue:45325", - "right": "issue:46315", + "left": "issue:41720", + "right": "issue:43322", "accept": false, - "reason": "Qwen2.5-VL rope-index scaling bug is unrelated to tokenizer conversion speed regression." + "reason": "Qwen3 auto device-mapping assert and Llava Next segmentation fault are different model-loading failures." }, { "left": "issue:43749", "right": "issue:44534", "accept": false, - "reason": "FSDP CPU RAM efficient loading and junk-filled non-persistent buffers are separate regressions." + "reason": "FSDP CPU RAM efficient loading breakage and non-persistent buffer junk in v5 are not the same bug." }, { "left": "issue:44991", "right": "issue:46291", "accept": false, - "reason": "Tokenizer loading failure and kernels v0.15 import/version-revision failure do not share the same bug." + "reason": "Both are version-dependent loading failures, but one is tokenizer loading for a specific model and the other is kernels import requiring revision/version; different root causes." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel.py execution failure and FineGrainedFP8Config failure are different issues." + "reason": "3d_parallel.py runtime failure and Qwen3Moe FineGrainedFP8Config failure are unrelated errors." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Same model family, but different symptoms and likely different code paths (CUDA assert vs flash-attention generation issue)." + "reason": "Qwen3 cuda assert on A800 and Qwen3.5 flash-attention generation issues are separate inference paths and symptoms." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both concern saving/serialization, but one is tied lm_head.weight behavior and the other is incorrect visual encoder keys." + "reason": "Both involve save/load serialization around multimodal weights, but one is general lm_head tied-weight serialization and the other is Qwen3.5 visual encoder keys; not clearly the same defect." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug and Molmo missing all_tied_weights_keys are unrelated." + "reason": "BeitImageProcessorFast reduce_labels and Molmo tied-weights attribute error are unrelated image processor vs model attribute bugs." }, { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "Perplexity regression is not the same as version-dependent tokenization changes." - }, - { - "left": "issue:43881", - "right": "issue:44991", - "accept": false, - "reason": "Different models and different failure surfaces: model loading vs tokenizer loading." + "reason": "GPT-OSS perplexity and broad tokenization changes are too different to treat as one underlying bug." } ] }, @@ -4104,7 +4197,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36010", "issue:36331", @@ -4128,6 +4221,7 @@ "issue:43708", "issue:43720", "issue:43873", + "issue:43881", "issue:44008", "issue:44186", "issue:44387", @@ -4156,14 +4250,15 @@ ], "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:44877|issue:44991", + "issue:43577|issue:43708", "issue:36331|issue:42915", "issue:43450|issue:45701", + "issue:44008|issue:45823", "issue:43873|issue:44756", "issue:43550|issue:43720", "issue:43404|issue:45357", @@ -4173,220 +4268,244 @@ "issue:43550|issue:43606", "issue:44610|issue:44898", "issue:44186|issue:44610", - "issue:42915|issue:43317", "issue:43295|issue:45701", + "issue:42915|issue:43317", "issue:45632|issue:45684", "issue:43425|issue:46291", "issue:36010|issue:42617", "issue:41720|issue:45161", "issue:43525|issue:46291", + "issue:43531|issue:43572", "issue:45081|issue:45812", "issue:44671|issue:44743", - "issue:43531|issue:43572", "issue:43316|issue:44913" ], - "prepared_review_unit_hash": "a6b9e66582ea78f13181a010ad9bd07284df534df5e2bd50641768b19fcc7645", + "prepared_review_unit_hash": "1013963e773c095d7dc1d6de61d4455a0bed46a6987c5a331f28966cbdc4b89c", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35945, - "estimated_input_tokens": 8987, - "estimated_eval_tokens": 18230 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36769, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items span many unrelated regressions across different subsystems. None of the soft pairs look like the same underlying bug or change, so I would reject all duplicate merges. If an anchor is needed, issue 43873 is the broadest discussion-heavy issue, but it is not a true duplicate match for the others.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 43873 is the broadest and most actively discussed item here, with multiple inbound references and a recurring offloading/quantization theme, so it is the best available cluster anchor.", + "summary": "No genuine duplicate cluster emerges here: the soft pairs mostly connect unrelated model-, tokenizer-, trainer-, and serving-specific bugs that only share broad subsystem words like quantization, offloading, or tokenization.", + "confidence": 0.72, + "canonical_issue_reason": "issue:43873 is the broadest and most connected issue in this set (quantization/offloading with inbound references), so it is the best anchor if a single issue must represent the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the most suitable representative issue because it has the widest scope and the strongest evidence of ongoing community impact, even though this set does not contain a clear duplicate group.", + "best_issue_reason": "issue:43873 is the strongest global representative because it is active, broadly scoped, and externally referenced; the other issues are narrower point failures.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44387", "right": "issue:45704", "accept": false, - "reason": "Different problems: int4 quantization OOM vs T5/apex RMSNorm memory leak." + "reason": "Both are memory-related, but one is int4 quantization OOM and the other is a T5 apex RMSNorm leak; different code paths and root causes." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Different models and failures: glm-4v-9b loading vs est-roberta tokenizer loading." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Trainer custom compute_loss keyword mismatch is unrelated to failing to run a 3d_parallel script." + "reason": "Trainer compute_loss kwarg error is unrelated to a 3d_parallel.py runtime failure." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "Different SAM3 bugs: dummy Conv2d vs text encoder weights not loading." + "reason": "Same model family, but dummy Conv2d in Sam3PixelDecoder and missing text encoder weights are distinct bugs." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash on NER/padding is unrelated to dtype mismatch in bfloat16 inference." - }, - { - "left": "issue:43577", - "right": "issue:43708", - "accept": false, - "reason": "BLIP2 dtype propagation bug is unrelated to Trainer resume/max_steps recalculation." + "reason": "Tokenizer padding/NER crash is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." }, { - "left": "issue:44008", - "right": "issue:45823", + "left": "issue:44877", + "right": "issue:44991", "accept": false, - "reason": "Different Gemma models and failures: variable-name collision in forward vs PLE device mismatch." + "reason": "Config strictness for granite_speech is unrelated to est-roberta tokenizer loading." }, { - "left": "issue:44877", - "right": "issue:44991", + "left": "issue:43577", + "right": "issue:43708", "accept": false, - "reason": "Strict config loading for granite_speech is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Blip2 dtype propagation and Trainer resume/max_steps logic are separate issues." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "CustomTrainer loss signature mismatch is unrelated to Qwen3MoE FP8 failure." + "reason": "Unexpected compute_loss kwarg and Qwen3Moe FP8 failure are unrelated." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video processor batched-shape bug is unrelated to a broad tokenization version-change report." + "reason": "Video processor batch-shape bug is unrelated to tokenization changes." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma-related, but one is a variable name collision in forward and the other is a device-map mismatch in a different model." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Both involve loading/offloading, but one is quantization offload behavior and the other is mmap OOM on Strix Halo; not the same bug." + "reason": "Both mention offloading/memory, but one is quantization offload behavior and the other is mmap OOM on Strix Halo; not the same bug." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "Both are BitNet/SDPA vs packed-weight-unpacking issues; different failure paths and fixes." + "reason": "torch.compile with SDPA and BitNet packed-weight loading are different concrete failures." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Mistral3 lm_head tying bug is unrelated to Qwen3.5 visual encoder key saving." + "reason": "Mistral3 lm_head tying and Qwen3.5 visual-encoder key saving are unrelated." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "TensorFlow backend packaging issue is unrelated to kernels version/revision import failure in evaluate." + "reason": "Packaging/backend dependency selection and kernels version/revision requirements are different installation issues." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Qwen3 auto device-map cuda assert is unrelated to a script execution failure." + "reason": "Qwen3 device-mapping CUDA assert is unrelated to running 3d_parallel.py." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Continuous-batching multimodal crash is unrelated to Gemma4Processor missing _tokenizer in serve." - }, - { - "left": "issue:43550", - "right": "issue:43606", - "accept": false, - "reason": "Different subsystems and symptoms: torch.compile+SDPA in Bamba vs CPU offload device mismatch in bark-small." + "reason": "Both affect serve, but continuous-batching multimodal crash and missing Gemma4Processor tokenizer are separate code paths." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "OmDet processor output size bug is different from Perceiver non-default resolution failure." + "reason": "Both involve image size mismatches, but they are for different models and different failure mechanisms." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer/NER padding crash and image processor resolution mismatch are unrelated." + "reason": "LayoutLMv2 tokenizer crashes and OmDet-Turbo resolution mismatch are unrelated." }, { - "left": "issue:42915", - "right": "issue:43317", + "left": "issue:43295", + "right": "issue:45701", "accept": false, - "reason": "Qwen3Moe FP8 failure is unrelated to dequantized model loading with auto device_map offload." + "reason": "Processor.tokenizer regression and version-dependent tokenization changes are not the same specific bug." }, { - "left": "issue:43295", - "right": "issue:45701", + "left": "issue:42915", + "right": "issue:43317", "accept": false, - "reason": "Processor/tokenizer regression is unrelated to a general tokenization-version report." + "reason": "FP8 config failure and dequantized model offload failure are distinct quantization/loading problems." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Both touch save/load paths, but cache-path collisions and read-only permission propagation are different bugs." + "reason": "Cache-path collisions and read-only permission propagation from save_pretrained are unrelated." }, { "left": "issue:43425", "right": "issue:46291", "accept": false, - "reason": "Torch 2.10 incompatibility is unrelated to kernels version/revision requirements." + "reason": "Torch incompatibility and kernels dependency-version enforcement are different dependency issues." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import error is unrelated to failing to run 3d_parallel.py." + "reason": "GenerationMixin import error is unrelated to 3d_parallel.py execution." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Both mention model dispatching, but one is a cuda assert on Qwen3 and the other is only-TP not working on GPT-OSS MoE." + "reason": "Both touch parallelism/device mapping, but one is an A800 CUDA assert and the other is a GPT-OSS TP issue; not the same bug." }, { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is unrelated to kernels import/version requirements." + "reason": "Missing pad_token_id and kernels version/revision dependency errors are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Both mention missing padding-related config fields, but they affect different models and different attribute names." }, { "left": "issue:45081", "right": "issue:45812", "accept": false, - "reason": "Mistral regex patch crash is unrelated to Granite token ID regression." + "reason": "Mistral regex patch crash and Granite wrong token IDs are different tokenizer regressions." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked-LM predictions and Qwen3.5 recurrent-state handling are different code paths and failures." + "reason": "CamemBERT masked-LM prediction regression and Qwen3.5 recurrent-state reset are unrelated." }, { - "left": "issue:43531", - "right": "issue:43572", + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Duplicate of offloading/memory theme only; the underlying problems differ enough that they should stay separate." + }, + { + "left": "issue:43295", + "right": "issue:44991", + "accept": false, + "reason": "Processor API regression and tokenizer loading failure are not the same underlying defect." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "SDPA torch.compile failure and CPU-offload device mismatch are different model/runtime bugs." + }, + { + "left": "issue:44610", + "right": "issue:44186", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is unrelated to StableLmConfig missing pad_token_idx." + "reason": "Image processor resolution bug and LayoutLMv2 tokenizer crash are unrelated." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Gemma3TextConfig API mismatch is unrelated to GPTNeoX rotary_pct reload regression." + "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct reload regression are different config bugs." } ] }, @@ -4408,7 +4527,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:29159", "issue:30064", @@ -4461,8 +4580,8 @@ ], "soft_pairs": [ "issue:43540|issue:43782", - "issue:42915|issue:43322", "issue:44186|issue:44361", + "issue:42915|issue:43322", "issue:43421|issue:45812", "issue:36331|issue:41720", "issue:43526|issue:43881", @@ -4472,228 +4591,228 @@ "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:45406|issue:45464", + "issue:41720|issue:43746", "issue:43421|issue:45701", "issue:45357|issue:46402", "issue:43606|issue:43653", - "issue:44492|issue:44509", "issue:43531|issue:45759", "issue:34634|issue:41720", + "issue:44492|issue:44509", "issue:42490|issue:43329", "issue:34567|issue:35707", - "issue:44514|issue:44545", "issue:44625|issue:45479", "issue:44451|issue:44991", + "issue:44514|issue:44545", "issue:43531|issue:44589" ], - "prepared_review_unit_hash": "f6cc990cdb85435f074956b0b2f28365e4cd453d282d782424a3909e51fe0248", + "prepared_review_unit_hash": "53c3f91fdce4fcc41534ea0747acace67b9de523390f8016a1a2c5fc4acd462c", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36229, - "estimated_input_tokens": 9058, - "estimated_eval_tokens": 18372 + "serialized_chars": 36222, + "estimated_input_tokens": 9056, + "estimated_eval_tokens": 18368 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Only one soft edge looks like a true duplicate: the two Qwen2_5_VLProcessor.apply_chat_template crash reports. The rest are distinct bugs across different components and should stay separate.", - "confidence": 0.96, - "canonical_issue_reason": "issue:44514 is the cleanest canonical anchor for the batched apply_chat_template crash: it was filed slightly earlier and describes the failure mode directly. issue:44545 is essentially the same report with a near-identical title.", + "summary": "The cluster is mostly heterogeneous. There is one strong duplicate candidate: issues 44514 and 44545 share the same Qwen2_5_VLProcessor.apply_chat_template batched-input/padding=False title and appear to describe the same bug. The remaining pairs look like distinct issues in different models or code paths.", + "confidence": 0.76, + "canonical_issue_reason": "issue:44545 is the best canonical issue for the close duplicate pair because it matches issue:44514 on the same title and has slightly more inbound references/discussion activity.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44514 best represents the duplicated Qwen2_5_VLProcessor.padding=False crash because it is specific, actionable, and the duplicate pair maps to the same underlying code path.", + "best_issue_reason": "issue:44545 is the strongest issue representative in this cluster because it is the only accepted duplicate pair and is a direct match for issue:44514.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43540", "right": "issue:43782", "accept": false, - "reason": "Different models and different failure modes: Qwen3OmniMoe video-input ValueError vs Qwen3VL from_pretrained weight_only load error." + "reason": "Different models and failure modes: Qwen3OmniMoe video processing vs Qwen3VL from_pretrained weight_only loading." }, { - "left": "issue:42915", - "right": "issue:43322", + "left": "issue:44186", + "right": "issue:44361", "accept": false, - "reason": "Different multimodal model families and distinct load-time errors; no shared code-path evidence." + "reason": "Both are tokenizer-related, but the concrete failures differ: LayoutLMv2 NER/batching vs MLuke task AttributeError." }, { - "left": "issue:44186", - "right": "issue:44361", + "left": "issue:42915", + "right": "issue:43322", "accept": false, - "reason": "Both are tokenizer-related but the bugs differ: NER/batched padding crash in LayoutLMv2Tokenizer vs AttributeError in MLukeTokenizer." + "reason": "Unrelated loading failures: Qwen3Moe fine-grained FP8 config vs Llava Next segmentation fault." }, { "left": "issue:43421", "right": "issue:45812", "accept": false, - "reason": "Both involve tokenizers post-processing, but one is runtime special-token updates while the other is a Granite AutoTokenizer regression producing wrong IDs." + "reason": "Both involve tokenizers, but one is runtime post-processor syncing and the other is a Granite token-ID regression." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Unrelated areas: trainer compute_loss signature mismatch vs Qwen3 auto device mapping CUDA assert." + "reason": "Trainer loss-signature mismatch and Qwen3 auto device-map CUDA assert are unrelated bugs." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "Different bugs in different models: BeitImageProcessorFast label reduction vs glm-4v-9b loading failure." + "reason": "Different subsystems: BeitImageProcessorFast label reduction vs glm-4v model loading failure." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "Tokenizer inconsistency vs T5Tokenizer class/method conflict; these are not the same bug." + "reason": "Different tokenizer issues: fast/slow inconsistency vs T5 add_special_tokens name conflict." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "A MacOS bus error on a CLIP community model is unrelated to generation with inputs_embeds and past_key_values." + "reason": "MacOS CLIP bus error is unrelated to progressive generation with inputs_embeds/past_key_values." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Different subsystems and symptoms: gather_for_metrics label truncation vs SAM3 text/bbox behavior." + "reason": "Evaluation batching bug and SAM3 text/bounding-box behavior are different problems." }, { "left": "issue:30064", "right": "issue:34634", "accept": false, - "reason": "Both affect processors, but one is void segmentation-map handling and the other is BarkProcessor voice_preset support." + "reason": "Image segmentation-map processing and BarkProcessor voice_preset handling are unrelated." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "Different code paths: torch.compile+SDPA failure in Bamba vs tokenizer mask-token registration/empty decode issue in BigBirdTokenizer." + "reason": "torch.compile SDPA failure and BigBirdTokenizer special-token registration are different bugs." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "Both are CI/loading failures, but one is CPU offload device mismatch and the other is packed-weight unpacking in accelerate loading." + "reason": "Both involve loading/offload, but the underlying problems and models differ materially." }, { - "left": "issue:45406", - "right": "issue:45464", + "left": "issue:43873", + "right": "issue:45704", "accept": false, - "reason": "Different crash surfaces: Gemma4Processor missing _tokenizer in serve vs chat/completions streaming failure on Qwen3.5-0.8B." + "reason": "Quantization offloading behavior and a T5 RMSNorm memory leak are unrelated." }, { - "left": "issue:43873", - "right": "issue:45704", + "left": "issue:44530", + "right": "issue:46291", "accept": false, - "reason": "Quantization/offloading behavior is broad, but these are unrelated specific bugs: offloading with quantization vs T5 using apex.FusedRMSNorm memory leak." + "reason": "PagedAttentionCache linear_attention crash and kernels import/version validation are different code paths." }, { - "left": "issue:41720", - "right": "issue:43746", + "left": "issue:45406", + "right": "issue:45464", "accept": false, - "reason": "Different model families and causes: Qwen3 auto device mapping assert vs GraniteSpeech PEFT local checkpoint loading." + "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming inference failures are not the same bug." }, { - "left": "issue:44530", - "right": "issue:46291", + "left": "issue:41720", + "right": "issue:43746", "accept": false, - "reason": "PagedAttentionCache linear_attention group error is unrelated to kernels import/version requirements in evaluate." + "reason": "Qwen3 device-mapping CUDA assert is unrelated to GraniteSpeech PEFT local checkpoint loading." }, { "left": "issue:43421", "right": "issue:45701", "accept": false, - "reason": "Both involve tokenization behavior broadly, but one is runtime post-processor updates and the other is a version-dependent tokenization difference report." + "reason": "Runtime special-token/post-processor syncing is different from a broad tokenization-version regression." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Both concern Qwen3.5 checkpoints, but one is save_pretrained visual encoder keys and the other is loading text encoder weights; different operations and failures." + "reason": "Different save/load propagation bugs: Qwen3.5 visual encoder keys vs Sam3Model text encoder weights." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "No shared underlying bug: CPU offload device mismatch vs BigBirdTokenizer special-token registration." - }, - { - "left": "issue:44492", - "right": "issue:44509", - "accept": false, - "reason": "A typo in cache strategies is unrelated to documentation still mentioning removed pipeline tasks." + "reason": "CPU offload device mismatch is unrelated to BigBirdTokenizer mask-token registration." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Different Qwen MoE configuration issues: sliding_window support vs from_config not unwrapping text_config." + "reason": "Sliding-window behavior and composite config unwrapping are distinct Qwen-related issues." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "Different symptoms and model families: BarkProcessor voice_preset vs Qwen3 auto device mapping assert." + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping failure are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "These are different documentation issues: a typo vs stale pipeline-task references." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Local-path trailing-slash import warning is unrelated to a video-branch undefined-variable bug in multimodal token counting." + "reason": "Local-path import warning and multimodal token-counting video-branch bug are unrelated." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counting is unrelated to progressive generation with inputs_embeds/past_key_values." - }, - { - "left": "issue:44514", - "right": "issue:44545", - "accept": true, - "reason": "These are the same underlying crash report: Qwen2_5_VLProcessor.apply_chat_template fails on batched input when padding=False." + "reason": "TrainerState token counting and progressive generation cache handling are different problems." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Different classification bugs: num_labels propagation in Qwen3.5 vs degenerate zero loss across sequence-classification models." + "reason": "One is Qwen3.5 config propagation; the other is a general classification loss bug for num_labels=1." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer loading regressions, but they affect different checkpoints and likely different root causes." + "reason": "Both are tokenizer/model-loading regressions, but they affect different models and failure points." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same title and same reported Qwen2_5_VLProcessor.apply_chat_template batched-input crash with padding=False; this looks like the same bug." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Different model/config problems: Qwen3-MoE sliding_window vs missing Float8 storage type during loading." + "reason": "Sliding-window Qwen3-MoE bug and Float8 storage lookup error are unrelated." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft edge, issue:44514 \u2194 issue:44545, is a clear duplicate by title and timing. The remaining pairs are reasonably rejected as distinct bugs based on their titles and scopes, without overstating shared root cause evidence." + "accept": false, + "feedback": "The soft-edge rejections are mostly conservative and grounded in the issue titles, but the cluster-level conclusion overstates the evidence. In the packet, 44514 and 44545 have identical titles, which supports a strong duplicate candidate, but the summary goes beyond the packet by claiming they are the \"only true duplicate pair\" and that 44545 has the \"most engagement.\" That is not well supported: several other issues have higher discussion_activity, and the packet does not establish that the two reports describe the exact same crash beyond the matching title. The canonical-issue rationale also adds subjective claims like \"clearer/cleaner report\" that are not grounded in the packet. Because the summary and confidence are too strong for the available evidence, I would reject this cluster result." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -4709,7 +4828,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:30064", "issue:36010", @@ -4762,21 +4881,21 @@ ], "soft_pairs": [ "issue:44898|issue:45072", - "issue:44077|issue:44625", "issue:45759|issue:45859", + "issue:44077|issue:44625", "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:43782|issue:45381", "issue:43746|issue:46050", "issue:42886|issue:44843", "issue:45704|issue:45923", - "issue:43381|issue:43844", "issue:43421|issue:43653", + "issue:43381|issue:43844", "issue:45663|issue:45923", "issue:45704|issue:45859", "issue:41720|issue:45704", @@ -4786,229 +4905,229 @@ "issue:43746|issue:43873", "issue:45081|issue:45356", "issue:39692|issue:43873", - "issue:43540|issue:46294", "issue:43596|issue:44220", + "issue:43540|issue:46294", "issue:42915|issue:43746", "issue:45356|issue:45907", "issue:30064|issue:36010" ], - "prepared_review_unit_hash": "bf9844fa89c752951063787a1e740edb163382279cc2e30139c9bf4e35e187cf", + "prepared_review_unit_hash": "56c6001ee60254d943de2f43b006f895d733e20c3a9e4ddd3c6a5f34cd5be393", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 37078, - "estimated_input_tokens": 9270, - "estimated_eval_tokens": 18796 + "serialized_chars": 37083, + "estimated_input_tokens": 9271, + "estimated_eval_tokens": 18798 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous, with one strong duplicate pair around offline/Mistral-regex tokenizer loading (42886 \u2194 44843). A second Mistral-regex pair (45081 \u2194 45356) is related but too different in failure mode to treat as the same bug.", - "confidence": 0.84, - "canonical_issue_reason": "44843 is the best canonical issue for the only clearly duplicate subgroup: it names the concrete HF_HUB_OFFLINE regression in _patch_mistral_regex and matches 42886 closely.", + "summary": "This set is mostly a heterogeneous mix of unrelated Transformers issues. The only clear duplicate-style pair is the offline tokenizer/Mistral-regex regression; most other soft-similarity pairs share only a broad subsystem or model family, not the same concrete bug.", + "confidence": 0.79, + "canonical_issue_reason": "issue:44843 is the clearest root-cause report in the only near-duplicate pair: it pinpoints the offline regression to an unconditional `model_info()` call inside `_patch_mistral_regex`, while issue:42886 describes the same symptom at a higher level.", "canonical_pr_reason": null, - "best_issue_reason": "44843 is the strongest standalone issue here because it captures the specific offline tokenizer regression and serves as the clearest representative of the duplicate subgroup.", + "best_issue_reason": "issue:44843 is the best representative because it names the exact failing code path and bug trigger, making it the strongest canonical anchor for the cluster\u2019s only true duplicate subthread.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Different models and failure modes: image-size/interpolation behavior vs dtype mismatch in bfloat16." + "reason": "Different failures: image interpolation/resolution handling vs bfloat16 dtype mismatch." }, { - "left": "issue:44077", - "right": "issue:44625", + "left": "issue:45759", + "right": "issue:45859", "accept": false, - "reason": "Unrelated config issues: optional post_init validation vs num_labels propagation." + "reason": "Different code paths: config unwrapping for composite Qwen models vs CPU-offload incompatibility in rotary embedding." }, { - "left": "issue:45759", - "right": "issue:45859", + "left": "issue:44077", + "right": "issue:44625", "accept": false, - "reason": "Different code paths: composite config unwrapping vs CPU-offload incompatibility in rotary embedding." + "reason": "Unrelated bugs: optional `post_init` handling vs `num_labels` propagation in Qwen3.5." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Both tokenizer-related, but one is runtime post-processor updates and the other is save/load losing custom token IDs." + "reason": "Both are tokenizer-related, but one is runtime special-token/post-processor updates and the other is save/load token-ID loss." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "Flash-attention sinkless-model crash is unrelated to Qwen config unwrapping." + "reason": "Different components: flash-attention sink handling vs multimodal config unwrapping." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "AddedToken keyword conflict is not the same as failing to load a specific model." + "reason": "AddedToken constructor error is unrelated to the model-loading failure for ScandiBERT." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "Different regressions: LoRA warning about tie_word_embeddings vs incorrect visual encoder keys on save_pretrained." + "reason": "One is a tie-word-embeddings warning during LoRA fine-tuning; the other is incorrect visual encoder key saving." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Loading a specific model failing is not the same bug as PIL image processors requiring torchvision." + "reason": "Different problems: a specific model load regression vs PIL image processors wrongly requiring torchvision." }, { "left": "issue:44918", "right": "issue:44936", "accept": false, - "reason": "Different trainer failures: embedding unpacking in SFT vs evaluate() after train()." + "reason": "Different training/eval failures with different triggers and stacks." }, { - "left": "issue:43782", - "right": "issue:45381", + "left": "issue:39692", + "right": "issue:44610", "accept": false, - "reason": "Different multimodal bugs: weight_only loading error vs wrong video vision_position_ids." + "reason": "Doc-example errors for SigLIP2 vs an OmDet-Turbo processor size mismatch." }, { - "left": "issue:39692", - "right": "issue:44610", + "left": "issue:43782", + "right": "issue:45381", "accept": false, - "reason": "Both mention processor/model mismatch, but they involve different models and different failures." + "reason": "Both involve Qwen VL/video, but the reported bugs differ: weight_only loading vs video position IDs." }, { "left": "issue:43746", "right": "issue:46050", "accept": false, - "reason": "Local checkpoint loading with PEFT adapters is not the same as quantized-model save/load corruption." + "reason": "Local PEFT checkpoint loading issue vs quantized-model roundtrip corruption after save/load." }, { "left": "issue:42886", "right": "issue:44843", "accept": true, - "reason": "Same offline tokenizer regression: HF_HUB_OFFLINE breaks tokenizer loading via the unconditional model_info call in the Mistral regex path." + "reason": "Same offline tokenizer regression: both point to `AutoTokenizer.from_pretrained` / `_patch_mistral_regex` breaking when `HF_HUB_OFFLINE=1`; 44843 gives the exact root cause." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "Apex RMSNorm memory leak and missing supports_gradient_checkpointing are unrelated bugs." + "reason": "Completely different issues: apex RMSNorm memory leak vs missing gradient-checkpointing flag." }, { - "left": "issue:43381", - "right": "issue:43844", + "left": "issue:43421", + "right": "issue:43653", "accept": false, - "reason": "Eval-mode checkpointing restriction is a different issue from abnormal gradient growth under ZeRO-3." + "reason": "Both touch special tokens, but one is runtime post-processor updates and the other is BigBird mask-token registration/decoding." }, { - "left": "issue:43421", - "right": "issue:43653", + "left": "issue:43381", + "right": "issue:43844", "accept": false, - "reason": "Special-token post-processor updating and BigBird mask-token registration are related, but not the same underlying bug." + "reason": "One is an eval-mode restriction for gradient checkpointing; the other is a ZeRO-3 training instability." }, { "left": "issue:45663", "right": "issue:45923", "accept": false, - "reason": "FSDP2 shared_kv_states KeyError is unrelated to the gradient-checkpointing capability flag." + "reason": "Different model/runtime bugs: Gemma4 FSDP2 shared KV state rebuilding vs a trust_remote_code flag omission." }, { "left": "issue:45704", "right": "issue:45859", "accept": false, - "reason": "Apex RMSNorm memory leak and CPU-offload incompatibility in Qwen rotary embedding are unrelated." + "reason": "Unrelated: T5 apex RMSNorm leak vs Qwen3_5MoE rotary embedding CPU-offload incompatibility." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Auto device mapping cudaErrorAssert and an RMSNorm memory leak are different bugs." + "reason": "Different subsystems and symptoms: Qwen3 device-mapping CUDA assert vs T5 RMSNorm leak." }, { "left": "issue:43531", "right": "issue:43645", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to Jupyter custom-model initialization." + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to Jupyter custom-model initialization in Transformers 5.0." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Both are incorrect-output reports, but on different models and mechanisms." + "reason": "Both are model-quality regressions, but on different models and different failure modes." }, { "left": "issue:43421", "right": "issue:44186", "accept": false, - "reason": "Token post-processor updates and LayoutLMv2 NER/padding crashes are distinct tokenizer issues." + "reason": "Tokenizer special-token update bug is not the same as LayoutLMv2 crashing on NER padding/truncation." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "PEFT local checkpoint loading is unrelated to quantization/offloading behavior." + "reason": "PEFT checkpoint loading and quantization roundtrip differ from generic quantization offloading behavior." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both touch the Mistral-regex/tokenizer path, but the actual failures differ: backend_tokenizer AttributeError crash vs Kimi-K2.5 codec/warning regression." + "reason": "Related to Mistral regex/tokenizer handling, but one is an AttributeError crash and the other is a codec-handling regression; not the same concrete bug." }, { "left": "issue:39692", "right": "issue:43873", "accept": false, - "reason": "SigLIP2 doc/example problems are unrelated to quantization/offloading not working as expected." + "reason": "SigLIP2 doc example errors are unrelated to quantization/offloading behavior." }, { - "left": "issue:43540", - "right": "issue:46294", + "left": "issue:43596", + "right": "issue:44220", "accept": false, - "reason": "Both are video-input issues, but they affect different Qwen families and different failure points." + "reason": "Different low-level failures: ZeRO-3/BertModel indexing vs fbank feature extraction." }, { - "left": "issue:43596", - "right": "issue:44220", + "left": "issue:43540", + "right": "issue:46294", "accept": false, - "reason": "Deepspeed zero3/BertModel indexing and _torch_extract_fbank_features() are unrelated." + "reason": "Both are Qwen video-input issues, but the concrete failure is different: ValueError in processing vs chat-template token/feature mismatch." }, { "left": "issue:42915", "right": "issue:43746", "accept": false, - "reason": "Qwen3Moe FP8 configuration failure is unrelated to GraniteSpeech PEFT checkpoint loading." + "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to GraniteSpeech PEFT checkpoint loading." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Kimi-K2.5 tokenizer regression is unrelated to list_repo_files not forwarding revision." + "reason": "Mistral-regex/codec regression is unrelated to `list_repo_files` not forwarding `revision`." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Void segmentation map processing and a GenerationMixin import error are completely different." + "reason": "Image segmentation-map processing and GenerationMixin import errors are unrelated." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The clustering is mostly grounded, and 42886\u219444843 is a defensible duplicate link. But the accept on 45081\u219445356 is too strong for the packet: the titles show related Mistral/fix_mistral_regex code paths, yet one is a backend_tokenizer AttributeError crash and the other is a Kimi-K2.5 codec/warning regression. That is similar enough to group softly, but not enough evidence to call it a true duplicate with confidence. The overall summary should stay more conservative." + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the set as mostly heterogeneous and identifies issue:44843/issue:42886 as the only clearly duplicate-style pair. The soft-edge verdicts are generally conservative and the accepted pair is well-supported by the titles and stated root cause. No PR-mergeability concerns apply here." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -5024,7 +5143,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:30064", "issue:36331", @@ -5069,6 +5188,7 @@ "issue:45183", "issue:45381", "issue:45479", + "issue:45584", "issue:45685", "issue:45834", "issue:45987", @@ -5079,227 +5199,234 @@ "issue:43653|issue:43720", "issue:44220|issue:44661", "issue:43550|issue:45072", - "issue:43540|issue:44008", "issue:43295|issue:45834", + "issue:43540|issue:44008", "issue:43618|issue:43697", "issue:44062|issue:44488", "issue:43550|issue:43819", "issue:45183|issue:45479", "issue:44186|issue:44898", + "issue:43531|issue:43643", + "issue:43653|issue:43819", "issue:43322|issue:45685", "issue:43716|issue:44008", - "issue:43653|issue:43819", - "issue:43531|issue:43643", - "issue:43388|issue:46032", "issue:43335|issue:43756", - "issue:43950|issue:45987", + "issue:43388|issue:46032", "issue:43643|issue:44877", + "issue:43950|issue:45987", "issue:44843|issue:45081", "issue:30064|issue:36331", "issue:44448|issue:44671", "issue:41720|issue:44991", - "issue:43540|issue:45381", - "issue:43421|issue:44811", "issue:44821|issue:45084", + "issue:43421|issue:44811", "issue:43577|issue:43818", - "issue:39692|issue:45834" + "issue:43540|issue:45381", + "issue:39692|issue:45834", + "issue:45584|issue:45834" ], - "prepared_review_unit_hash": "6cf40e4c047972f2bde3ad60fcac2c15679e27eb684011772e0d36491a59fe25", + "prepared_review_unit_hash": "ec3da15150cec9a46618d9e042ecd6a4fa3f9e01b02e5aac0ef096c609213cff", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35960, - "estimated_input_tokens": 8990, - "estimated_eval_tokens": 18236 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36828, + "estimated_input_tokens": 9207, + "estimated_eval_tokens": 18670 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab-bag of unrelated bug reports spanning tokenizers, multimodal models, cache/dtype issues, and model-specific regressions. None of the soft pairs look like the same underlying bug or change, so I would not merge any of them into a duplicate cluster.", - "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue fits this cluster: the items target different models, different code paths, and different failure modes.", + "summary": "This cluster is mostly a loose collection of unrelated Transformers bugs that only overlap lexically around tokenizers/processors, loading, and model-specific regressions. The soft edges look like name-similarity rather than true duplicate bugs, so I would not merge any of them.", + "confidence": 0.32, + "canonical_issue_reason": "Issue 44843 is the most central representative of the tokenizer/loading regression subset: it targets a concrete shared code path (_patch_mistral_regex / AutoTokenizer loading), has nearby similar reports, and is broader than the more specific follow-ups.", "canonical_pr_reason": null, - "best_issue_reason": "No global best issue; the reports are heterogeneous and none is a clear umbrella for the rest.", + "best_issue_reason": "Issue 44843 is the best single issue representative because it sits at the center of the tokenizer-loading/Mistral-regex cluster and is the most generally actionable code-path bug among the items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "BitNet packed-weight loading during accelerate vs DAC latent/STE mismatch; different models and different code paths." + "reason": "Unrelated bugs: BitNet accelerate loading vs DAC STE/forward mismatch. Same broad ML area only." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "BigBirdTokenizer special-token decode bug is unrelated to BitNet packed-weight unpacking in accelerate loading." + "reason": "Tokenizer special-token decoding bug vs BitNet packed-weight loading bug; no shared code path." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio fbank feature extraction vs add-new-model-like tokenizer mapping; unrelated failures." + "reason": "Audio feature extraction issue vs tokenizer-mapping add-new-model-like failure; different subsystems." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "One is Bamba torch.compile with SDPA, the other is SwitchTransformers/TimmWrapperModel bfloat16 dtype mismatch; not the same bug." + "reason": "Both mention dtype/SDPA-ish model behavior, but they are model-specific and different failures." }, { - "left": "issue:43540", - "right": "issue:44008", + "left": "issue:43295", + "right": "issue:45834", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError and Gemma3n audio_mel_mask name collision are different model-specific issues." + "reason": "Processor/tokenizer regression on v4.57.5 vs Kosmos2.5 OCR index error; not the same bug." }, { - "left": "issue:43295", - "right": "issue:45834", + "left": "issue:43540", + "right": "issue:44008", "accept": false, - "reason": "processor.tokenizer regression in v4.57.5 is unrelated to Kosmos2.5 OCR index handling." + "reason": "Video-input processing error in Qwen3OmniMoe vs Gemma3n variable name collision; distinct code paths." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "Missing CLIPOutput attentions and RTDetrV2 output drift in v5 are different regressions." + "reason": "CLIPOutput attribute regression vs RTDetrV2 output-difference regression; unrelated model behaviors." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken special-arg duplication is unrelated to loading cjvt/sleng-bert." + "reason": "AddedToken keyword conflict vs tokenizer loading failure for a specific model; different root causes." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Bamba SDPA/torch.compile failure and DAC from_latents mismatch are unrelated." + "reason": "Bamba torch.compile/SDPA failure vs DAC latent/forward mismatch; no common concrete bug." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Input validation XOR messaging is unrelated to single-label-classification zero-loss behavior." + "reason": "Input validation messaging bug vs degenerate loss for single-label classification; different issues." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "LayoutLMv2Tokenizer NER/padding crash and Perceiver resize/interpolate issue are different subsystems." - }, - { - "left": "issue:43322", - "right": "issue:45685", - "accept": false, - "reason": "Llava Next loading segfault and MPS histogram int error do not share a concrete code-path." + "reason": "LayoutLMv2 tokenizer padding/NER crash vs Perceiver non-default resolution failure; not duplicates." }, { - "left": "issue:43716", - "right": "issue:44008", + "left": "issue:43531", + "right": "issue:43643", "accept": false, - "reason": "Mistral-3 image-preprocessor dtype mismatch and Gemma3n variable collision are separate bugs." + "reason": "Qwen3-MoE sliding-window behavior vs AutoConfig trust_remote_code field loss; unrelated regressions." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "BigBirdTokenizer mask-token decode bug and DAC latent/forward mismatch are unrelated." + "reason": "BigBird special-token decode bug vs DAC latent/forward mismatch; no overlap in bug mechanics." }, { - "left": "issue:43531", - "right": "issue:43643", + "left": "issue:43322", + "right": "issue:45685", "accept": false, - "reason": "Qwen3-MoE sliding_window bug and trust_remote_code missing-fields regression are different issues." + "reason": "Llava Next loading segfault vs MPS histogram int error in moe code; unrelated symptoms and paths." }, { - "left": "issue:43388", - "right": "issue:46032", + "left": "issue:43716", + "right": "issue:44008", "accept": false, - "reason": "gather_for_metrics tuple-label trimming and Mamba2Mixer use_cache corruption are not the same underlying defect." + "reason": "Mistral-3 dtype mismatch vs Gemma3n audio mask attribute collision; different model-specific bugs." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Sparse-layer creation with zero layers and Smollm3 RoPE-layer dropping are distinct model-configuration bugs." + "reason": "SwitchTransformers sparse-layer construction bug vs Smollm3 RoPE layer count mismatch; not the same defect." }, { - "left": "issue:43950", - "right": "issue:45987", + "left": "issue:43388", + "right": "issue:46032", "accept": false, - "reason": "Non-persistent buffer corruption on from_pretrained is unrelated to StaticCache.get_seq_length returning a tensor." + "reason": "Metrics gathering truncation bug vs Mamba2 use_cache incorrect results; distinct runtime paths." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "Missing fields with trust_remote_code and strict config blocking granite_speech are separate config-loading problems." + "reason": "AutoConfig trust_remote_code missing fields vs strict config blocking granite_speech; both loading-related but not same bug." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "Non-persistent buffer corruption on from_pretrained vs StaticCache.get_seq_length returning wrong type; unrelated cache/model-loading bugs." }, { "left": "issue:44843", "right": "issue:45081", "accept": false, - "reason": "Both touch Mistral regex patching, but one is offline model_info calls and the other is backend_tokenizer attribute access; different failure points and fixes." + "reason": "Both involve _patch_mistral_regex, but one is offline/network access and the other is an attribute error on tokenizer wrapping; not the same concrete bug." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Void segmentation maps in Image Processor vs CustomTrainer.compute_loss kwarg change are unrelated." + "reason": "Image processor void segmentation maps vs CustomTrainer compute_loss signature mismatch; completely different areas." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Both mention v4/v5 output differences, but they concern different models and different prediction paths." + "reason": "Pegasus output drift across versions vs CamemBERT masked LM prediction errors; different model regressions." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "Qwen3 auto-device-map CUDA assert and est-roberta tokenizer loading failure are unrelated." + "reason": "Qwen3 auto device mapping CUDA assert vs tokenizer loading failure for est-roberta; not the same bug." }, { - "left": "issue:43540", - "right": "issue:45381", + "left": "issue:44821", + "right": "issue:45084", "accept": false, - "reason": "Qwen3OmniMoe video-input ValueError and Qwen2.5-VL vision_position_ids bug are model-specific and not the same fix." + "reason": "AutoImageProcessor URL loading vs template-node compilation error; no shared code path." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Runtime post-processor updates for special tokens and Whisper batch_decode skip_special_tokens behavior are different tokenizer features." + "reason": "Tokenizer backend post-processor update bug vs Whisper batch_decode skip_special_tokens behavior; related to tokenizers, but different functionality." }, { - "left": "issue:44821", - "right": "issue:45084", + "left": "issue:43577", + "right": "issue:43818", "accept": false, - "reason": "AutoImageProcessor URL loading and template-node compilation errors are unrelated." + "reason": "Blip2 dtype propagation bug vs Video-LLaVA temporal-attention/weight-sharing bug; unrelated." }, { - "left": "issue:43577", - "right": "issue:43818", + "left": "issue:43540", + "right": "issue:45381", "accept": false, - "reason": "BLIP2 dtype propagation and Video-LLaVA temporal-attention/weight-sharing issues are different model families and bugs." + "reason": "Qwen3OmniMoe video processing error vs Qwen2.5-VL video vision_position_ids issue; both video-related, but distinct model/code-path bugs." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "SigLIP2 doc example errors and Kosmos2.5 OCR index error are unrelated." + "reason": "SigLIP2 doc example errors vs Kosmos2.5 OCR index error; documentation/model usage issue vs runtime model bug." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription generation bug vs Kosmos2.5 long OCR input index error; different models and failures." } ] }, @@ -5321,7 +5448,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:28218", "issue:29159", @@ -5334,6 +5461,7 @@ "issue:43582", "issue:43606", "issue:43644", + "issue:43653", "issue:43720", "issue:43749", "issue:43782", @@ -5353,6 +5481,7 @@ "issue:44671", "issue:44805", "issue:44857", + "issue:44912", "issue:44936", "issue:44977", "issue:45003", @@ -5362,7 +5491,6 @@ "issue:45325", "issue:45372", "issue:45406", - "issue:45584", "issue:45701", "issue:45834", "issue:45987", @@ -5373,237 +5501,240 @@ ], "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:34634|issue:36010", "issue:44805|issue:44936", + "issue:34634|issue:36010", "issue:43577|issue:43582", "issue:44485|issue:44509", "issue:45987|issue:46192", "issue:28218|issue:29159", "issue:44561|issue:45020", - "issue:43720|issue:45072", "issue:44610|issue:45072", + "issue:43720|issue:45072", "issue:44671|issue:45701", - "issue:43782|issue:45325", "issue:43825|issue:43827", + "issue:43782|issue:45325", "issue:43582|issue:44857", - "issue:43906|issue:46209", "issue:43749|issue:44291", "issue:45081|issue:45406", "issue:42175|issue:45081", + "issue:43906|issue:46209", "issue:45372|issue:46326", "issue:43262|issue:43819", "issue:34634|issue:36331", - "issue:45003|issue:46285" + "issue:45003|issue:46285", + "issue:43653|issue:44912" ], - "prepared_review_unit_hash": "f7b90a0e6c326257821ca03dfbb61cd3df486e89da0fe5d01c7f4fa68fc29141", + "prepared_review_unit_hash": "e6b69807cf9287ac92cefb3de1369dffe1dc1bdc9eb75eb6eb96530274c96b14", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35685, - "estimated_input_tokens": 8922, - "estimated_eval_tokens": 18100 + "serialized_chars": 36290, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most items are unrelated point bugs across tokenizers, model loading, training, vision/audio processors, and docs. All soft-similarity pairs look like false positives, so no merges are warranted.", - "confidence": 0.79, - "canonical_issue_reason": "issue:45020 is the broadest open-ended regression report in the set ('Recent transformers versions break models using `remote_code`'), so it is the closest thing to an umbrella issue; however, the cluster overall is mixed and it does not actually subsume the other reports.", + "summary": "Mostly heterogeneous issue cluster with several weak thematic similarities. I found one likely true duplicate pair around `remote_code` breakage, while the tokenizer, loading, docs, and model-specific items appear unrelated enough to keep separate.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 44561 is the clearest root-cause report: it names the concrete v5 regression (`is_torch_fx_available` removal) and the affected surface (`trust_remote_code` models), which makes it the best canonical anchor for the remote_code failure mode also described more broadly in 45020.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best single representative because it is the most generic version-regression report and could attract related breakages; the other issues are narrower and largely unrelated.", + "best_issue_reason": "Issue 44561 is the most actionable and specific representative among the cluster items; it pins down the failure mechanism rather than only describing the downstream symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Different failures: Qwen3VL `weight_only=True` loading error vs Qwen3.5 flash-attention generation bug." - }, - { - "left": "issue:45584", - "right": "issue:45834", - "accept": false, - "reason": "Whisper empty-transcription bug after `align_special_tokens` is unrelated to Kosmos2.5 long-OCR index errors." + "reason": "Different model/runtime failures: Qwen3VL weight-only loading vs Qwen3.5 flash-attention generation." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "LayoutLMv2 tokenizer NER/padding crash and AutoTokenizer failing to load a different tokenizer are separate code paths." + "reason": "Both are tokenizer-loading bugs, but for different tokenizers and failure modes; not the same underlying defect." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Non-persistent buffer junk initialization is not the same as FSDP CPU RAM efficient loading being broken." + "reason": "One is non-persistent buffer initialization junk; the other is FSDP CPU RAM efficient loading. Related area, different bug." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "CPU-offload device mismatch for Bark is a different bug from dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Different problems: Bark CPU offload device mismatch vs dtype mismatch in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "Qwen VL video-input regression and Kosmos2.5 OCR index error are unrelated model-specific failures." + "reason": "Unrelated model families and symptoms: Qwen video regression vs Kosmos2.5 OCR index error." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "Wrong pipeline error messaging about removed translation tasks is not the same as SigLIP2 producing bad outputs." + "reason": "Both mention pipeline(), but one is an error-message issue and the other is a model-output correctness bug." }, { - "left": "issue:34634", - "right": "issue:36010", + "left": "issue:44805", + "right": "issue:44936", "accept": false, - "reason": "BarkProcessor `voice_preset` issue and `GenerationMixin` import failure are unrelated." + "reason": "Generic tensor shape/index error vs trainer evaluate-after-train failure; no shared code-path evidence." }, { - "left": "issue:44805", - "right": "issue:44936", + "left": "issue:34634", + "right": "issue:36010", "accept": false, - "reason": "Mask shape/indexing error is a different failure than `trainer.evaluate()` breaking after `train()`." + "reason": "BarkProcessor voice_preset bug and GenerationMixin import error are unrelated." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "BLIP2 dtype staying float32 and Apple Silicon warmup `TypeError` are unrelated problems." + "reason": "BLIP2 dtype propagation and Apple Silicon allocator warmup TypeError are different subsystems." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "GLM-5 RoPE discussion and docs still mentioning removed pipeline tasks are separate issues." + "reason": "GLM-5 RoPE discussion and v5 pipeline docs drift are distinct issues." }, { "left": "issue:45987", "right": "issue:46192", "accept": false, - "reason": "`StaticCache.get_seq_length()` returning a tensor and `return_tensors` being ignored are different API bugs." + "reason": "Cache sequence-length return type and return_tensors handling are separate API bugs." }, { "left": "issue:28218", "right": "issue:29159", "accept": false, - "reason": "Extra space after added token and slow/fast tokenizer inconsistency are related only at a very high level, not the same concrete bug." + "reason": "Both are tokenizer-related, but the titles point to different symptoms; not enough evidence they are the same whitespace bug." }, { "left": "issue:44561", "right": "issue:45020", - "accept": false, - "reason": "Both mention remote-code breakage, but one is specifically `is_torch_fx_available` removal and the other is a broader version-regression report." + "accept": true, + "reason": "Both describe the same remote_code regression in recent Transformers versions; 44561 gives the concrete cause (`is_torch_fx_available` removal) and 45020 the broader symptom." }, { - "left": "issue:43720", + "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is not the same as bf16 dtype mismatches." + "reason": "Processor image-size mismatch vs dtype mismatch in inference; unrelated failures." }, { - "left": "issue:44610", + "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "OmDet-Turbo processor producing 640px inputs is unrelated to SwitchTransformers/TimmWrapperModel dtype mismatch." + "reason": "BitNet packed-weight loading and dtype mismatch issues are different code paths." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "CamemBERT masked-LM prediction regression and general tokenization changes are not the same underlying bug." - }, - { - "left": "issue:43782", - "right": "issue:45325", - "accept": false, - "reason": "Qwen3VL `weight_only` load failure and Qwen2.5-VL RoPE scaling bug affect different models and code paths." + "reason": "CamemBERT masked-LM prediction regression and a general tokenization-change report are not the same bug." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "An incorrect runtime error message and stale docs references are separate artifacts, not one duplicate bug." + "reason": "Error message about removed pipeline tasks vs docs still referencing removed tasks; same theme, but not one concrete duplicate bug." }, { - "left": "issue:43582", - "right": "issue:44857", + "left": "issue:43782", + "right": "issue:45325", "accept": false, - "reason": "Apple Silicon warmup `TypeError` and LwDetrImageLoss AMP/CUDA crash are unrelated." + "reason": "Qwen3VL weight-only loading error and Qwen2.5-VL RoPE position-id scaling are separate regressions." }, { - "left": "issue:43906", - "right": "issue:46209", + "left": "issue:43582", + "right": "issue:44857", "accept": false, - "reason": "An isolated reproduction of an earlier issue and an FP8 B200 inference failure do not describe the same bug." + "reason": "Allocator warmup TypeError and AMP/CUDA loss crash are unrelated." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "FSDP CPU RAM loading failure and `init_empty_weights` unexpected `_is_hf_initialized` argument are distinct initialization/loading regressions." + "reason": "FSDP CPU RAM loading breakage and init_empty_weights argument mismatch are distinct loading bugs." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Mistral tokenizer regex patch crash and Gemma4Processor missing `_tokenizer` are different failures." + "reason": "Mistral tokenizer patch crash and Gemma4 processor `_tokenizer` attribute crash are different failures." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Package install backend omission and Mistral tokenizer regex crash are unrelated." + "reason": "Backend-install packaging issue vs tokenizer regex patch crash; no shared underlying bug." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated reproduction of an older issue is not the same as the B200 FP8 inference failure." }, { "left": "issue:45372", "right": "issue:46326", "accept": false, - "reason": "`ReasoningEffort` import failure in Gemma 4 processor loading and `continue_final_message` behavior for mistral_common tokenizers are different bugs." + "reason": "Both involve `mistral_common`, but one is an import failure during Gemma 4 loading and the other is a tokenizer behavior bug." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and DAC `from_latents` mismatch are unrelated." + "reason": "Audio chat-template sampling-rate default and DAC latent reconstruction mismatch are unrelated." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "BarkProcessor voice preset and `CustomTrainer.compute_loss()` signature mismatch are unrelated." + "reason": "BarkProcessor voice preset and custom trainer `compute_loss` signature mismatch are unrelated." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Unsafe `sys.modules` access in `modeling_utils` and CLIPTextModel structure inconsistency are different code-path regressions." + "reason": "Unsafe `sys.modules` access and CLIP text-model structure inconsistency are different regressions." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug and MXFP4 quantization fallback are unrelated." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The cluster is indeed heterogeneous, and the lone accepted pair (`44561`/`45020`) is a plausible same-family `remote_code` regression grounded in the titles. The other soft-edge rejections are framed cautiously enough from the available evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -5618,7 +5749,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36010", "issue:36331", @@ -5637,7 +5768,6 @@ "issue:43638", "issue:43645", "issue:43650", - "issue:43653", "issue:43696", "issue:43697", "issue:43720", @@ -5670,24 +5800,23 @@ "issue:46424" ], "soft_pairs": [ - "issue:43653|issue:44912", "issue:43526|issue:43697", "issue:43742|issue:45020", "issue:44060|issue:44368", "issue:45084|issue:45538", "issue:43526|issue:44554", - "issue:43650|issue:43906", "issue:43329|issue:45600", "issue:44589|issue:45042", + "issue:43650|issue:43906", "issue:43452|issue:44991", - "issue:43746|issue:44811", "issue:43638|issue:44936", + "issue:43746|issue:44811", "issue:43550|issue:44912", "issue:44387|issue:44617", "issue:43819|issue:43927", + "issue:36010|issue:43645", "issue:43257|issue:43329", "issue:46421|issue:46424", - "issue:36010|issue:43645", "issue:43854|issue:44977", "issue:43720|issue:44610", "issue:38175|issue:44610", @@ -5695,213 +5824,201 @@ "issue:36010|issue:36331", "issue:43531|issue:45042", "issue:45459|issue:45803", - "issue:43454|issue:43577", "issue:42490|issue:43257", + "issue:43454|issue:43577", "issue:43262|issue:44811", "issue:43257|issue:43696" ], - "prepared_review_unit_hash": "102ba651a4c4e03f09d06d01e4cbef642b08cfa43b40d1bef95b3f2780e90c84", + "prepared_review_unit_hash": "a4115a465f1614ccf4105a96e1228ef39dfdbfa317eeb0c7277b417e7b4c1498", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36606, - "estimated_input_tokens": 9152, - "estimated_eval_tokens": 18560 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35717, + "estimated_input_tokens": 8930, + "estimated_eval_tokens": 18116 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set does not form one duplicate cluster; it contains many unrelated issues across tokenizers, loading, quantization, cache, and multimodal processors. All soft-similarity pairs look like false positives and should be kept separate.", - "confidence": 0.97, - "canonical_issue_reason": "No single issue is a good canonical representative because the items are thematically unrelated rather than duplicates of one underlying bug.", + "summary": "These are mostly unrelated issue reports across different models, loaders, processors, and training paths. None of the soft pairs look like safe duplicate matches, and there is no single clear canonical issue or PR for the set.", + "confidence": 0.95, + "canonical_issue_reason": "No canonical issue is safe here because the items span heterogeneous bugs rather than one underlying defect cluster.", "canonical_pr_reason": null, - "best_issue_reason": "None of the issues is suitable as a global canonical issue for this set; the strongest matches still describe different code paths or different products/models.", + "best_issue_reason": "No issue in this list is a good global representative; the closest-looking pairs still differ in concrete code path and symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43653", - "right": "issue:44912", - "accept": false, - "reason": "Different bugs: BigBirdTokenizer special-token registration vs GPT-OSS/gguf quantization loading." - }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "Different subsystems and symptoms: BEiT label reduction vs RTDetrV2 output drift." + "reason": "Different models and failures: BEiT label reduction vs RTDetrV2 output divergence." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "One is a MobileLLM key error; the other is a broad remote_code regression affecting many models." + "reason": "Specific MobileLLM key error is not the same as a broad remote_code loading regression." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Both mention Qwen warnings, but one is a tied-weights bug and the other is a fine-tuning config warning; not the same defect." + "reason": "Both mention Qwen tied-weight warnings, but they describe different model/config paths and symptoms." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Unrelated: template compilation error vs CLIPTokenizer max length constant." + "reason": "Unrelated: a compile/template-node error versus a tokenizer max_length constant issue." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Unrelated vision preprocessing bug vs MPS attention correctness issue." - }, - { - "left": "issue:43650", - "right": "issue:43906", - "accept": false, - "reason": "No meaningful relation; placeholder issue text vs a reproduction of a specific upstream bug." + "reason": "Image label reduction bug and MPS attention correctness bug are unrelated." }, { "left": "issue:43329", "right": "issue:45600", "accept": false, - "reason": "Different areas: multimodal token counting vs auto_mappings/config cleanup." + "reason": "Multimodal token counting bug is unrelated to stale auto_mappings config references." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Storage type import error is unrelated to PIL/image-processor torchvision dependency handling." + "reason": "Float8 storage lookup failure is unrelated to an unwanted torchvision dependency." }, { - "left": "issue:43452", - "right": "issue:44991", + "left": "issue:43650", + "right": "issue:43906", "accept": false, - "reason": "Both involve loading, but one is gguf_file support and the other is a tokenizer regression for a specific model; not the same code path." + "reason": "A placeholder issue title is not the same as a reproduction of another bug." }, { - "left": "issue:43746", - "right": "issue:44811", + "left": "issue:43452", + "right": "issue:44991", "accept": false, - "reason": "Different models and APIs: GraniteSpeech PEFT checkpoint loading vs Whisper batch_decode token handling." + "reason": "Both are loading-related, but one is gguf_file/from_pretrained handling and the other is a specific tokenizer load failure." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Different training/evaluation failures with different root causes; not a duplicate." + "reason": "Different trainer failures: Deepspeed ZeRO3 BERT indexing vs evaluate-after-train behavior." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT checkpoint loading and Whisper batch_decode token filtering are unrelated code paths." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "Different model/runtime issues: torch.compile+SDPA vs gguf/MXFP4 quantized loading." + "reason": "torch.compile/SDPA failure and MXFP4 quantization fallback are different issues." }, { "left": "issue:44387", "right": "issue:44617", "accept": false, - "reason": "Both can OOM, but one is a Transformers 5.x int4 memory regression and the other is a model-specific runtime OOM." + "reason": "Both mention OOM, but one is int4 quantization memory regression and the other is a model-specific video OOM." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "Different failures: DAC latent reconstruction mismatch vs DiaConfig save/load token ID loss." + "reason": "DAC latent mismatch and DiaConfig token-ID persistence are unrelated bugs." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Import path breakage is not the same as Jupyter notebook custom-model initialization failure." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Completely different areas: MoE weight conversion/loading vs multimodal token counting video path." + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to multimodal token counting." }, { "left": "issue:46421", "right": "issue:46424", "accept": false, - "reason": "Related cache area, but one is lazy init causing recompilation and the other is max_cache_length handling for quantized cache; not the same bug." - }, - { - "left": "issue:36010", - "right": "issue:43645", - "accept": false, - "reason": "Import error for GenerationMixin is unrelated to Jupyter custom-model initialization in Transformers 5.0.0." + "reason": "Both concern cache behavior, but one is recompilation from lazy init and the other is max_cache_length handling for quantized cache only." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Different model-specific generation/load problems; no shared concrete code-path defect." + "reason": "Different models and symptoms: GLM-4.7-Flash loading tests vs Qwen3.5 flash-attention generation." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Unrelated: packed-weight unloading in accelerate vs image processor input-size mismatch." + "reason": "BitNet packed-weight loading and OmDet-Turbo image sizing are unrelated." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "Different model families and failure modes; zero probabilities vs wrong processor size." + "reason": "SigLIP2 zero probabilities and OmDet-Turbo processor/model size mismatch are unrelated." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Documentation/example errors are unrelated to the gguf/MXFP4 loading regression." + "reason": "Documentation example errors for SigLIP2 do not match the MXFP4 quantization loading bug." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Different APIs and failures: import path breakage vs Trainer custom-loss signature change." + "reason": "ImportError for GenerationMixin and unexpected compute_loss kwarg are different API breakages." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding window behavior is unrelated to PIL image-processor torchvision requirements." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the torchvision requirement regression." }, { "left": "issue:45459", "right": "issue:45803", "accept": false, - "reason": "Both are exception-handling issues, but they affect different code paths and different error types; not the same bug." - }, - { - "left": "issue:43454", - "right": "issue:43577", - "accept": false, - "reason": "Different model-loading/data-type problems: lm_head tying vs dtype propagation to qformer." + "reason": "Both involve exception handling, but they are different functions and different failure mechanisms." }, { "left": "issue:42490", "right": "issue:43257", "accept": false, - "reason": "Trailing-slash path warning is unrelated to Qwen3 MOE weight conversion with accelerate+deepspeed." + "reason": "Trailing-slash local path warning is unrelated to Qwen3 MoE weight conversion." }, { - "left": "issue:43262", - "right": "issue:44811", + "left": "issue:43454", + "right": "issue:43577", "accept": false, - "reason": "Different processor APIs and symptoms: audio chat-template sample-rate default vs Whisper decode token filtering." + "reason": "AyaVision weight-tying failure and BLIP2 dtype casting are distinct bugs." }, { - "left": "issue:43257", - "right": "issue:43696", + "left": "issue:43262", + "right": "issue:44811", "accept": false, - "reason": "MoE weight conversion/loading is unrelated to GPT-oss CUDA OOM." + "reason": "Audio chat-template sampling-rate defaults are unrelated to Whisper batch_decode skip_special_tokens handling." } ] }, @@ -5923,7 +6040,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:24643", "issue:28282", @@ -5978,263 +6095,266 @@ "issue:43643|issue:45759", "issue:43329|issue:44479", "issue:43329|issue:43450", - "issue:43606|issue:45859", "issue:34567|issue:40990", + "issue:43606|issue:45859", "issue:43749|issue:44756", - "issue:44568|issue:45356", "issue:29159|issue:34567", + "issue:44568|issue:45356", "issue:43257|issue:43931", "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:43697|issue:45357", "issue:42915|issue:45538", "issue:44849|issue:45701", "issue:44568|issue:45488", - "issue:41720|issue:43540", "issue:33357|issue:34567", + "issue:41720|issue:43540", "issue:42175|issue:43824", - "issue:44479|issue:45381", "issue:43317|issue:45538", - "issue:43540|issue:44479", + "issue:44479|issue:45381", "issue:44062|issue:44991", - "issue:43540|issue:44560", + "issue:43540|issue:44479", "issue:28282|issue:34689", + "issue:43540|issue:44560", "issue:45137|issue:45405", - "issue:24643|issue:29159", "issue:43742|issue:44291", + "issue:24643|issue:29159", "issue:44987|issue:45907" ], - "prepared_review_unit_hash": "fa24daff671f1bba2faeecf50b14c9b9edb5f34004add87710825838738a0d4c", + "prepared_review_unit_hash": "043bb84f30b681794768465644b9dad79f5991e8ee639289f7066ad456c960c2", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37328, - "estimated_input_tokens": 9332, - "estimated_eval_tokens": 18920 + "serialized_chars": 37330, + "estimated_input_tokens": 9333, + "estimated_eval_tokens": 18922 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes unrelated tokenizer, video/VLM, loading, offload, and training-state bugs. I found no soft pair that clearly represents the same underlying fix, so all pairs should be kept separate.", - "confidence": 0.24, - "canonical_issue_reason": "Issue 45701 is the broadest, most umbrella-like report in the set and has inbound references, but there is no truly strong canonical duplicate cluster here because the topics are too mixed.", + "summary": "Mostly a heterogeneous set of closed Transformer bugs spanning tokenizer regressions, model-loading failures, multimodal/video processing, offload, and quantization. I found one likely true duplicate-style soft edge in the Qwen video regression area; the rest are distinct issues.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45701 is the broadest umbrella item in the cluster and best matches the tokenization-regression theme shared by several nearby reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45701 is the closest thing to a representative issue for the cluster because it is the most general regression report and could absorb adjacent tokenization/version-related complaints, though it does not cover the whole set.", + "best_issue_reason": "issue:45701 is the most representative issue overall because it is broad, user-facing, and central to the tokenization/version-regression subset.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43643", "right": "issue:45759", "accept": false, - "reason": "Both are config/model-construction complaints, but one is AutoConfig trust_remote_code metadata loss and the other is AutoModelForCausalLM.from_config not unwrapping text_config. Different code paths and failure modes." + "reason": "Different APIs and failures: missing fields from AutoConfig with trust_remote_code vs text_config unwrapping in AutoModelForCausalLM.from_config." }, { "left": "issue:43329", "right": "issue:44479", "accept": false, - "reason": "Both involve video/VLM regressions, but 43329 is an undefined-name bug in a helper path while 44479 is a broader v5.3.0 video-input regression across several Qwen variants. Not the same concrete bug." + "reason": "Both are video-related, but one is an undefined-name bug in _get_num_multimodal_tokens and the other is a broader 5.3.0 regression across Qwen video models." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "One is token counting with undefined video-branch symbols; the other is batched video processor shape handling. Related subsystem, but different underlying defects." + "reason": "Undefined helper names in video token counting vs batched video processor shape errors; different code paths." }, { - "left": "issue:43606", - "right": "issue:45859", + "left": "issue:34567", + "right": "issue:40990", "accept": false, - "reason": "Both mention CPU offload, but they hit different models and different code paths: Bark device mismatch versus Qwen3.5 MoE rotary embedding compatibility." + "reason": "Unrelated topics: TrainerState token counter update vs high perplexity on gpt-oss-20b." }, { - "left": "issue:34567", - "right": "issue:40990", + "left": "issue:43606", + "right": "issue:45859", "accept": false, - "reason": "TrainerState token accounting is unrelated to perplexity on gpt-oss-20b." + "reason": "Both involve offload, but one is a bark-small device mismatch and the other is a Qwen rotary-embedding CPU-offload incompatibility." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Both are memory/loading related, but FSDP CPU RAM-efficient loading and Strix Halo mmap OOM are distinct problems." + "reason": "Different problems: FSDP CPU RAM efficient loading vs mmap OOM on Strix Halo." }, { - "left": "issue:44568", - "right": "issue:45356", + "left": "issue:29159", + "right": "issue:34567", "accept": false, - "reason": "Both are tokenizer regressions, but one is add_special_tokens not adding BOS/EOS and the other is codec/warning behavior for Kimi-K2.5. Different bugs." + "reason": "Tokenizer inconsistency vs TrainerState token-count tracking; not the same bug." }, { - "left": "issue:29159", - "right": "issue:34567", + "left": "issue:44568", + "right": "issue:45356", "accept": false, - "reason": "Tokenizer slow/fast inconsistency is unrelated to TrainerState num_input_tokens_seen not updating." + "reason": "Separate tokenizer regressions affecting different models and failure modes." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "Both are Qwen3 loading issues, but one is deepspeed/accelerate MoE weight conversion and the other is a Qwen3-VL shape mismatch. Not the same fix." + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed vs Qwen3-VL weight-shape mismatch on load." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "Qwen3Moe FP8 failure and MobileLLM key error are unrelated loading failures." + "reason": "Different load-time failures on different models: FineGrainedFP8Config vs a MobileLLM key error." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized and PIL backend torchvision dependency are different code paths with no shared concrete bug." + "reason": "Unexpected _is_hf_initialized argument during init_empty_weights is unrelated to PIL backend requiring torchvision." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "FP8 Qwen3Moe loading and AutoImageProcessor-from-URL are unrelated." + "reason": "Unrelated: Qwen3Moe FP8 config failure vs AutoImageProcessor URL loading." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Loading physical-intelligence/fast on v5.1.0 and CLIPTokenizer model_max_length are separate issues." + "reason": "Different issues: loading physical-intelligence/fast vs CLIPTokenizer model_max_length being 10**30." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "gguf_file tokenizer/model loading breakage and AppleSilicon caching_allocator_warmup TypeError are unrelated." + "reason": "gguf_file loading breakage is unrelated to the Apple Silicon caching_allocator_warmup TypeError." }, { - "left": "issue:43697", - "right": "issue:45357", + "left": "issue:43927", + "right": "issue:44912", "accept": false, - "reason": "RTDetrV2 output changes and Qwen3.5 save_pretrained visual-key regression are different model families and different failure modes." + "reason": "Different failure modes: DiaConfig custom token IDs persistence vs MXFP4 quantization fallback on gpt-oss-20b." }, { - "left": "issue:43927", - "right": "issue:44912", + "left": "issue:43697", + "right": "issue:45357", "accept": false, - "reason": "DiaConfig save/load token ID loss is unrelated to MXFP4 quantization fallback on git-oss-20b." + "reason": "RTDetrV2 output mismatch and Qwen3.5 save_pretrained key regression are separate bugs." }, { "left": "issue:42915", "right": "issue:45538", "accept": false, - "reason": "Qwen3Moe FP8 loading and CLIPTokenizer max length are not the same bug." + "reason": "No overlap beyond both being tokenizer/model issues; one is Qwen3Moe FP8, the other CLIPTokenizer max length." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 output_hidden_states regression is a model-forward bug, while 45701 is a generic tokenization/version-change report." + "reason": "Qwen3.5 output_hidden_states bug is distinct from the broad tokenization regression report." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer regressions, but one concerns special-token insertion and the other concerns overriding tokenizer.json pre-tokenizer settings. Different code paths." - }, - { - "left": "issue:41720", - "right": "issue:43540", - "accept": false, - "reason": "cudaErrorAssert on auto device mapping for Qwen3 and video-processing ValueError in Qwen3OmniMoe are distinct issues." + "reason": "Both tokenizer-related, but the BOS/EOS special-token bug and the LlamaTokenizer ByteLevel/Metaspace override are different code paths." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS bus error with community CLIP and TrainerState token tracking are unrelated." + "reason": "MacOS bus error in CLIP model loading is unrelated to TrainerState token tracking." }, { - "left": "issue:42175", - "right": "issue:43824", + "left": "issue:41720", + "right": "issue:43540", "accept": false, - "reason": "TensorFlow backend packaging and missing Qwen2.5-VL import are different concerns." + "reason": "Different Qwen failures: auto device mapping cudaErrorAssert vs video-input ValueError in Qwen3OmniMoe." }, { - "left": "issue:44479", - "right": "issue:45381", + "left": "issue:42175", + "right": "issue:43824", "accept": false, - "reason": "Both are video-input regressions, but 45381 is specifically wrong vision_position_ids in qwen2.5-vl, while 44479 is a broader multi-model regression. Too little evidence for one bug." + "reason": "Backend packaging issue for TensorFlow is unrelated to the Qwen2.5-VL import error." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "Dequantized model loading with device_map=auto/offload and CLIPTokenizer model_max_length are unrelated." + "reason": "Device-map/offload loading failure is unrelated to CLIPTokenizer max_length behavior." }, { - "left": "issue:43540", - "right": "issue:44479", - "accept": false, - "reason": "Qwen3OmniMoe video ValueError and the broader v5.3.0 video-input regression are not clearly the same underlying defect." + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "These describe the same Qwen video-input regression family around 5.3.0; 45381 is a specific qwen2.5-vl manifestation of the broader multi-model bug in 44479." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken special-argument conflict and tokenizer loading for EMBEDDIA/est-roberta are different tokenizer bugs." + "reason": "Different tokenizer-loading TypeErrors with different roots and models." }, { "left": "issue:43540", - "right": "issue:44560", + "right": "issue:44479", "accept": false, - "reason": "Both are Qwen video-related, but one is a processing ValueError and the other is a StopIteration in qwen3-vl embedding. Different concrete failures." + "reason": "Both touch Qwen video inputs, but one is a Qwen3OmniMoe processing error and the other is a broader v5.3.0 regression across several Qwen models." }, { "left": "issue:28282", "right": "issue:34689", "accept": false, - "reason": "Missing PyTorch import and Llama 3.2 Vision model loading breakage are unrelated load-time errors." + "reason": "ImportError for missing PyTorch is unrelated to Llama 3.2 90B Vision model-loading breakage." }, { - "left": "issue:45137", - "right": "issue:45405", + "left": "issue:43540", + "right": "issue:44560", "accept": false, - "reason": "DeepSpeed ZeRO3 deque underflow and a PEFT version bump are not the same issue." + "reason": "Different video bugs: a ValueError in Qwen3OmniMoe processing vs a StopIteration in Qwen3-VL embedding video handling." }, { - "left": "issue:24643", - "right": "issue:29159", + "left": "issue:45137", + "right": "issue:45405", "accept": false, - "reason": "DeepSpeed weight dimensionality error and tokenizer inconsistency are unrelated." + "reason": "DeepSpeed ZeRO3 deque underflow is unrelated to the PEFT minimum-version bump." }, { "left": "issue:43742", "right": "issue:44291", "accept": false, - "reason": "Key error loading MobileLLM and init_empty_weights unexpected argument are separate model-loading failures." + "reason": "Different loading failures: MobileLLM key error vs init_empty_weights/_is_hf_initialized TypeError." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "Training with DeepSpeed weight-shape error is unrelated to tokenizer slow/fast inconsistency." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Loading physical-intelligence/fast and list_repo_files not forwarding revision are unrelated regressions." + "reason": "Loading physical-intelligence/fast is unrelated to list_repo_files needing to forward revision." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet, and the only accepted soft edge (44479 \u2194 45381) is conservative and supported by the titles. The rejected pairs are generally reasoned at the level of distinct APIs, models, or failure modes, without overstating duplication evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -6249,7 +6369,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:27957", "issue:28282", @@ -6301,48 +6421,48 @@ "issue:46399" ], "soft_pairs": [ - "issue:43577|issue:45072", "issue:28282|issue:29127", + "issue:43577|issue:45072", "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:44737|issue:45588", "issue:44291|issue:44821", "issue:34567|issue:36331", - "issue:43404|issue:45987", "issue:43526|issue:45479", + "issue:43404|issue:45987", "issue:44062|issue:45081", - "issue:43906|issue:45381", "issue:27957|issue:28282", "issue:43756|issue:46399", "issue:44913|issue:46121", "issue:44792|issue:44964", + "issue:43906|issue:45381", "issue:43756|issue:45325", "issue:45042|issue:45081", - "issue:43452|issue:44488", "issue:43742|issue:44589", + "issue:43452|issue:44488", "issue:45357|issue:45920", "issue:43531|issue:44821", "issue:45405|issue:45600", "issue:43643|issue:45375" ], - "prepared_review_unit_hash": "618bee7731f22f39421c60b7c68da478310cbf9bc4bfed4afd0506c08d6b7471", + "prepared_review_unit_hash": "a30be84874f22047c547fcb2d59678d4e8d490b1feb05e4dc3017ef795c05bea", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36188, + "serialized_chars": 36187, "estimated_input_tokens": 9047, "estimated_eval_tokens": 18350 }, @@ -6351,180 +6471,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous: the issues span unrelated models, loaders, dtype paths, tokenizer bugs, and config serialization problems. Only the SigLIP2 pair (43493/43994) looks like a plausible duplicate; the rest are distinct reports and should not be merged.", - "confidence": 0.95, - "canonical_issue_reason": null, + "summary": "This set is mostly heterogeneous and not a duplicate cluster. The only strong soft-duplicate pair is the SigLIP2 regression reports (43493/43994); the rest describe different models, different errors, or different code paths.", + "confidence": 0.34, + "canonical_issue_reason": "issue:43493 is the best anchor for the one real duplicate pair because it describes the underlying SigLIP2 implementation discrepancy at a root-cause level, not just a symptom.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:43994 is the most actionable issue overall: it names the exact model, the user-facing failure mode, and the affected entry points (AutoModel and pipeline).", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43577", - "right": "issue:45072", + "left": "issue:28282", + "right": "issue:29127", "accept": false, - "reason": "Both mention dtype mismatches, but they affect different models and code paths (BLIP2 load vs SwitchTransformers/TimmWrapper inference)." + "reason": "Different bugs: PyTorch missing import vs. LayoutLMv3 error-message clarity." }, { - "left": "issue:28282", - "right": "issue:29127", + "left": "issue:43577", + "right": "issue:45072", "accept": false, - "reason": "One is an ImportError about missing PyTorch; the other is a layoutlmv3 error-message clarity issue. No shared bug." + "reason": "Both mention dtype, but they are different models and different failures; not the same concrete code-path bug." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Qwen3Moe FP8 failure and FSDP CPU RAM efficient loading are different mechanisms and different failure modes." + "reason": "Unrelated: Qwen3 MoE FP8 loading vs. FSDP CPU RAM efficient loading." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "Both are image-model correctness reports, but SigLIP2 zero probabilities and Perceiver non-default-resolution failure are distinct bugs." + "reason": "Different vision-model failures with different symptoms and likely different model code paths." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "Packaging/backend dependency coverage is unrelated to a model-loading failure for cjvt/sleng-bert." + "reason": "Backend dependency packaging issue vs. a model-loading failure for a specific checkpoint." }, { "left": "issue:43257", "right": "issue:43756", "accept": false, - "reason": "Different Qwen3/Smollm3 bugs: weight conversion under accelerate+deepspeed vs RoPE layer dropping." + "reason": "Different Qwen/Smollm model bugs; one is load/conversion, the other is RoPE-layer dropping." }, { "left": "issue:43582", "right": "issue:43782", "accept": false, - "reason": "AppleSilicon allocator TypeError and Qwen3VL weight_only loading error are unrelated." - }, - { - "left": "issue:44737", - "right": "issue:45588", - "accept": false, - "reason": "XLNet CPU-device placement bug and flash_attention sink-less crash are different code paths." + "reason": "Apple Silicon allocator TypeError vs. Qwen3VL weight_only load error." }, { "left": "issue:43493", "right": "issue:43994", "accept": true, - "reason": "Both describe the same SigLIP2 model producing incorrect outputs; 43994 looks like a user-facing manifestation of the implementation discrepancy in 43493." + "reason": "Same underlying SigLIP2 regression: both report bad outputs for google/siglip2-base-patch16-224, one as HF-vs-JAX discrepancy and the other as nonsensical AutoModel/pipeline results." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU placement bug vs. flash-attention sinkless-model AttributeError." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Loading with init_empty_weights and AutoImageProcessor-from-URL failure are unrelated." + "reason": "init_empty_weights/_is_hf_initialized TypeError is unrelated to AutoImageProcessor URL loading." }, { "left": "issue:34567", "right": "issue:36331", "accept": false, - "reason": "TrainerState token counting and CustomTrainer.compute_loss signature mismatch are different trainer bugs." - }, - { - "left": "issue:43404", - "right": "issue:45987", - "accept": false, - "reason": "Mistral3 lm_head tying and StaticCache.get_seq_length return type are unrelated." + "reason": "Trainer state token-count tracking vs. custom compute_loss signature breakage; different trainer subsystems." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug and sequence-classification zero-loss bug are different subsystems and symptoms." + "reason": "BEiT image label reduction bug vs. sequence-classification zero-loss bug; not the same issue." }, { - "left": "issue:44062", - "right": "issue:45081", + "left": "issue:43404", + "right": "issue:45987", "accept": false, - "reason": "Tokenizer AddedToken argument duplication and Mistral regex patch backend_tokenizer crash are unrelated tokenizer bugs." + "reason": "Mistral3 lm_head tying bug vs. StaticCache seq-length return-type bug." }, { - "left": "issue:43906", - "right": "issue:45381", + "left": "issue:44062", + "right": "issue:45081", "accept": false, - "reason": "An isolated reproduction of an earlier issue and a Qwen2.5-VL video position-id regression are not the same bug." + "reason": "Tokenizer AddedToken duplication vs. mistral regex patch crashing on a missing backend_tokenizer attribute." }, { "left": "issue:27957", "right": "issue:28282", "accept": false, - "reason": "XLMRoberta Flash Attention 2 support and missing PyTorch environment detection are unrelated." + "reason": "Flash Attention 2 support for XLM-Roberta vs. missing PyTorch ImportError." }, { "left": "issue:43756", "right": "issue:46399", "accept": false, - "reason": "Both involve RoPE-adjacent models, but Smollm3 RoPE-layer loss and Qwen3.5 kernel registration failure are different code paths." + "reason": "Different model-family RoPE/kernel bugs with different failure modes." }, { "left": "issue:44913", "right": "issue:46121", "accept": false, - "reason": "GPTNeoX rotary_pct reload regression and JSON-list handling in rope validation are separate config bugs." + "reason": "GPTNeoX config reload regression vs. rope-validation list TypeError; related area but different bug." }, { "left": "issue:44792", "right": "issue:44964", "accept": false, - "reason": "Janus image-generation test failure and Phi-4 multimodal load failure are different model-specific problems." + "reason": "Janus test failure vs. Phi-4 multimodal load failure; different models and symptoms." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen2.5-VL related reports with different reproduction paths and affected behavior." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Smollm3 RoPE-layer dropping and Qwen2.5-VL still-image temporal position scaling are unrelated." + "reason": "Both involve positional encodings, but they affect different models and different code paths." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "A PIL-backend torchvision dependency regression is unrelated to the Mistral tokenizer regex crash." + "reason": "Image-processor torchvision dependency regression vs. mistral tokenizer regex crash." }, { - "left": "issue:43452", - "right": "issue:44488", + "left": "issue:43742", + "right": "issue:44589", "accept": false, - "reason": "gguf_file loading breakage and inability to load cjvt/sleng-bert are both loading issues, but not the same concrete failure." + "reason": "MobileLLM key error vs. Float8 storage lookup failure are unrelated loading problems." }, { - "left": "issue:43742", - "right": "issue:44589", + "left": "issue:43452", + "right": "issue:44488", "accept": false, - "reason": "MobileLLM-125M key error and missing Float8 storage type are unrelated load/runtime issues." + "reason": "gguf_file support breakage for tokenizer/model loading vs. a specific checkpoint load failure." }, { "left": "issue:45357", "right": "issue:45920", "accept": false, - "reason": "Qwen3.5 save_pretrained visual-key regression and AutoTokenizer wrong token IDs across multiple models are different regressions." + "reason": "Qwen3.5 save_pretrained visual-key regression vs. AutoTokenizer wrong token IDs for unrelated 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 bug vs. AutoImageProcessor URL loading issue." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "A PEFT version bump problem and auto_mappings referencing removed configs are separate CI/package issues." + "reason": "PEFT version pinning problem vs. broken auto_mappings reference to removed configs." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Generic trust_remote_code config field loss and Qwen3_5MoeVisionConfig missing deepstack_visual_indexes are both config issues, but not the same bug." + "reason": "Generic trust_remote_code config-field loss vs. a specific Qwen3_5MoeVisionConfig field-drop bug; not the same underlying change." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analysis is grounded in the packet and the conclusions are conservative overall. The one accepted pair, 43493/43994, is a plausible duplicate about SigLIP2 incorrect outputs, while the rest are reasonably rejected as different models, code paths, or bug classes. No soft-edge verdict looks overstated based on the provided titles and evidence." + "feedback": "The summary is grounded in the packet: most pairs are clearly different models, errors, or code paths, and the lone accepted SigLIP2 pair is the only plausible soft duplicate. The verdicts stay conservative overall, with no obvious overclaim beyond what the titles support." }, "evaluator_used": true, "retried": false, @@ -6543,12 +6663,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:24643", "issue:27957", "issue:28282", "issue:29127", + "issue:29159", "issue:30064", "issue:32090", "issue:33357", @@ -6583,7 +6704,6 @@ "issue:44936", "issue:45005", "issue:45216", - "issue:45290", "issue:45356", "issue:45372", "issue:45381", @@ -6597,21 +6717,21 @@ "soft_pairs": [ "issue:44625|issue:44877", "issue:43720|issue:44898", - "issue:24643|issue:29127", "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:45372|issue:46170", + "issue:43819|issue:44811", "issue:44936|issue:45823", + "issue:43906|issue:46202", "issue:28282|issue:33357", "issue:43295|issue:45356", "issue:44361|issue:44442", @@ -6621,222 +6741,219 @@ "issue:45542|issue:46178", "issue:24643|issue:27957", "issue:43653|issue:44568", - "issue:44514|issue:45290", + "issue:44464|issue:44743", "issue:43577|issue:43697", - "issue:44464|issue:44743" + "issue:29159|issue:33357" ], - "prepared_review_unit_hash": "ddb3e76b0adf2c727e6872060f515c70de4e711429a863445b2412e82387fa25", + "prepared_review_unit_hash": "84e9592430723f51acb15c241bc3f875eee529089a6adf01d70b0249ed35b56c", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36685, - "estimated_input_tokens": 9172, - "estimated_eval_tokens": 18600 + "serialized_chars": 36660, + "estimated_input_tokens": 9165, + "estimated_eval_tokens": 18586 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated bug reports across tokenizer, processor, training, quantization, and serialization paths. Only one soft duplicate looks plausible: the tied-weights save/serialization regression pair (44466 \u2194 45005).", - "confidence": 0.87, - "canonical_issue_reason": "Issue 45005 is the best umbrella report: it describes the broader v5 tied-weights serialization regression, which fits the more specific lm_head.weight device-dependent serialization symptom in 44466.", + "summary": "The cluster is heterogeneous: the soft pairs mostly share broad vocabulary (tokenizers, configs, loading, training, serialization) but do not describe the same underlying bug or change. No duplicate edges should be accepted.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out because the items are not one duplicate set; they span unrelated regressions across tokenization, loading, training, and serialization.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45005 is the strongest cluster representative because it is the broadest and most actionable issue about the underlying tied-weights/save_pretrained bug.", + "best_issue_reason": "issue:43749 is the strongest standalone report in the set, with clear impact and the richest discussion, but it is not a representative duplicate umbrella for the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44625", "right": "issue:44877", "accept": false, - "reason": "Both are config/loading regressions, but they affect different models and different failure modes; not the same underlying bug." + "reason": "Both are config-related, but one is Qwen3.5 num_labels propagation and the other is a strict-config load failure for granite_speech; different model families and failure modes." }, { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "Unrelated code paths: BitNet accelerate weight unpacking vs Perceiver positional interpolation." + "reason": "BitNet packed-weight loading and Perceiver positional-encoding interpolation are unrelated code paths and bugs." }, { - "left": "issue:24643", - "right": "issue:29127", + "left": "issue:43697", + "right": "issue:45216", "accept": false, - "reason": "DeepSpeed training weight-shape error and a LayoutLMv3 error-message clarity issue are not the same bug." + "reason": "One is a v5 output-difference regression in RTDetrV2; the other is broken checkpoint saving for Qwen3.5. Same release window, different underlying issue." }, { - "left": "issue:43697", - "right": "issue:45216", + "left": "issue:24643", + "right": "issue:29127", "accept": false, - "reason": "Both are v5 regressions, but one is RTDetrV2 output drift and the other is Qwen3.5 checkpoint saving; different code paths." + "reason": "DeepSpeed training shape error vs. a request to improve a LayoutLMv3 error message; not the same bug." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "Gemma3 ZeRO-3 loading failure and trainer.evaluate-after-train behavior are different problems." + "reason": "Gemma3 loading with ZeRO-3 and trainer.evaluate() after train() are different Trainer/DeepSpeed failures." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Same model family, but one is chat-template batching and the other is video vision_position_ids; not the same concrete bug." + "reason": "Both involve Qwen2.5-VL, but one is a batched chat-template crash and the other is wrong vision_position_ids for video input; distinct defects." }, { "left": "issue:44493", "right": "issue:46050", "accept": false, - "reason": "Unexpected position-id keys vs sinq quantized model save/load failure are unrelated." + "reason": "Unexpected position-id keys during model loading is not the same as sinq quantization breaking save/load roundtrips." }, { "left": "issue:44466", "right": "issue:45005", - "accept": true, - "reason": "Both report the v5 tied-weights serialization/save_pretrained regression; 44466 is a more specific manifestation of the broader bug in 45005." + "accept": false, + "reason": "Both mention tied weights/serialization, but one is device-dependent lm_head serialization and the other is translation-model tied-weight issues; too broad to merge." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "Tokenizer/AutoModel loading with gguf_file is unrelated to Blip2 dtype propagation." - }, - { - "left": "issue:43906", - "right": "issue:46202", - "accept": false, - "reason": "A reproduction of an older issue and a Gemma pipeline roundtrip bug are not the same underlying defect." + "reason": "gguf_file loading failure and BLIP2 dtype propagation are unrelated loading bugs." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "Both involve AutoTokenizer loading, but the gguf_file breakage and offline model_info call are distinct code paths." + "reason": "Both touch from_pretrained loading, but gguf_file breakage and unconditional model_info() in _patch_mistral_regex are different code paths." }, { "left": "issue:30064", "right": "issue:33666", "accept": false, - "reason": "Void segmentation map processing is unrelated to Qwen2-VL multi-GPU training." + "reason": "Void segmentation-map processing and Qwen2-VL multi-GPU training are unrelated." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Auxiliary-loss normalization and FSDP CPU RAM efficient loading are different issues." + "reason": "Aux-loss normalization in OLMoE/GPT Oss is unrelated to FSDP CPU-RAM-efficient loading." }, { - "left": "issue:43819", - "right": "issue:44811", + "left": "issue:28282", + "right": "issue:32090", "accept": false, - "reason": "DAC latent decoding mismatch and Whisper batch_decode skip_special_tokens handling are unrelated." + "reason": "PyTorch import detection failure and a NoneType broadcast error in Trainer are different problems." }, { "left": "issue:45372", "right": "issue:46170", "accept": false, - "reason": "Gemma 4 processor import failure and Windows/ROCm torch.distributed import issues are different backend problems." + "reason": "Gemma 4 processor loading depends on mistral_common's ReasoningEffort import, while the other is a Windows/ROCm torch.distributed dependency issue." }, { - "left": "issue:28282", - "right": "issue:32090", + "left": "issue:43819", + "right": "issue:44811", "accept": false, - "reason": "Missing PyTorch import and Trainer NoneType broadcast are unrelated errors." + "reason": "DAC STE mismatch and Whisper batch_decode skip_special_tokens behavior are unrelated." }, { "left": "issue:44936", "right": "issue:45823", "accept": false, - "reason": "Trainer lifecycle failure and Gemma4 device_map forward mismatch are different code paths." + "reason": "Trainer.evaluate() after train() and Gemma4 PLE device mismatch under device_map='auto' are different failures." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "An isolated reproduction of another issue is not the same as a Gemma pipeline roundtrip bug." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "General PyTorch-import error and a MacOS bus error with CLIP are not the same bug." + "reason": "Missing PyTorch import and a MacOS bus error in a CLIP model are unrelated." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Both are regressions, but one is custom processor/tokenizer behavior and the other is Kimi-K2.5 codec/regex handling." + "reason": "One is a processor/tokenizer regression around custom model code; the other is a Kimi-K2.5 tokenizer codec/mistral-regex regression." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "Different tokenizer failures on different models; no shared underlying defect is evident." + "reason": "Different tokenizer failures on different models; not the same underlying tokenizer bug." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBird special-token registration and MLukeTokenizer task AttributeError are unrelated." + "reason": "BigBird special-token registration and MLukeTokenizer task failures are unrelated tokenizer bugs." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "save_pretrained shard-size question and AutoProcessor kwargs propagation are not the same issue." + "reason": "save_pretrained shard-size question and AutoProcessor kwargs forwarding are unrelated." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "Siglip2 zero probabilities and processor.tokenizer regression are unrelated." + "reason": "Zero-probabilities in SigLIP2 and a processor.tokenizer regression are different model behaviors." }, { "left": "issue:45542", "right": "issue:46178", "accept": false, - "reason": "Both touch backend detection, but they are different missing-backend/registration failures." + "reason": "Both are backend-dependency errors, but tensorboard-without-TensorFlow and tensorflow_text mapping failures are different issues." }, { "left": "issue:24643", "right": "issue:27957", "accept": false, - "reason": "DeepSpeed training weight-shape error and XLMRoberta Flash Attention 2 support are different issues." + "reason": "DeepSpeed weight-shape error and XLMRoberta FlashAttention2 support are unrelated." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "Both involve token special handling, but the BigBird mask-token bug and mDeBERTa BOS/EOS regression are distinct." + "reason": "BigBird mask-token registration and mDeBERTa BOS/EOS add_special_tokens regression are different tokenizer special-token bugs." }, { - "left": "issue:44514", - "right": "issue:45290", + "left": "issue:44464", + "right": "issue:44743", "accept": false, - "reason": "Both are chat/template related, but one is batched input padding and the other is tool-call assistant messages; different bugs." + "reason": "Chunked generation with compiled forward and recurrent-state reset with cache are separate generation/cache problems." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "Blip2 dtype propagation and RTDetrV2 output differences are unrelated." + "reason": "BLIP2 dtype loading and RTDetrV2 output divergence are unrelated." }, { - "left": "issue:44464", - "right": "issue:44743", + "left": "issue:29159", + "right": "issue:33357", "accept": false, - "reason": "Both involve generation/cache behavior, but one is chunked generation with compiled forward and the other is recurrent-state reset in modular_qwen3_5; not enough to call the same bug." + "reason": "Slow/fast tokenizer inconsistency and a MacOS CLIP bus error are not the same bug." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The cluster summary is well grounded in the packet, and the only accepted soft edge (44466 \u2194 45005) is conservative and plausibly the same tied-weights/save_pretrained regression. The rejected pairs are also appropriately cautious given the titles alone." - }, - "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 }, @@ -6851,11 +6968,9 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ - "issue:29159", "issue:32090", - "issue:33357", "issue:33666", "issue:34567", "issue:36010", @@ -6882,6 +6997,7 @@ "issue:44464", "issue:44466", "issue:44479", + "issue:44514", "issue:44554", "issue:44561", "issue:44610", @@ -6890,6 +7006,7 @@ "issue:44843", "issue:44977", "issue:44991", + "issue:45290", "issue:45440", "issue:45491", "issue:45588", @@ -6899,262 +7016,270 @@ "issue:46002", "issue:46050", "issue:46209", + "issue:46365", "issue:46402" ], "soft_pairs": [ - "issue:29159|issue:33357", "issue:44843|issue:44991", - "issue:43646|issue:43749", + "issue:44514|issue:45290", "issue:43531|issue:43742", + "issue:43646|issue:43749", "issue:43756|issue:46402", - "issue:43688|issue:45440", "issue:43688|issue:45704", - "issue:45491|issue:46002", + "issue:43688|issue:45440", "issue:43653|issue:43927", + "issue:45491|issue:46002", "issue:43550|issue:43927", "issue:43688|issue:45701", - "issue:44464|issue:44977", "issue:44479|issue:44625", + "issue:44464|issue:44977", "issue:43550|issue:44361", "issue:43643|issue:44561", "issue:41720|issue:44464", - "issue:44977|issue:46209", "issue:32090|issue:34567", - "issue:43577|issue:45588", + "issue:44977|issue:46209", "issue:43577|issue:44554", + "issue:43577|issue:45588", "issue:44466|issue:46050", "issue:32090|issue:33666", "issue:43504|issue:43720", - "issue:43504|issue:43550", "issue:33666|issue:36010", + "issue:43504|issue:43550", "issue:43504|issue:44610", "issue:43742|issue:44821", - "issue:45823|issue:46050", + "issue:44554|issue:46365", + "issue:43720|issue:43927", "issue:43262|issue:43295", "issue:43452|issue:44451", - "issue:43720|issue:43927", - "issue:43606|issue:43927" + "issue:43606|issue:43927", + "issue:45823|issue:46050" ], - "prepared_review_unit_hash": "3c8e171f028b544ed6b1500c5bf520e77889959b48f87162912e819d481c11dc", + "prepared_review_unit_hash": "35f52b83821daadacd5f48652e8c6a7ee0a463a68152f764ee0c12155144ddfd", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 32, - "serialized_chars": 36691, - "estimated_input_tokens": 9173, - "estimated_eval_tokens": 18602 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37603, + "estimated_input_tokens": 9401, + "estimated_eval_tokens": 19058 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of false-positive soft matches across unrelated bugs/regressions (tokenization, model loading, training, generation, quantization, and device-mapping). There are no PRs in the input. If forced to anchor the cluster, issue:43927 is the most connected and concrete save/load regression, but it does not represent a true duplicate umbrella for the whole set.", - "confidence": 0.34, - "canonical_issue_reason": "issue:43927 has the highest soft-link connectivity among the candidates and is a concrete, actionable config/save-load bug around token IDs. It is the best anchor only in a weak clustering sense, not because the whole set is one duplicate.", + "summary": "The cluster is highly heterogeneous: most items are unrelated bugs across loading, tokenization, training, generation, and model-specific processor paths. The soft pairs look like surface-level similarity rather than the same underlying fix, so none should be merged.", + "confidence": 0.95, + "canonical_issue_reason": "No clear canonical issue: the items span different subsystems and concrete failure modes, so there is no single bug that plausibly represents the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43927 is the strongest overall issue candidate by centrality and clarity of repro; it is broad enough to serve as the cluster anchor, though the surrounding issues are not true duplicates.", + "best_issue_reason": "No strong best issue anchor: even the closest-looking pairs are distinct enough that picking one representative issue would be misleading.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:29159", - "right": "issue:33357", - "accept": false, - "reason": "Tokenizer inconsistency vs MacOS bus error in a CLIP model; different subsystems and failure modes." - }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Both involve tokenizer loading, but one is offline Mistral regex patching and the other is a specific tokenizer load failure; not the same bug." + "reason": "Both involve tokenizer loading, but one is an offline Mistral regex/hub call bug and the other is a specific tokenizer load failure for EMBEDDIA/est-roberta; different code paths." }, { - "left": "issue:43646", - "right": "issue:43749", + "left": "issue:44514", + "right": "issue:45290", "accept": false, - "reason": "Custom model initialization breakage vs FSDP CPU RAM-efficient loading; unrelated code paths." + "reason": "Both crash in apply_chat_template, but one is batched padding=False handling and the other is assistant tool-call messages with no content; distinct edge cases." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior vs MobileLLM key error during load; different issues." + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key-error loading are unrelated failures." }, { - "left": "issue:43756", - "right": "issue:46402", + "left": "issue:43646", + "right": "issue:43749", "accept": false, - "reason": "Smollm3 RoPE layer selection bug vs Sam3 text encoder weights not loading; unrelated model internals." + "reason": "Custom model init breakage and FSDP CPU RAM efficient loading are different failure modes, not one shared bug." }, { - "left": "issue:43688", - "right": "issue:45440", + "left": "issue:43756", + "right": "issue:46402", "accept": false, - "reason": "Aux-loss normalization bug vs DeepseekV3 implementation divergence; not the same underlying defect." + "reason": "Smollm3 RoPE-layer handling and Sam3 text encoder weight loading are separate model-specific issues." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Aux-loss normalization vs T5 apex memory leak; unrelated." + "reason": "Aux-loss normalization in MoE models and a T5 RMSNorm memory leak are unrelated." }, { - "left": "issue:45491", - "right": "issue:46002", + "left": "issue:43688", + "right": "issue:45440", "accept": false, - "reason": "Gemma3 sliding-window NaNs vs classification-head NaNs on init; different root causes." + "reason": "Different model families and different defects: auxiliary-loss normalization vs remote/native implementation divergence." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "BigBirdTokenizer special-token registration issue vs DiaConfig token-ID persistence loss; both token-related but different mechanisms." + "reason": "Tokenizer special-token registration vs config token-ID persistence after save/load are different bugs." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both mention NaNs, but the causes differ: Gemma3 sliding-window attention vs classification head init/checkpoint behavior." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "Bamba torch.compile SDPA failure vs config token-ID persistence; unrelated code paths." + "reason": "torch.compile SDPA failure in Bamba and DiaConfig token-ID loss are unrelated." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "Aux-loss normalization bug vs version-dependent tokenization changes; not the same bug." + "reason": "MoE auxiliary-loss normalization and tokenization version changes are not the same issue." }, { - "left": "issue:44464", - "right": "issue:44977", + "left": "issue:44479", + "right": "issue:44625", "accept": false, - "reason": "Compiled chunked generation inconsistency vs Qwen3.5 flash-attention generation failure; different generation paths." + "reason": "A video-input regression in Qwen vision-language models and num_labels propagation in Qwen3.5 are different bugs." }, { - "left": "issue:44479", - "right": "issue:44625", + "left": "issue:44464", + "right": "issue:44977", "accept": false, - "reason": "Video-input regression vs num_labels propagation bug; unrelated." + "reason": "Chunked generation inconsistency with compiled forward is not the same as a flash-attention generation regression." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "torch.compile SDPA failure vs MLukeTokenizer AttributeError; different components." + "reason": "Bamba SDPA/compile failure and MLukeTokenizer task AttributeError are unrelated." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Missing fields from trust_remote_code config loading vs removal of is_torch_fx_available breaking trust_remote_code models; same area, different root cause." + "reason": "Both involve trust_remote_code, but one is missing fields in AutoConfig output and the other is removal of a helper breaking remote-code models; not one concrete bug." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Qwen3 auto device-map cuda assert vs chunked generation with compiled forward; unrelated." + "reason": "CUDA assert on auto device mapping and chunked-generation inconsistency are different code paths and symptoms." }, { - "left": "issue:44977", - "right": "issue:46209", + "left": "issue:32090", + "right": "issue:34567", "accept": false, - "reason": "Qwen3.5 flash-attention generation issue vs FP8 inference failure on B200; different hardware/code paths." + "reason": "Trainer broadcast TypeError and num_input_tokens_seen not updating are separate Trainer bugs." }, { - "left": "issue:32090", - "right": "issue:34567", + "left": "issue:44977", + "right": "issue:46209", "accept": false, - "reason": "Trainer broadcast TypeError vs num_input_tokens_seen not updating; both Trainer-related but not the same bug." + "reason": "Flash-attention generation issues in Qwen3.5 and FP8 inference failures on B200 are unrelated." }, { "left": "issue:43577", - "right": "issue:45588", + "right": "issue:44554", "accept": false, - "reason": "BLIP2 dtype propagation bug vs flash_attention s_aux=None crash; unrelated." + "reason": "Blip2 dtype loading and an MPS attention correctness issue do not share a concrete fix path." }, { "left": "issue:43577", - "right": "issue:44554", + "right": "issue:45588", "accept": false, - "reason": "BLIP2 dtype issue vs MPS attention correctness issue; different defects." + "reason": "A dtype-loading bug and a flash-attention s_aux=None crash are distinct." }, { "left": "issue:44466", "right": "issue:46050", "accept": false, - "reason": "lm_head serialization inconsistency vs quantized save/load breakage; both serialization-related but not the same failure." + "reason": "Serialization of tied lm_head weights and sinq quantized save/load breakage are different bugs." }, { "left": "issue:32090", "right": "issue:33666", "accept": false, - "reason": "Trainer NoneType broadcast error vs Qwen2-VL multi-GPU training issue; unrelated." + "reason": "Generic Trainer _gpu_broadcast TypeError and Qwen2-VL multi-GPU training are too broad to treat as the same bug." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Beit legacy-field loading bug vs BitNet packed-weight unpacking during accelerate loading; both loading issues but different models and fixes." + "reason": "Legacy field loading in Beit and BitNet packed-weight unpacking during accelerate loading are different root causes." }, { - "left": "issue:43504", - "right": "issue:43550", + "left": "issue:33666", + "right": "issue:36010", "accept": false, - "reason": "Beit pretrained-load issue vs Bamba torch.compile SDPA failure; unrelated." + "reason": "Multi-GPU Qwen2-VL training and a GenerationMixin import error are unrelated." }, { - "left": "issue:33666", - "right": "issue:36010", + "left": "issue:43504", + "right": "issue:43550", "accept": false, - "reason": "Qwen2-VL multi-GPU training vs GenerationMixin import error; unrelated." + "reason": "Different models and failures: Beit preset loading vs Bamba SDPA/compile behavior." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Beit legacy-field load failure vs OmDet processor/model input-size mismatch; unrelated." + "reason": "Beit legacy-field loading and OmDet-Turbo processor size mismatch are unrelated." }, { "left": "issue:43742", "right": "issue:44821", "accept": false, - "reason": "MobileLLM key error on load vs AutoImageProcessor URL loading; different components." + "reason": "Loading a specific model with a key error and loading AutoImageProcessor from URL are different loading paths." }, { - "left": "issue:45823", - "right": "issue:46050", + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "An MPS attention correctness bug and a Trainer import issue caused by datasets are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:43927", "accept": false, - "reason": "Gemma4 device-mismatch during forward vs sinq quantized save/load breakage; unrelated." + "reason": "BitNet packed-weight loading and DiaConfig token-ID persistence are not the same bug." }, { "left": "issue:43262", "right": "issue:43295", "accept": false, - "reason": "Audio chat-template sample-rate default vs processor.tokenizer/image regression; both processor issues but not the same bug." + "reason": "Audio chat-template sampling-rate defaulting and processor.tokenizer regression are different processor bugs." }, { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "gguf_file loading failure vs ScandiBERT load failure; unrelated model-loading problems." + "reason": "gguf_file loading failures and inability to load a specific tokenizer/model are not the same underlying issue." }, { - "left": "issue:43720", + "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "BitNet packed-weight loading bug vs DiaConfig token-ID persistence loss; different code paths." + "reason": "CPU-offload device mismatch in bark-small and DiaConfig token-ID loss are unrelated." }, { - "left": "issue:43606", - "right": "issue:43927", + "left": "issue:45823", + "right": "issue:46050", "accept": false, - "reason": "CPU-offload device mismatch in bark-small vs config token-ID persistence; unrelated." + "reason": "Gemma4 PLE device mismatch and sinq save/load quantization breakage are different defects." } ] }, @@ -7176,7 +7301,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:28218", "issue:33357", @@ -7228,8 +7353,8 @@ "issue:46424" ], "soft_pairs": [ - "issue:43994|issue:46192", "issue:33666|issue:36331", + "issue:43994|issue:46192", "issue:43525|issue:43572", "issue:44849|issue:46209", "issue:43278|issue:43688", @@ -7237,14 +7362,14 @@ "issue:44625|issue:46002", "issue:43493|issue:43825", "issue:28218|issue:33357", - "issue:39692|issue:43493", "issue:44964|issue:45405", + "issue:39692|issue:43493", "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:43540|issue:44977", + "issue:45200|issue:45229", "issue:44488|issue:44843", "issue:39692|issue:43295", "issue:43278|issue:43844", @@ -7255,231 +7380,237 @@ "issue:43257|issue:43299", "issue:44589|issue:45520", "issue:43824|issue:44589", - "issue:45229|issue:45823", "issue:43927|issue:44361", "issue:43504|issue:43606", + "issue:45229|issue:45823", "issue:43572|issue:44291" ], - "prepared_review_unit_hash": "c02270b36815665c85dbff5cbf9a77bd9285b0b6741fc5b06755219b00919931", + "prepared_review_unit_hash": "516889d22131dbf2fa51054fe07d354c468648fb8134919635bd66506cbff2d0", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37111, - "estimated_input_tokens": 9278, - "estimated_eval_tokens": 18812 + "serialized_chars": 37122, + "estimated_input_tokens": 9281, + "estimated_eval_tokens": 18818 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly false-positive similarity: the issues span unrelated model families and failure modes. The only broad commonality is \u201crecent Transformers/v5 regressions,\u201d but the concrete bugs differ too much to merge as duplicates.", - "confidence": 0.82, - "canonical_issue_reason": "If one issue has to anchor the cluster, `issue:43646` is the broadest umbrella (`Transformers 5.0.0 breaks custom model initialization`) and best matches the general v5-regression theme seen in a few nearby reports. Still, the cluster is not a true duplicate set.", + "summary": "These candidates are not a duplicate cluster: they span unrelated tokenizer, model-loading, training, config, inference, and documentation problems across many model families. I would not merge any soft pair.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the items are heterogeneous and do not share one underlying bug or change.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:43646` is the most representative issue because it is phrased at the right abstraction level and covers a general regression class rather than a single model-specific symptom.", + "best_issue_reason": "No single issue is a good global representative because the set is broad and mostly unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43994", - "right": "issue:46192", + "left": "issue:33666", + "right": "issue:36331", "accept": false, - "reason": "Different bugs: SigLIP2 inference correctness vs `return_tensors` being ignored when `text_kwargs` is passed." + "reason": "Both are training-related, but one is Qwen2-VL multi-GPU usage and the other is a CustomTrainer signature regression; different bugs and code paths." }, { - "left": "issue:33666", - "right": "issue:36331", + "left": "issue:43994", + "right": "issue:46192", "accept": false, - "reason": "Different areas and symptoms: multi-GPU training setup vs `compute_loss()` signature breakage in a custom trainer." + "reason": "One is a SigLIP2 model output-quality issue, the other is a preprocessing argument being ignored; no shared underlying failure." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Both are missing token-id config fields after v5, but they affect different model configs (`Llama4Config` vs `StableLmConfig`) and likely separate fixes." + "reason": "Both mention missing pad-token fields, but they affect different config classes and different attributes (`pad_token_id` vs `pad_token_idx`)." }, { "left": "issue:44849", "right": "issue:46209", "accept": false, - "reason": "Qwen3.5 hidden-state behavior vs FP8 inference failure on B200 are unrelated code paths." + "reason": "Qwen3.5 hidden-state behavior and FP8 inference on B200 are unrelated issues." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Embedding dtype drift during evaluate is unrelated to MoE auxiliary-loss normalization." + "reason": "Embedding dtype drift at eval and auxiliary-loss normalization are separate model-behavior bugs." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "Packed-weight loading bug in BitNet vs tokenizer/task AttributeError in MLuke are different subsystems." + "reason": "Packed-weight loading in BitNet and a tokenizer/task AttributeError are unrelated." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "Config propagation for Qwen3.5 labels is not the same as NaNs in classification heads after checkout/init." + "reason": "One is config propagation for Qwen3.5, the other is NaNs in classification heads; different symptoms and paths." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "SigLIP2 implementation mismatch is unrelated to the pipeline error-message regression." + "reason": "SigLIP2 implementation discrepancy and a pipeline error-message regression are unrelated." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "Tokenizer spacing bug and MacOS CLIP bus error do not share the same underlying failure." + "reason": "Tokenizer spacing behavior and a MacOS bus error in a CLIP model are not the same bug." }, { - "left": "issue:39692", - "right": "issue:43493", + "left": "issue:44964", + "right": "issue:45405", "accept": false, - "reason": "Docs-example mistakes are not the same bug as a SigLIP2 implementation discrepancy." + "reason": "Model loading failure and a dependency-version bump are not duplicates." }, { - "left": "issue:44964", - "right": "issue:45405", + "left": "issue:39692", + "right": "issue:43493", "accept": false, - "reason": "Model loading failure and a dependency-version bump issue are unrelated." + "reason": "One is a docs/example page with multiple errors, the other is a core SigLIP2 implementation mismatch; not one concrete fix." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 regressions: hidden states vs incorrect visual-encoder keys on save." + "reason": "Hidden-states output bug and incorrect save_pretrained key serialization are different regressions." }, { "left": "issue:44568", "right": "issue:45812", "accept": false, - "reason": "Both affect tokenization, but one is about special tokens on a specific tokenizer and the other is a broad Granite token-ID regression." + "reason": "Both are tokenizer regressions, but they affect different model families and different tokenization behaviors." }, { "left": "issue:44464", "right": "issue:46424", "accept": false, - "reason": "Chunked-generation inconsistency with compiled forward is not the same as cache-length handling for quantized cache." + "reason": "Chunked generation with compiled forward and cache-length handling for quantized cache are unrelated." }, { - "left": "issue:45200", - "right": "issue:45229", + "left": "issue:43493", + "right": "issue:43844", "accept": false, - "reason": "Gemma4 text-only token-type defaults and multi-GPU OOM are different problems." + "reason": "A model-implementation discrepancy and a ZeRO-3 gradient anomaly are separate issues." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Video-input processing in Qwen3OmniMoe is unrelated to flash-attention generation behavior in Qwen3.5." + "reason": "Video-input processing in Qwen3OmniMoe and Qwen3.5 flash-attention generation bugs do not share a concrete code path." }, { - "left": "issue:43493", - "right": "issue:43844", + "left": "issue:45200", + "right": "issue:45229", "accept": false, - "reason": "SigLIP2 model discrepancy and gradient blow-up under DeepSpeed/ZeRO-3 are different code paths." + "reason": "Text-only mm_token_type_ids handling and Gemma4 multi-GPU OOM are unrelated." }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "Loading a specific model and offline-mode breakage from `model_info()` are unrelated." + "reason": "Model load failure for a specific checkpoint and an offline-mode bug in AutoTokenizer are different problems." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "Documentation example errors are not the same bug as the processor/tokenizer regression in v4.57.5." + "reason": "Docs/example mistakes and a regression around `processor.tokenizer` access are not the same underlying bug." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Training/eval dtype change is unrelated to ZeRO-3 gradient inflation." + "reason": "A dtype change during evaluation and abnormal gradient growth under ZeRO-3 are distinct training issues." }, { "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Both mention missing pad-token fields, but they are different configs and symptoms (`pad_token_id` vs `pad_token_idx`)." + "reason": "Both involve missing pad-token config fields, but they are model-specific config regressions with different attributes and loaders." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization breakage is not the same as device-dependent `lm_head.weight` serialization." + "reason": "Custom model initialization breakage and device-dependent serialization of `lm_head.weight` are unrelated." }, { "left": "issue:33357", "right": "issue:33666", "accept": false, - "reason": "MacOS CLIP bus error and Qwen2-VL multi-GPU training are unrelated." + "reason": "A bus error on MacOS and Qwen2-VL multi-GPU training are not the same failure mode." }, { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Missing `pad_token_id` in Qwen3-VL config is unrelated to `from_config` not unwrapping `text_config`." + "reason": "Missing `pad_token_id` in Qwen3-VL and failure to unwrap `text_config` in composite Qwen 3.5/3.6 configs are different bugs." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both involve Qwen3 MoE, but one is an accelerate/deepspeed weight-conversion bug and the other is a load-time regression in dev; not the same concrete fix." + "reason": "Different Qwen3 MOE loading regressions with different triggers and symptoms." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Float8 storage lookup failure and missing `flash_attn` import key on Python 3.13 are unrelated." + "reason": "An unsupported float8 storage type and a flash_attn import KeyError on Python 3.13 are unrelated." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "Missing Qwen2.5-VL symbol export and Float8 storage lookup are separate failures." - }, - { - "left": "issue:45229", - "right": "issue:45823", - "accept": false, - "reason": "Gemma4 OOM under multi-GPU inference is not the same as a PLE device-mismatch forward bug." + "reason": "Importing Qwen2.5-VL and float8 storage lookup failures are completely separate." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "Dia save/load token-ID loss and MLuke tokenizer task AttributeError are different issues." + "reason": "Config save/load token-ID loss and MLukeTokenizer task AttributeErrors are not the same bug." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field loading failure in Beit semantic segmentation is unrelated to CPU-offload device mismatch in bark-small." + "reason": "Legacy field handling in Beit loading and a CPU offload device mismatch in bark-small are unrelated." }, { - "left": "issue:46424", - "right": "issue:44464", + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 OOM and Gemma4 device-map mismatch are both in the same model family but are different concrete failures." + }, + { + "left": "issue:45357", + "right": "issue:45759", + "accept": false, + "reason": "Incorrect visual-encoder keys on save and missing `text_config` unwrapping on from_config are different composite-model regressions." + }, + { + "left": "issue:43844", + "right": "issue:43688", "accept": false, - "reason": "Cache max-length being ignored for quantized cache is not the same bug as inconsistent chunked generation with compiled forward." + "reason": "Gradient explosion under ZeRO-3 and auxiliary-loss normalization are distinct bugs." }, { "left": "issue:43572", "right": "issue:44291", "accept": false, - "reason": "Missing `pad_token_idx` in one config and unexpected `_is_hf_initialized` during `init_empty_weights` are different regressions." + "reason": "A missing config field after v5 migration and an unexpected `_is_hf_initialized` argument in init_empty_weights are unrelated." } ] }, @@ -7501,7 +7632,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:15354", "issue:28218", @@ -7562,8 +7693,8 @@ "issue:43994|issue:44493", "issue:43873|issue:46050", "issue:45588|issue:46050", - "issue:44423|issue:44734", "issue:43065|issue:46369", + "issue:44423|issue:44734", "issue:29942|issue:30333", "issue:42490|issue:43696", "issue:45406|issue:46121", @@ -7575,195 +7706,195 @@ "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:45200|issue:45823", + "issue:44521|issue:45290", "issue:37428|issue:43957" ], - "prepared_review_unit_hash": "37099c17afa29e0f5b4fd755c2e7f7462bf987199ad1eb6c44ec6e6f99e17f49", + "prepared_review_unit_hash": "0878fbbede52490904ac9fa4ad05df24b3c95b20651a9e412256ffd205a0913d", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 35585, - "estimated_input_tokens": 8897, - "estimated_eval_tokens": 18050 + "serialized_chars": 35579, + "estimated_input_tokens": 8895, + "estimated_eval_tokens": 18046 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: it mixes unrelated tokenizer, training, serving, quantization, attention, and model-loading bugs. None of the proposed soft edges look like the same underlying issue, so all should be rejected. No PRs are present.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "This cluster is heterogeneous: it spans unrelated tokenizer, model-loading, serving, attention, quantization, and CI issues. None of the soft pairs look like true duplicates of the same underlying bug/change, so all soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out; the cluster is too broad and the soft pairs do not share a common root cause.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:44521 is the strongest standalone bug report here because it is open, has the most inbound attention, and describes a concrete, reproducible multimodal chat-template failure.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "One is BF16/FP32 dtype drift in evaluation; the other is a resume_from_checkpoint key mismatch. Different code paths and failure modes." + "reason": "Different problems: embedding dtype drift during eval vs resume_from_checkpoint key mismatch." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "Tensor-parallel OOM while loading a large model is unrelated to a PEFT minimum-version packaging issue." + "reason": "OOM during tensor-parallel model load is unrelated to a PEFT version bump." }, { "left": "issue:44493", "right": "issue:46285", "accept": false, - "reason": "Unexpected position-id key handling is not the same as a CLIP text-model structure/API change." + "reason": "Unexpected position-id key behavior is not the same as CLIP text-model structural inconsistency." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "A failing Janus image-generation test is unrelated to the PEFT version bump problem." + "reason": "Janus image-generation test failure and PEFT minimum-version issue are unrelated." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Legacy field loading in Beit segmentation and BigBirdTokenizer special-token registration are unrelated bugs." + "reason": "Legacy preset loading for BEiT segmentation is a different bug from BigBirdTokenizer special-token registration." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Both mention device-related CI failures, but one is CPU offload mismatch and the other is a stale GraniteSpeech test; not the same bug." + "reason": "Both mention device behavior, but one is a CPU-offload mismatch and the other is a CI test regression." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 incorrect outputs and unexpected position-id key warnings are different model-level issues." + "reason": "Siglip2 bad outputs and unexpected position-id keys are different model-path failures." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Quantized offloading behavior and sinq save/load corruption are different quantization problems with different failure points." + "reason": "Quantization/offloading behavior is not the same as broken save_pretrained/from_pretrained after sinq quantization." }, { "left": "issue:45588", "right": "issue:46050", "accept": false, - "reason": "A flash-attention integration crash on s_aux=None is unrelated to sinq serialization breakage." + "reason": "Flash-attention crash on sink-less models is unrelated to quantized model serialization." }, { - "left": "issue:44423", - "right": "issue:44734", + "left": "issue:43065", + "right": "issue:46369", "accept": false, - "reason": "Both affect serve, but one is multimodal continuous batching and the other is KV-cache continuation indexing; different concrete code paths." + "reason": "Dummy Conv2d placeholder in Sam3PixelDecoder is a different issue from a memory overflow in sam3-litetext." }, { - "left": "issue:43065", - "right": "issue:46369", + "left": "issue:44423", + "right": "issue:44734", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and Sam3-litetext memory overflow are distinct SAM3 issues." + "reason": "Two separate serve crashes with different symptoms and code paths: string .to() error vs KV-cache tensor indexing." }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 test failures and MLFlow job reporting on failure are unrelated subsystems." + "reason": "Flash Attention 2 test failures and MLflow reporting on failed jobs are unrelated." }, { "left": "issue:42490", "right": "issue:43696", "accept": false, - "reason": "Trailing-slash import warnings are unrelated to a CUDA OOM in GPT-oss-20b." + "reason": "Import warning for trailing-slash model paths is unrelated to GPT-oss CUDA OOM." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and JSON-loaded list handling in rope validation are unrelated bugs." + "reason": "Gemma4Processor missing _tokenizer is not the same as rope-parameter type conversion failing on a JSON list." }, { "left": "issue:43844", "right": "issue:44964", "accept": false, - "reason": "ZeRO-3 gradient growth during training is unrelated to loading Phi-4 multimodal models." + "reason": "ZeRO-3 gradient explosion and Phi-4 multimodal loading failure are different bugs." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing after added tokens and TrainerState num_input_tokens_seen not updating are different tokenizer/trainer behaviors." + "reason": "Added-token spacing in tokenizers is unrelated to TrainerState token-count tracking." }, { "left": "issue:29159", "right": "issue:32090", "accept": false, - "reason": "Slow-vs-fast tokenizer inconsistency is unrelated to a Trainer GPU broadcast NoneType error." + "reason": "Slow/fast tokenizer inconsistency does not match a Trainer broadcast NoneType error." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Template-node compilation failure and a flash_attn import KeyError are separate import/compile problems." + "reason": "Non-template-node compilation failure is unrelated to flash_attn import key errors on Python 3.13." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "An attention correctness bug on MPS is unrelated to a flex_attention PyTorch version check import bug." + "reason": "MPS attention correctness and a flex_attention PyTorch version check are distinct issues." }, { "left": "issue:43582", "right": "issue:44977", "accept": false, - "reason": "Apple Silicon allocator TypeError and Qwen3.5 flash-attention generation issues are different failures." + "reason": "Apple Silicon allocator warmup TypeError is unrelated to Qwen3.5 flash-attention generation failure." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp export failure is unrelated to unexpected zero probabilities in SigLIP2." + "reason": "TorchScript GeneratorExp export error and Siglip2 zero-probability outputs are unrelated." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Video-input processing in Qwen3OmniMoe and Gemma4 mm_token_type_ids defaults are different multimodal model issues." + "reason": "Video-input processing failure in Qwen3OmniMoe is different from Gemma4 mm_token_type_ids defaults." }, { - "left": "issue:44521", - "right": "issue:45290", + "left": "issue:45072", + "right": "issue:45198", "accept": false, - "reason": "All-zero assistant masks for multimodal inputs and tool-call messages with no content are both chat-template related, but they are distinct edge cases." + "reason": "BF16 dtype mismatches in inference are not the same as Wav2Vec2 save/tokenization failures." }, { "left": "issue:45200", "right": "issue:45823", "accept": false, - "reason": "Missing mm_token_type_ids defaults in fine-tuning is unrelated to a device_map auto forward mismatch in Gemma4 PLE." + "reason": "Gemma4 token-type-id defaults and Gemma4 PLE device mismatch are separate model bugs." }, { - "left": "issue:45072", - "right": "issue:45198", + "left": "issue:44521", + "right": "issue:45290", "accept": false, - "reason": "bfloat16 dtype mismatches in inference and Wav2Vec2 save_pretrained/tokenization failures are unrelated." + "reason": "Both touch apply_chat_template, but one is all-zero assistant masks for multimodal inputs while the other is a crash on tool-call messages without content." }, { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Flash-support import errors and meta-device loading regressions are separate model-loading issues." + "reason": "Flash-support import error and meta-device loading regressions are different issues with different triggers." } ] }, @@ -7785,17 +7916,17 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:28218", "issue:35141", "issue:37428", "issue:38175", "issue:39692", - "issue:41720", "issue:42490", "issue:42915", "issue:43232", + "issue:43278", "issue:43295", "issue:43334", "issue:43388", @@ -7809,6 +7940,7 @@ "issue:43701", "issue:43824", "issue:43844", + "issue:43856", "issue:43873", "issue:43950", "issue:43957", @@ -7833,20 +7965,18 @@ "issue:46178", "issue:46192", "issue:46209", - "issue:46285", - "issue:46365" + "issue:46285" ], "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:45381|issue:46209", "issue:43295|issue:44610", - "issue:44554|issue:46365", - "issue:39692|issue:46192", "issue:43572|issue:44987", + "issue:39692|issue:46192", "issue:43295|issue:43873", "issue:44554|issue:45356", "issue:43572|issue:43957", @@ -7859,223 +7989,217 @@ "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:42915|issue:43824", "issue:43824|issue:43975", "issue:44792|issue:45563", "issue:43388|issue:43526", "issue:28218|issue:35141", - "issue:38175|issue:43493" + "issue:38175|issue:43493", + "issue:43278|issue:43856" ], - "prepared_review_unit_hash": "802bc70d31a9cdd6cf52ade4a8c038a6bcd6264dba55fd36ed9ef14e36fd8592", + "prepared_review_unit_hash": "bad98418b19254e26236040eecb911b778cb7d52f2b350f553baa8137f62a8c5", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 37015, - "estimated_input_tokens": 9254, - "estimated_eval_tokens": 18764 + "soft_pair_count": 29, + "serialized_chars": 36717, + "estimated_input_tokens": 9180, + "estimated_eval_tokens": 18616 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous. The closest items form a loose processor/tokenizer-regression subgroup, but none of the soft pairs look like the same concrete bug or change. Issue 43295 is the most central representative of that subgroup.", - "confidence": 0.88, - "canonical_issue_reason": "Issue 43295 is the broadest, most central item among the few related processor/tokenizer regression reports and has multiple soft links, making it the best cluster anchor.", + "summary": "This cluster is a loose mix of unrelated Transformers bug reports across tokenization, multimodal processing, model loading, quantization, and generation. None of the soft pairs look like true duplicates; they mostly share broad subsystem vocabulary rather than the same concrete bug. The most representative issue is issue:43295 because it is a central regression report with several nearby similarity links, but it is only a cluster representative, not a duplicate of the others.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43295 is the most central representative in this set: it is a clear regression report, touches a user-facing multimodal code path, and has multiple similarity links to other processor/model-loading issues. It is still not a duplicate of the rest, just the best cluster anchor.", "canonical_pr_reason": null, - "best_issue_reason": "43295 best represents the cluster\u2019s main recurring theme (processor/tokenizer regressions) and is the most general, well-scoped issue among the candidates.", + "best_issue_reason": "issue:43295 is the best single issue representative because it sits near several soft-linked reports and describes a concrete regression affecting custom multimodal model code.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "Different problems: SigLIP2 implementation discrepancy vs. resume-from-checkpoint key mismatch." + "reason": "Different bugs: SigLIP2 implementation discrepancy vs resume_from_checkpoint key mismatch." }, { "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "Both involve tokenization/processor APIs, but the crashes/bugs are in different code paths and are not the same underlying issue." + "reason": "Different failures: LayoutLMv2 tokenizer crash on NER/padding vs return_tensors being ignored when text_kwargs is passed." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to the Qwen3.5 hidden-states bug." + "reason": "Different generation bugs: sync_gpus kwargs handling vs Qwen3.5 output_hidden_states behavior." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Device-mapping forward mismatch is unrelated to silently ignored return_tensors handling." + "reason": "Different subsystems: Gemma4 device_map/PLE device mismatch vs tokenizer argument handling in processors." }, { - "left": "issue:45381", - "right": "issue:46209", + "left": "issue:43295", + "right": "issue:43450", "accept": false, - "reason": "Qwen2.5-VL video position-id bug and FP8 inference failure on B200 are different code-path failures." + "reason": "Both involve multimodal code, but one is processor.tokenizer regression and the other is batched video processor shape output; not the same bug." }, { - "left": "issue:43295", - "right": "issue:43450", + "left": "issue:45381", + "right": "issue:46209", "accept": false, - "reason": "Both are processor-related, but one is about processor.tokenizer access/passing images and the other is batched video shape output; not the same bug." + "reason": "Different issues: Qwen2.5-VL vision_position_ids for video input vs FP8 inference failure on B200." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Different model-specific processor bugs: custom tokenizer/image handling vs. OmDet image size mismatch." + "reason": "Custom processor/tokenizer regression is unrelated to OmDet-Turbo processor producing the wrong image size." }, { - "left": "issue:44554", - "right": "issue:46365", + "left": "issue:43572", + "right": "issue:44987", "accept": false, - "reason": "MPS attention correctness and Trainer import failure are unrelated." + "reason": "Both are load-time regressions, but one is missing pad_token_idx in StableLmConfig and the other is a model-loading failure for physical-intelligence/fast." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "Documentation/model mismatch and return_tensors being ignored are separate issues." - }, - { - "left": "issue:43572", - "right": "issue:44987", - "accept": false, - "reason": "Missing pad_token_idx in a config and a model-loading failure for a specific repo are not the same bug." + "reason": "SigLIP2 doc/example errors are unrelated to the tokenizer return_tensors bug." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Processor/tokenizer regression is unrelated to quantization offloading behavior." + "reason": "Custom processor/tokenizer breakage is unrelated to quantization offloading not working as expected." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "Attention on MPS and Kimi-K2.5 tokenizer regression are unrelated." + "reason": "Different model-specific correctness bugs: MPS attention value-head issue vs Kimi-K2.5 tokenizer regression." }, { "left": "issue:43572", "right": "issue:43957", "accept": false, - "reason": "Both are loading/config regressions, but different models and failure modes." + "reason": "Both are loading-related, but StableLmConfig pad token metadata and meta-device loading failures are distinct problems." }, { "left": "issue:45520", "right": "issue:46178", "accept": false, - "reason": "Different import/backend lookup failures in separate modules." + "reason": "Different import/backend problems: flash_attn import on Python 3.13 vs tensorflow_text backend mapping error." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Distinct import errors for different symbols/models." + "reason": "Both are import errors, but for completely different missing symbols and code paths." }, { "left": "issue:44466", "right": "issue:45127", "accept": false, - "reason": "Both touch tied-weight behavior, but one is serialization inconsistency and the other is LoRA merge collapse; not the same concrete issue." + "reason": "Different weight-serialization bug vs LoRA merge collapse with extended vocabulary." }, { "left": "issue:43950", "right": "issue:46285", "accept": false, - "reason": "Non-persistent buffer corruption is unrelated to CLIPTextModel structure changes." + "reason": "Non-persistent buffer corruption in from_pretrained is unrelated to CLIP class structure inconsistency." }, { "left": "issue:44898", "right": "issue:45412", "accept": false, - "reason": "Perceiver resize failure and RT-DETR memory retention are unrelated." + "reason": "Perceiver non-default resolution failure is unrelated to RT-DETR memory release behavior." }, { "left": "issue:43550", "right": "issue:43844", "accept": false, - "reason": "torch.compile/SDPA failure is unrelated to gradient blow-up under ZeRO-3." + "reason": "Different training/runtime issues: torch.compile SDPA failure on Bamba vs gradient explosion under ZeRO-3." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Text-only mm_token_type_ids defaults and Qwen2.5-VL rope indexing are different bugs." + "reason": "Different multimodal position/defaulting bugs in different models; one is mm_token_type_ids defaults, the other is still-image temporal position_ids scaling." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Similar symptom name, but different model families and code paths; not the same underlying fix." + "reason": "Same symptom style, but different model families and code paths: Qwen3-VL vs Llama4Config pad_token_id." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video processor batching bug is unrelated to Qwen2-57B TP OOM." - }, - { - "left": "issue:42915", - "right": "issue:43824", - "accept": false, - "reason": "FP8 config failure for Qwen3Moe is unrelated to the Qwen2.5-VL import error." + "reason": "Video processor shape bug and Qwen2-57B-A14B-instruct OOM are unrelated." }, { "left": "issue:42490", "right": "issue:43493", "accept": false, - "reason": "Local-path warning behavior is unrelated to the SigLIP2 discrepancy." + "reason": "Import warning for local paths ending with slash is unrelated to SigLIP2 implementation discrepancy." }, { - "left": "issue:41720", - "right": "issue:46209", + "left": "issue:42915", + "right": "issue:43824", "accept": false, - "reason": "CUDA assert under auto device mapping and FP8 inference failure on B200 are different hardware/runtime issues." + "reason": "Qwen3 MoE fine-grained FP8 failure is unrelated to Qwen2.5-VL import error." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "Importing a class and tokenizer detokenization are unrelated." + "reason": "ImportError for a missing class and detokenization regression are different bugs." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "Failed image generation test and stale warning text are not the same bug." + "reason": "Failed janus image-generation test is unrelated to a stale generate() warning." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "Different label-handling layers: metrics gathering vs. image processor reduction." + "reason": "Both concern label handling, but one is gather_for_metrics truncation with tuple labels and the other is BeitImageProcessorFast reduce_labels behavior." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resize." + "reason": "Tokenizer extra-space detokenization bug is unrelated to embedding reinitialization after resize." }, { "left": "issue:38175", "right": "issue:43493", "accept": false, - "reason": "Same model family, but one is zero probabilities and the other is a broader implementation discrepancy; not enough to treat as the same bug." + "reason": "Unexpected zero probabilities in SigLIP2 is not the same as the HF-vs-JAX discrepancy report, even if related to the same model family." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype changing to FP32 in eval is unrelated to Qwen3 MoE memory inefficiency." } ] }, @@ -8097,14 +8221,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:29942", "issue:32090", "issue:37428", "issue:38175", + "issue:41720", "issue:43262", - "issue:43278", "issue:43335", "issue:43425", "issue:43450", @@ -8115,13 +8239,12 @@ "issue:43653", "issue:43819", "issue:43827", - "issue:43856", "issue:43901", "issue:44112", "issue:44206", + "issue:44351", "issue:44442", "issue:44448", - "issue:44493", "issue:44514", "issue:44521", "issue:44561", @@ -8143,27 +8266,28 @@ "issue:45759", "issue:46170", "issue:46192", + "issue:46209", "issue:46234", "issue:46285", "issue:46326", "issue:46402" ], "soft_pairs": [ - "issue:43278|issue:43856", - "issue:43526|issue:46192", + "issue:41720|issue:46209", "issue:43653|issue:44112", + "issue:43526|issue:46192", "issue:44442|issue:45198", - "issue:45290|issue:46326", "issue:38175|issue:45412", + "issue:45290|issue:46326", "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:44514|issue:44521", "issue:43819|issue:44112", "issue:43827|issue:43901", + "issue:44206|issue:45584", "issue:43819|issue:45083", "issue:44857|issue:45593", "issue:43335|issue:43550", @@ -8175,205 +8299,212 @@ "issue:45200|issue:45588", "issue:45127|issue:45245", "issue:43335|issue:43653", - "issue:43819|issue:44493", - "issue:44855|issue:45083" + "issue:44855|issue:45083", + "issue:37428|issue:44351", + "issue:43550|issue:44112" ], - "prepared_review_unit_hash": "ec29193c84294c7e88b4afd86360a9942a412df55eed712685de61f8e747aded", + "prepared_review_unit_hash": "8df2be93658c732dfdb26259f0f020e777f698f619a69b91cd80bc51eaa08c08", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 36344, - "estimated_input_tokens": 9086, - "estimated_eval_tokens": 18428 + "soft_pair_count": 29, + "serialized_chars": 36621, + "estimated_input_tokens": 9156, + "estimated_eval_tokens": 18568 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a grab-bag of mostly unrelated issue reports. There are a few loose thematic overlaps (chat-template bugs, tokenizer/model-loading regressions, import-utils symbols), but none are the same underlying defect, so no soft-edge pair should be merged.", - "confidence": 0.91, - "canonical_issue_reason": "issue:44521 is the strongest issue anchor here: it is open, high-activity, and has the broadest concrete reproduction around multimodal apply_chat_template behavior.", + "summary": "This cluster is a broad false-positive mix of unrelated Transformers issues: chat templating, tokenizer/model loading, vision/audio processors, import/platform regressions, and a few model-specific training/inference bugs. None of the soft pairs look like the same underlying bug or a mergeable PR-equivalent change.", + "confidence": 0.95, + "canonical_issue_reason": "No safe canonical issue: the apparent similarities are mostly superficial (shared APIs or broad subsystem names) and do not converge on one underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the best overall issue candidate because it is active, well-scoped, and central enough to serve as the representative issue for this mixed set.", + "best_issue_reason": "issue:44521 is the strongest standalone report in the set: it has the most discussion/inbound references and a clear, user-facing repro for multimodal assistant masks. It is representative, but it should not absorb the other issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43278", - "right": "issue:43856", - "accept": false, - "reason": "Different bugs: eval dtype drift vs MoE training memory usage." - }, - { - "left": "issue:43526", - "right": "issue:46192", + "left": "issue:41720", + "right": "issue:46209", "accept": false, - "reason": "Different processor bugs: label reduction vs return_tensors handling." + "reason": "Both are GPU-related, but one is an auto device-mapping cuda assert on A800 and the other is FP8 inference failure on B200; different code paths and root causes." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "Tokenizer special-token registration issue vs GraniteSpeech CI device-override test failure." + "reason": "Tokenizer special-token decode bug vs a stale GraniteSpeech CI device-override test; different components and failure modes." }, { - "left": "issue:44442", - "right": "issue:45198", + "left": "issue:43526", + "right": "issue:46192", "accept": false, - "reason": "Both audio-related, but different failure modes and different model/tokenizer stacks." + "reason": "BeitImageProcessorFast label reduction bug vs return_tensors being ignored when text_kwargs is passed; unrelated processor behaviors." }, { - "left": "issue:45290", - "right": "issue:46326", + "left": "issue:44442", + "right": "issue:45198", "accept": false, - "reason": "Both touch apply_chat_template, but one is tool-call empty-content crashing and the other is mistral_common continue_final_message behavior." + "reason": "AutoTokenizer load failure for FastSpeech2ConformerTokenizer vs wav2vec2 save_pretrained/tokenization failure; different tokenizer families and bugs." }, { "left": "issue:38175", "right": "issue:45412", "accept": false, - "reason": "Different subsystems: unexpected probabilities vs RT-DETR memory release." + "reason": "Unexpected zero probabilities in SigLIP2 vs RT-DETR memory not released on delete; not the same underlying issue." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both involve apply_chat_template-like flows, but one crashes on assistant messages with tool calls and no content while the other is about continue_final_message for mistral_common tokenizers." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "Audio sampling-rate default bug vs video batch-shape bug." + "reason": "Audio processor sampling-rate default bug vs video processor batched-shape bug; different modalities and code paths." }, { "left": "issue:45520", "right": "issue:46170", "accept": false, - "reason": "Unrelated platform/import problems: flash_attn lookup on Python 3.13 vs Windows ROCm torch.distributed dependency." + "reason": "Python 3.13 flash_attn import KeyError vs Windows ROCm torch.distributed dependency issue; unrelated platform/import regressions." }, { "left": "issue:45759", "right": "issue:46285", "accept": false, - "reason": "Both concern composite model structure, but the concrete failures differ: text_config unwrapping vs CLIP .text_model inconsistency." - }, - { - "left": "issue:44514", - "right": "issue:44521", - "accept": false, - "reason": "Both are VL/chat-template issues, but one is a padding=False batch crash and the other is all-zero assistant masks." + "reason": "Qwen multimodal config unwrapping bug vs CLIP text-model structure inconsistency; different model families and different fixes." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Different model and bug class: processor image size mismatch vs AMP loss crash." + "reason": "OmDet processor size mismatch vs LwDetrImageLoss AMP float16 crash; both are vision bugs but not the same defect." }, { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "Both mention import_utils availability helpers, but they are missing different symbols with different breakage paths." + "reason": "Removal of is_torch_fx_available breaking trust_remote_code models vs missing is_falcon_mamba_ssm_available export; different missing symbols." }, { - "left": "issue:44206", - "right": "issue:45584", + "left": "issue:44514", + "right": "issue:44521", "accept": false, - "reason": "Feature-extractor arg regression vs Whisper empty-transcription generation failure." + "reason": "Batched apply_chat_template crash with padding=False vs all-zero assistant_masks for multimodal inputs; related API area but different symptom and trigger." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "Different defects: DAC latent parity vs GraniteSpeech CI device override." + "reason": "DAC.from_latents / STE mismatch vs GraniteSpeech stale device-override CI failure; unrelated bugs." }, { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both docs issues, but they target different pipeline APIs and different deprecations." + "reason": "Both are docs issues, but they cover different deprecated APIs (pipeline removal vs return_all_scores); not duplicates." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Unsupported center arg in LasrFeatureExtractor vs Whisper empty-transcription behavior after align_special_tokens; different audio bugs." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "Unrelated audio/codebook latent mismatch vs helper length-calculation behavior." + "reason": "DAC latent reconstruction mismatch vs qwen3_omni_moe feature-length helper behavior; no shared underlying code path." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "Different detection-model bugs: float16 AMP crash vs auxiliary-loss omission." + "reason": "Float16 AMP crash in LwDetrImageLoss vs D-FINE not using auxiliary losses when denoising is off; different training logic." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "Config construction bug vs torch.compile/SDPA runtime incompatibility." + "reason": "SwitchTransformersConfig sparse-layer creation bug vs Bamba torch.compile SDPA failure; different models and fixes." }, { "left": "issue:43611", "right": "issue:46402", "accept": false, - "reason": "Both affect model loading, but one is base_model_prefix handling and the other is missing Sam3 text-encoder weights." + "reason": "Base-model-prefix loading break in Transformers 5.0 vs Sam3Model text encoder weights not loading; both are loading issues but not the same one." }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "Specific missing flash_attention_utils symbol vs broad Torch 2.10 incompatibility." + "reason": "Missing _flash_supports_window_size import vs broad Torch 2.10 incompatibility; one specific import regression, one version-compatibility report." }, { "left": "issue:43606", "right": "issue:45083", "accept": false, - "reason": "CPU-offload device mismatch vs audio feature-length helper behavior; not the same bug." + "reason": "CPU-offload device mismatch in suno/bark-small vs qwen3_omni_moe output-length helper bug; unrelated subsystems." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures are unrelated to NoneType GPU broadcast errors in Trainer." + "reason": "Flash Attention 2 test failures vs Trainer _gpu_broadcast_one NoneType TypeError; different failures and code paths." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Different generation regressions: Pegasus v4/v5 output drift vs qwen3_5 recurrent state reset with cache." + "reason": "Pegasus v4/v5 output difference vs Qwen3.5 recurrent-state reset with cache; different model families and root causes." }, { "left": "issue:45200", "right": "issue:45588", "accept": false, - "reason": "Different multimodal processor bugs: missing mm_token_type_ids default vs flash_attention.py s_aux=None crash." + "reason": "Gemma 4 mm_token_type_ids defaulting bug vs flash_attention integration crash on s_aux=None; not the same issue." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA/tied-embedding collapse is unrelated to category-count overflow." + "reason": "LoRA merge collapse with extended vocab/tied embeddings vs a category-count limit error; unrelated failure modes." }, { "left": "issue:43335", "right": "issue:43653", "accept": false, - "reason": "Sparse-layer config bug vs tokenizer special-token registration bug." + "reason": "SwitchTransformers sparse-layer config bug vs BigBirdTokenizer special-token registration bug; different components entirely." }, { - "left": "issue:43819", - "right": "issue:44493", + "left": "issue:44855", + "right": "issue:45083", "accept": false, - "reason": "Different symptoms and code paths: DAC latent mismatch vs unexpected position-id keys across models." + "reason": "Python 3.13 @torch.jit.script parsing error in DebertaV2Model vs qwen3_omni_moe helper behavior; no relation." }, { - "left": "issue:44855", - "right": "issue:45083", + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Two import errors, but for different missing symbols (_flash_supports_window_size vs HybridCache); not duplicates." + }, + { + "left": "issue:43550", + "right": "issue:44112", "accept": false, - "reason": "Python 3.13 decorator parsing/import error vs qwen3_omni_moe feature-length helper bug." + "reason": "Bamba torch.compile SDPA failure vs GraniteSpeech stale CI device override; unrelated bugs." } ] }, @@ -8395,15 +8526,15 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:28218", "issue:29127", "issue:29942", "issue:32090", + "issue:33357", "issue:34689", "issue:35141", - "issue:37428", "issue:38175", "issue:39692", "issue:40444", @@ -8412,16 +8543,15 @@ "issue:43335", "issue:43504", "issue:43526", - "issue:43550", "issue:43606", "issue:43701", "issue:43720", "issue:43746", + "issue:43819", "issue:43825", "issue:43854", "issue:44112", "issue:44206", - "issue:44351", "issue:44360", "issue:44448", "issue:44493", @@ -8447,19 +8577,18 @@ "issue:46234" ], "soft_pairs": [ - "issue:37428|issue:44351", - "issue:43550|issue:44112", + "issue:43819|issue:44493", "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:29127|issue:32090", "issue:39692|issue:43526", + "issue:45290|issue:45381", "issue:45310|issue:45478", - "issue:39692|issue:44493", - "issue:45459|issue:46234", "issue:43335|issue:45083", + "issue:45459|issue:46234", + "issue:39692|issue:44493", "issue:43720|issue:44112", "issue:44448|issue:44779", "issue:43335|issue:44112", @@ -8468,231 +8597,238 @@ "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:38175|issue:44493", "issue:28218|issue:32090", + "issue:44448|issue:44977", "issue:40444|issue:43526", - "issue:43746|issue:46209", + "issue:44977|issue:45229", "issue:43701|issue:45083", - "issue:29942|issue:35141" + "issue:43746|issue:46209", + "issue:29942|issue:35141", + "issue:32090|issue:33357" ], - "prepared_review_unit_hash": "ace45230a82d5a3e03ce3f2620a34902461580ac751fcf51d15de5f351044844", + "prepared_review_unit_hash": "0c1f514c07c6bc608ae1e0ed8c7b5621ed678a62927dd7635426790f190be709", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 37048, - "estimated_input_tokens": 9262, - "estimated_eval_tokens": 18780 + "serialized_chars": 36464, + "estimated_input_tokens": 9116, + "estimated_eval_tokens": 18488 }, "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. The only clear duplicate pair is the repeated Qwen3.5 MoE `from_pretrained` error report; the rest differ by model, subsystem, or failure mode and should stay separate.", - "confidence": 0.93, - "canonical_issue_reason": "issue:45310 is the earlier, broader report of the same Qwen3.5 MoE `from_pretrained` regression as issue:45478, so it is the best issue to keep as canonical among the candidates.", + "summary": "This cluster is mostly heterogeneous and does not form a single duplicate family: it mixes tokenizer regressions, model-loading bugs, CI failures, config issues, and offload/dtype problems. The only clear duplicate pair is the two identical Qwen3.5 MoE from_pretrained reports (issue:45310 and issue:45478).", + "confidence": 0.9, + "canonical_issue_reason": "issue:45920 is the strongest representative issue: it is open, has the broadest impact, and describes a multi-model AutoTokenizer regression rather than a narrow one-off failure.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45310 best represents the only true duplicate theme in this set: the Qwen3.5 MoE `from_pretrained` regression. It is the earlier report and captures the concrete failure more cleanly than the later repost.", + "best_issue_reason": "issue:45920 is the best issue to keep as the cluster representative because it covers a widespread v5 tokenizer regression across several models and has inbound references.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:37428", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but they concern different missing symbols (`_flash_supports_window_size` vs `HybridCache`) and likely different code paths." - }, - { - "left": "issue:43550", - "right": "issue:44112", + "left": "issue:43819", + "right": "issue:44493", "accept": false, - "reason": "Different models and failures: torch.compile/SDPA behavior in Bamba versus a GraniteSpeech CI device-override issue." + "reason": "Different bugs: DAC.from_latents/forward mismatch vs unexpected position_id keys across models." }, { "left": "issue:44206", "right": "issue:45823", "accept": false, - "reason": "One is a feature-extractor argument regression; the other is a Gemma4 device-mapping mismatch during forward. Not the same bug." - }, - { - "left": "issue:45290", - "right": "issue:45381", - "accept": false, - "reason": "Chat template tool-call crash and Qwen2.5-VL video position-id bug are separate failure modes." + "reason": "Different components and failure modes: feature extractor center arg crash vs Gemma4 device mismatch." }, { "left": "issue:43335", "right": "issue:43504", "accept": false, - "reason": "Different architectures and symptoms: SwitchTransformers config construction versus BEiT pretrained preset loading." + "reason": "Unrelated model/config bug vs legacy-field pretrained loading bug." }, { - "left": "issue:29127", - "right": "issue:32090", + "left": "issue:43278", + "right": "issue:43825", "accept": false, - "reason": "An error-message clarity issue in LayoutLMv3 is unrelated to the Trainer `_gpu_broadcast_one` TypeError." + "reason": "Embedding dtype drift in train/eval is unrelated to a pipeline error-message regression." }, { "left": "issue:45588", "right": "issue:45859", "accept": false, - "reason": "Both touch model internals, but one is a flash-attention crash on `s_aux=None` and the other is a CPU-offload incompatibility in rotary embeddings." + "reason": "Both are offload-related, but one is a flash_attention aux-state crash and the other is a rotary-embedding CPU-offload incompatibility." }, { - "left": "issue:43278", - "right": "issue:43825", + "left": "issue:29127", + "right": "issue:32090", "accept": false, - "reason": "A dtype change between train/eval is unrelated to a pipeline warning about translation tasks in v5." + "reason": "Different errors in different subsystems: LayoutLMv3 box validation vs Trainer GPU broadcast NoneType." }, { "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "Documentation/example errors for SigLIP2 are unrelated to BEiT Fast label reduction returning only one label." + "reason": "Documentation example problems do not match a BeitImageProcessorFast reduce_labels bug." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash is unrelated to Qwen2.5-VL video position-id computation." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "These are the same Qwen3.5 MoE `from_pretrained` error report; the titles and described failure match closely." + "reason": "Same underlying bug: identical Qwen3.5 MoE from_pretrained failure on transformers>=5.4.0." }, { - "left": "issue:39692", - "right": "issue:44493", + "left": "issue:43335", + "right": "issue:45083", "accept": false, - "reason": "A SigLIP2 doc example problem is not the same as the broad positional-id warning regression." + "reason": "SwitchTransformers sparse-layer config bug is unrelated to qwen3_omni_moe output-length helper behavior." }, { "left": "issue:45459", "right": "issue:46234", "accept": false, - "reason": "Tokenizer protobuf error handling and a missing import-utils symbol are unrelated issues." + "reason": "Tokenizer protobuf decode error hiding vs a missing import_utils symbol are different issues." }, { - "left": "issue:43335", - "right": "issue:45083", + "left": "issue:39692", + "right": "issue:44493", "accept": false, - "reason": "SwitchTransformers sparse-layer construction and qwen3_omni_moe feature-length computation are different code paths." + "reason": "SigLIP2 docs example errors are unrelated to a position-id key mismatch." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "Packed-weight loading in BitNet is unrelated to the GraniteSpeech stale device-override CI failure." + "reason": "Packed-weights loading bug and GraniteSpeech CI stale override failure are different problems." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Both are output regressions, but one is Pegasus generation drift and the other is DeepSeek tokenizer tokenization; not the same bug." + "reason": "Both are version regressions, but they affect different models and code paths." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "SwitchTransformers config creation bug is unrelated to GraniteSpeech CI device override behavior." }, { "left": "issue:43335", "right": "issue:45593", "accept": false, - "reason": "A SwitchTransformers config bug is not the same as D-FINE missing auxiliary losses when denoising is off." + "reason": "Sparse-layer config bug is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "Both are runtime failures, but one is CPU-offload device mismatch and the other is AMP/CUDA loss behavior in LwDetrImageLoss." + "reason": "CPU offload device mismatch is not the same as an AMP/CUDA loss crash." }, { "left": "issue:42490", "right": "issue:43854", "accept": false, - "reason": "A local-path trailing-slash warning is unrelated to a GLM-4.7-Flash load failure in tests." + "reason": "Local-path import warning is unrelated to the GLM-4.7-Flash test loading failure." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Llama 3.2 Vision loading breakage and embedding reinitialization on resize are different issues." + "reason": "Model-loading breakage and embedding reinitialization after resize are different bugs." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "A DSA indexer activation omission is unrelated to the removal of `is_torch_fx_available` breaking `trust_remote_code` models." - }, - { - "left": "issue:38175", - "right": "issue:44493", - "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and unexpected key/position-id warnings are not the same underlying defect." + "reason": "DSA indexer architecture issue is unrelated to the removal of is_torch_fx_available." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Both mention precision/device behavior, but they concern different models and different failure modes." + "reason": "Different failures: LwDetrImageLoss AMP crash vs dtype mismatch in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:45920", "right": "issue:46192", "accept": false, - "reason": "Wrong token IDs across multiple models is a tokenizer regression; silently ignoring `return_tensors` is a separate API bug." + "reason": "AutoTokenizer wrong token IDs is not the same as return_tensors being ignored when text_kwargs is passed." }, { - "left": "issue:44448", - "right": "issue:44977", + "left": "issue:38175", + "right": "issue:44493", "accept": false, - "reason": "Pegasus v4/v5 output drift and Qwen3.5 flash-attention generation failure are distinct model-specific problems." + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to position-id key warnings." }, { - "left": "issue:44977", - "right": "issue:45229", + "left": "issue:28218", + "right": "issue:32090", "accept": false, - "reason": "Flash-attention generation failure in Qwen3.5 is unrelated to Gemma4 multi-GPU OOM." + "reason": "Tokenizer spacing after added token is unrelated to Trainer broadcast errors." }, { - "left": "issue:28218", - "right": "issue:32090", + "left": "issue:44448", + "right": "issue:44977", "accept": false, - "reason": "Tokenizer spacing after added tokens has no connection to the Trainer broadcast TypeError." + "reason": "Pegasus v4/v5 output drift is a different issue from Qwen3.5 flash-attention generation failure." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "Multi-image IterableDataset fine-tuning failure is unrelated to BEiT Fast label reduction behavior." + "reason": "Multi-image IterableDataset finetuning failure is unrelated to BeitImageProcessorFast label reduction." }, { - "left": "issue:43746", - "right": "issue:46209", + "left": "issue:44977", + "right": "issue:45229", "accept": false, - "reason": "PEFT adapter loading from local checkpoints and FP8 inference on B200 are separate issues." + "reason": "Flash-attention generation bug and multi-GPU OOM are distinct failure modes." }, { "left": "issue:43701", "right": "issue:45083", "accept": false, - "reason": "Checkpoint key mismatch and qwen3_omni_moe output-length helper behavior are different bugs." + "reason": "Checkpoint key mismatch is unrelated to qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "PEFT local checkpoint loading and FP8 inference on B200 are different issues." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures and embedding reinitialization after resize are unrelated." + "reason": "Flash Attention 2 test failures do not match an embedding resize reinitialization bug." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer NoneType broadcast and MacOS bus error with CLIP are different code paths." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative: it only treats the exact Qwen3.5 MoE `from_pretrained` pair as a duplicate and keeps the other soft pairs separate based on different titles/failure modes. The canonical-issue choice is also supported by the earlier identical report. " + "feedback": "Grounded overall. The cluster summary is conservative and matches the packet: it is mostly heterogeneous, and the only clearly identical duplicate pair is issue:45310 / issue:45478. The choice of issue:45920 as representative is reasonable given its broad multi-model v5 tokenizer regression title and inbound references." }, "evaluator_used": true, "retried": false, @@ -8711,7 +8847,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:10000", "issue:15354", @@ -8719,7 +8855,6 @@ "issue:27957", "issue:32090", "issue:33290", - "issue:33357", "issue:34689", "issue:37428", "issue:41720", @@ -8734,6 +8869,7 @@ "issue:43873", "issue:43927", "issue:43976", + "issue:44186", "issue:44206", "issue:44448", "issue:44466", @@ -8763,7 +8899,6 @@ "issue:46399" ], "soft_pairs": [ - "issue:32090|issue:33357", "issue:44855|issue:46399", "issue:43550|issue:45083", "issue:42490|issue:43575", @@ -8771,230 +8906,243 @@ "issue:43927|issue:45698", "issue:44561|issue:44779", "issue:37428|issue:44987", - "issue:43873|issue:45229", "issue:43278|issue:43329", + "issue:43873|issue:45229", "issue:44857|issue:45588", - "issue:41720|issue:45229", "issue:27957|issue:34689", + "issue:41720|issue:45229", "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:45372|issue:45406", "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", - "issue:44871|issue:44977" + "issue:44206|issue:44479", + "issue:44871|issue:44977", + "issue:44186|issue:45198" ], - "prepared_review_unit_hash": "abdbe7037d7b45237b9c7a8dc6b4f66123782354f9dd893356811d33e4da4fc8", + "prepared_review_unit_hash": "19b71d1e9e6c434daf4b3f30015cd676b2391e6268153198c8aa4338a6d96f22", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36700, - "estimated_input_tokens": 9175, - "estimated_eval_tokens": 18606 + "serialized_chars": 36702, + "estimated_input_tokens": 9176, + "estimated_eval_tokens": 18608 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a mixed bag of unrelated Transformers issues rather than a duplicate set. No soft-edge pair looks like the same underlying bug or change. The most broadly representative issue is the save/load custom-module regression (#45698).", - "confidence": 0.82, - "canonical_issue_reason": "#45698 is the clearest broad regression with concrete repro and multiple references: `from_pretrained` loading the wrong custom module after `save_pretrained`.", + "summary": "Loose overcluster of mostly unrelated bug reports; the soft pairs share superficial keywords but not the same concrete failure or fix path. No soft edge should be merged.", + "confidence": 0.97, + "canonical_issue_reason": "No single issue is a clear duplicate hub; the cluster is heterogeneous.", "canonical_pr_reason": null, - "best_issue_reason": "#45698 is the strongest issue candidate overall because it is specific, actionable, and affects a common serialization/loading path.", + "best_issue_reason": "If an anchor is required, issue:45698 is the broadest and most discussion-heavy loading/serialization report, but it still does not represent the rest of the cluster well.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:32090", - "right": "issue:33357", - "accept": false, - "reason": "Different failures in different paths: Trainer GPU broadcast NoneType vs CLIP model bus error on macOS." - }, { "left": "issue:44855", "right": "issue:46399", "accept": false, - "reason": "Both mention parsing/registration problems, but they affect different models and unrelated code paths." + "reason": "Both mention decorator/script-related failures, but one is a Python 3.13 import/parser issue and the other is a Qwen3.5 registration error; different code paths." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "Unrelated bugs: torch.compile/SDPA failure in Bamba vs helper output-length logic in qwen3_omni_moe." + "reason": "Torch.compile/SDPA crash in Bamba is unrelated to a feature-length helper behavior change in qwen3_omni_moe." }, { "left": "issue:42490", "right": "issue:43575", "accept": false, - "reason": "Import warning for local paths ending with slash is unrelated to Qwen2 TP OOM." + "reason": "Local-path trailing slash warning and multi-GPU TP OOM are different problems with no shared fix path." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "Completely different models and issues: DistilBert error vs torch.jit GeneratorExp export failure." + "reason": "German DistilBERT error and ViT torch.jit GeneratorExp issue are unrelated model/export failures." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Both involve save/load, but one loses custom token IDs while the other loads the wrong custom module; different bugs." + "reason": "Both involve save/load, but one loses custom token IDs while the other loads the wrong custom module; different serialization bugs." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Removal of `is_torch_fx_available` breaks remote-code models; Deepseek tokenizer regression is a separate tokenizer issue." + "reason": "Removal of is_torch_fx_available breaks trust_remote_code, while DeepSeek tokenizer regression is a separate tokenizer behavior bug." }, { "left": "issue:37428", "right": "issue:44987", "accept": false, - "reason": "One is a missing import symbol in flash attention utils; the other is a model-loading failure for a specific repo." + "reason": "Missing flash-support import and loading physical-intelligence/fast may both touch model loading, but the titles indicate different failures and root causes." }, { - "left": "issue:43873", - "right": "issue:45229", + "left": "issue:43278", + "right": "issue:43329", "accept": false, - "reason": "Quantized offloading behavior and multi-GPU CUDA OOM are different problems." + "reason": "Embedding dtype drift in eval and undefined variables in the video token-count branch are unrelated." }, { - "left": "issue:43278", - "right": "issue:43329", + "left": "issue:43873", + "right": "issue:45229", "accept": false, - "reason": "BF16-to-FP32 dtype drift in eval is unrelated to the undefined-variable bug in multimodal token counting." + "reason": "Quantization/offloading behavior and Gemma4 multi-GPU CUDA OOM are not the same concrete bug." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "Different crash sites: AMP image loss on CUDA vs flash_attention integration on sink-less models." + "reason": "float16 AMP loss crash and flash_attention s_aux=None crash are different subsystems and failures." }, { - "left": "issue:41720", - "right": "issue:45229", + "left": "issue:27957", + "right": "issue:34689", "accept": false, - "reason": "Auto device-map assert on Qwen3 is not the same as Gemma4 multi-GPU OOM." + "reason": "Flash Attention 2 issue for XLM-R and Llama 3.2 model-loading regression are unrelated." }, { - "left": "issue:27957", - "right": "issue:34689", + "left": "issue:41720", + "right": "issue:45229", "accept": false, - "reason": "Flash Attention 2 support issue and Llama 3.2 model loading regression are unrelated." + "reason": "cudaErrorAssert on auto device mapping is a different failure mode from an OOM during Gemma4 inference." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "Generation bug with flash-attention is unrelated to a stale warning in paged generate()." + "reason": "Qwen3.5 flash-attention generation bug and a stale warning for num_return_sequences are unrelated." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token handling and Gemma3 NaN embeddings are different failures." + "reason": "Tokenizer BOS/EOS omission and Gemma3 NaN embeddings under sliding-window attention are different bugs." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma4 processor import dependency error is not the same as flash_attention.py crashing on `s_aux=None`." + "reason": "Gemma4 processor import failure and flash_attention s_aux handling are separate code paths." }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Dtype mismatch during eval is unrelated to the gradient-checkpointing eval-mode restriction." + "reason": "BF16-to-FP32 dtype change and gradient-checkpointing-in-eval restriction are unrelated." }, { "left": "issue:43412", "right": "issue:45588", "accept": false, - "reason": "Executorch export failure and flash_attention `s_aux=None` crash are different code paths." + "reason": "Executorch export guard/erfinv support problem is unrelated to the flash_attention None handling crash." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "Both are Gemma4-related, but one is a flash-attention attribute crash and the other is a device-map forward mismatch." + "reason": "Both are Gemma4-related, but one is a flash-attention sink-less-model crash and the other is a device-mismatch forward bug; not the same fix." }, { - "left": "issue:45372", - "right": "issue:45406", + "left": "issue:44448", + "right": "issue:44466", "accept": false, - "reason": "ImportError from `mistral_common` and missing `_tokenizer` on Gemma4Processor are different regressions." + "reason": "Pegasus v4/v5 output regression and tied-weight serialization inconsistency are distinct issues." }, { - "left": "issue:44448", - "right": "issue:44466", + "left": "issue:45372", + "right": "issue:45406", "accept": false, - "reason": "Sequence-output differences and inconsistent `lm_head.weight` serialization are not the same bug." + "reason": "Gemma4 processor import error and serve-time missing _tokenizer attribute are different root causes." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python-version compatibility failure and DeepseekV3MoE divergence are unrelated." + "reason": "Python version compatibility issue and DeepSeekV3MoE divergence are unrelated." }, { "left": "issue:45198", "right": "issue:45803", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure and Fuyu bare-except cleanup bug do not share the same underlying defect." + "reason": "Wav2Vec2 save/tokenization failure and Fuyu's bare except swallowing errors are separate bugs." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline `top_k` shape change is unrelated to the Trainer broadcast error." + "reason": "Pipeline top_k nesting behavior and Trainer _gpu_broadcast_one NoneType error do not share a concrete code path." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count runtime limit and stale warning in paged generate are unrelated." + "reason": "Category-limit runtime error and stale warning for num_return_sequences are unrelated." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "Tokenizer special-token registration bug and recurrent-state reset in Qwen3.5 are unrelated." + "reason": "BigBirdTokenizer special-token registration bug and Qwen3.5 recurrent-state reset bug are unrelated." }, { - "left": "issue:44206", - "right": "issue:44479", + "left": "issue:32090", + "right": "issue:33290", "accept": false, - "reason": "Feature-extractor center-arg regression and video-input regression are different issues." + "reason": "Trainer broadcast NoneType error and AdaFactor/DeepSpeed OOM are different training failures." }, { - "left": "issue:32090", - "right": "issue:33290", + "left": "issue:44206", + "right": "issue:44479", "accept": false, - "reason": "Unsupported NoneType broadcast and Adafactor/deepspeed OOM are unrelated training failures." + "reason": "Unsupported center arg in LasrFeatureExtractor and Qwen VL video-input regression are different components." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma eos_token mismatch is unrelated to Qwen3.5 flash-attention generation failure." + "reason": "Gemma eos_token_id mismatch and Qwen3.5 flash-attention generation failure are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2 tokenizer crash on NER inputs and Wav2Vec2 save/tokenization failure are separate tokenizer/model bugs." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Same as above: runtime category limit vs stale warning are not the same underlying issue." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer decode bug and recurrent-state reset under cache are unrelated." } ] }, @@ -9016,7 +9164,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:27957", "issue:29127", @@ -9040,7 +9188,6 @@ "issue:43653", "issue:43701", "issue:43756", - "issue:43761", "issue:43824", "issue:43873", "issue:43994", @@ -9052,6 +9199,7 @@ "issue:44462", "issue:44464", "issue:44625", + "issue:45072", "issue:45198", "issue:45372", "issue:45412", @@ -9062,27 +9210,25 @@ "issue:45593", "issue:45712", "issue:45834", - "issue:45859", "issue:46121", "issue:46369", "issue:46399" ], "soft_pairs": [ - "issue:43065|issue:45412", - "issue:44186|issue:45198", "issue:43611|issue:46121", "issue:42490|issue:43873", "issue:44361|issue:45198", - "issue:33357|issue:39692", + "issue:43065|issue:45412", "issue:29127|issue:34567", - "issue:43065|issue:43994", + "issue:33357|issue:39692", "issue:45459|issue:45712", + "issue:43065|issue:43994", "issue:42915|issue:42994", "issue:43066|issue:44462", + "issue:43278|issue:43575", "issue:45372|issue:45459", "issue:29942|issue:30064", "issue:45412|issue:46369", - "issue:43278|issue:43575", "issue:43994|issue:45412", "issue:44265|issue:45198", "issue:43381|issue:43701", @@ -9092,176 +9238,171 @@ "issue:27957|issue:29942", "issue:29127|issue:33357", "issue:43824|issue:44279", - "issue:43994|issue:46369", "issue:45440|issue:45593", + "issue:43994|issue:46369", "issue:43526|issue:45834", "issue:44464|issue:45584", "issue:46121|issue:46399", "issue:43526|issue:44186", "issue:29127|issue:35141", "issue:44625|issue:45584", - "issue:43761|issue:45859" + "issue:43575|issue:43701", + "issue:44442|issue:45072" ], - "prepared_review_unit_hash": "18588924c3926efe7337de00b0e0cc79a046b9feab2e680f40e235031901e1d1", + "prepared_review_unit_hash": "a553b2db2c508fc39388a5d308db8f6c9b997dc38b43d690cdb59bafb2d4c5b1", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 33, - "serialized_chars": 37493, - "estimated_input_tokens": 9374, - "estimated_eval_tokens": 19004 + "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": "False-positive cluster: the issues span many unrelated models, tokenizers, and failure modes. None of the soft-edge pairs look like the same underlying bug or a mergeable duplicate PR case.", - "confidence": 0.97, - "canonical_issue_reason": "No safe canonical issue stands out; the set is heterogeneous and does not center on one underlying defect.", + "summary": "This cluster is a loose semantic grouping of mostly unrelated Transformers bug reports across tokenizers, quantization/offloading, model loading, vision processors, training state, and generation. I don\u2019t see any soft pair that looks like the same underlying bug or change, so no duplicate merges are justified.", + "confidence": 0.34, + "canonical_issue_reason": "issue:43873 is the strongest representative here: it has substantial discussion/inbound references, a concrete reproducible runtime bug, and broad user impact around quantization/offloading.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good global representative because the cluster mixes unrelated bugs across different subsystems and model families.", + "best_issue_reason": "issue:43873 is the best single issue to anchor this cluster because it is well-scoped, actively discussed, and broadly relevant compared with the more niche reports in the set.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43065", - "right": "issue:45412", - "accept": false, - "reason": "Both mention model internals and memory, but one is a dummy Conv2d placeholder and the other is a garbage-collection memory leak; different bugs." - }, - { - "left": "issue:44186", - "right": "issue:45198", - "accept": false, - "reason": "Different tokenizer/model families and different failure paths; no shared code-path bug." - }, { "left": "issue:43611", "right": "issue:46121", "accept": false, - "reason": "Loading with base_model_prefix and rope-param dict conversion are unrelated issues." + "reason": "Different failures: model loading with base_model_prefix vs rope-parameter validation on a JSON-loaded list." }, { "left": "issue:42490", "right": "issue:43873", "accept": false, - "reason": "Local-path import warning and quantization/offloading behavior are unrelated." + "reason": "Trailing-slash local path import warning is unrelated to quantization/offloading behavior." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "Both are tokenizer-related, but they affect different models with different symptoms and causes." + "reason": "Different tokenizer/model code paths: MLukeTokenizer AttributeError vs Wav2Vec2 save/tokenization failure." }, { - "left": "issue:33357", - "right": "issue:39692", + "left": "issue:43065", + "right": "issue:45412", "accept": false, - "reason": "MacOS bus error with CLIP is unrelated to SigLIP2 documentation/example errors." + "reason": "Dummy Conv2d in Sam3PixelDecoder and RT-DETR memory release are distinct bugs in different models." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error handling and TrainerState token counting are unrelated." + "reason": "LayoutLMv3 box validation messaging and TrainerState token counting are unrelated." }, { - "left": "issue:43065", - "right": "issue:43994", + "left": "issue:33357", + "right": "issue:39692", "accept": false, - "reason": "Dummy Conv2d placeholder issue is not the same as quantized model saving failure." + "reason": "MacOS bus error with a community CLIP model is not the same as SigLIP2 docs/model mismatch." }, { "left": "issue:45459", "right": "issue:45712", "accept": false, - "reason": "Tokenizer exception masking is unrelated to leftover dummy classes." + "reason": "Tokenizer protobuf exception masking and leftover dummy classes are unrelated maintenance issues." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d is unrelated to SigLIP2 producing bad outputs with AutoModel/pipeline." }, { "left": "issue:42915", "right": "issue:42994", "accept": false, - "reason": "Both involve quantization-adjacent topics, but one is FP8 loading and the other is quantized saving; different code paths." + "reason": "Quantization config failure during load and quantized model saving failure are different stages with different root causes." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Wrong decoder type in v5 and ignoring tokenizer.json are separate tokenizer-loading bugs." + "reason": "Wrong tokenizer decoder type in v5 is a different bug from ignoring tokenizer.json in the repo." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "BF16/FP32 dtype drift in eval is unrelated to Qwen2-57B TP loading OOM." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Gemma 4 processor import failure and protobuf-related tokenizer error masking are unrelated." + "reason": "Gemma processor import failure from ReasoningEffort and hidden tokenizer errors from protobuf absence are different problems." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures and void segmentation-map processing are different subsystems." + "reason": "Flash Attention 2 test failures and void-segmentation-map processing failures are unrelated." }, { "left": "issue:45412", "right": "issue:46369", "accept": false, - "reason": "Memory leak on deletion is not the same as memory overflow when calling sam3-litetext." - }, - { - "left": "issue:43278", - "right": "issue:43575", - "accept": false, - "reason": "Embedding dtype drift and tensor-parallel OOM are different problems." + "reason": "Both mention memory, but RT-DETR GC behavior and SAM3 model memory overflow are different bugs." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "Quantized model saving failure is unrelated to model memory release on deletion." + "reason": "SigLIP2 incorrect outputs and RT-DETR memory leak are not the same code-path problem." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export failure from torch_compilable_check is unrelated to tokenizer/save_pretrained issues." + "reason": "torch.export failures with torch_compilable_check are unrelated to Wav2Vec2 save/tokenization errors." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Eval-mode gradient checkpointing and checkpoint resume key mismatch are different issues." + "reason": "Eval-mode gradient checkpointing and resume_from_checkpoint key mismatch are separate training issues." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Both concern model implementation discrepancies, but they are for different model families and different missing behavior." + "reason": "Smollm3 RoPE layer-count mismatch and DeepSeekV3 implementation divergence are different model-specific bugs." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBird special-token decode bug and AutoTokenizer loading failure are distinct tokenizer bugs." + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer issues." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Runtime post-processor updates and Gemma3 NaN embeddings are unrelated." + "reason": "Runtime special-token post-processor updates and Gemma3 NaN embeddings are unrelated." }, { "left": "issue:27957", "right": "issue:29942", "accept": false, - "reason": "XLMRoberta Flash Attention 2 compatibility and a broader FA2 test failure are not clearly the same bug." + "reason": "XLMRoberta Flash Attention 2 support and failing Flash Attention 2 tests are not the same concrete bug." }, { "left": "issue:29127", @@ -9273,61 +9414,67 @@ "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "Missing Qwen2.5-VL import and a generic dependency issue do not match the same defect." + "reason": "Qwen2.5-VL import error and generic dependency issue do not share a concrete root cause." }, { - "left": "issue:43994", - "right": "issue:46369", + "left": "issue:45440", + "right": "issue:45593", "accept": false, - "reason": "Quantized saving failure and SAM3 memory overflow are unrelated." + "reason": "DeepSeekV3 implementation divergence and D-FINE auxiliary-loss behavior are unrelated model bugs." }, { - "left": "issue:45440", - "right": "issue:45593", + "left": "issue:43994", + "right": "issue:46369", "accept": false, - "reason": "DeepSeekV3 implementation divergence and D-FINE auxiliary-loss behavior are different model bugs." + "reason": "SigLIP2 bad outputs and SAM3-LiteText memory overflow are distinct issues." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "BeitImageProcessorFast label reduction and Kosmos2.5 OCR index error are different image-processing failures." + "reason": "BeitImageProcessorFast label reduction and Kosmos2.5 long-OCR index error are different failures." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Chunked generation with compiled forward and Whisper empty-transcription failure are different generation paths." + "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-param dict conversion TypeError and missing decorator for rotary-pos-emb registration are unrelated." + "reason": "RoPE validation TypeError from a loaded list and missing-decorator crash on rotary-pos embedding registration are different bugs." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "BeitImageProcessorFast label reduction and LayoutLMv2Tokenizer NER/padding crashes are unrelated." + "reason": "BeitImageProcessorFast reduce_labels and LayoutLMv2Tokenizer NER/padding crashes are unrelated tokenizer/processor bugs." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "LayoutLMv3 error messaging and embedding reinitialization after resize are different bugs." + "reason": "LayoutLMv3 box error handling and output embedding reinitialization in post_init are unrelated." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Whisper empty transcription are unrelated." + "reason": "Qwen3.5 num_labels propagation and Whisper empty transcription after token alignment are unrelated." }, { - "left": "issue:43761", - "right": "issue:45859", + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "TP loading OOM for Qwen2-57B and resume_from_checkpoint key mismatch are distinct training/runtime issues." + }, + { + "left": "issue:44442", + "right": "issue:45072", "accept": false, - "reason": "CLIPVision hidden_states regression and Qwen3_5Moe CPU-offload incompatibility are different code paths." + "reason": "FastSpeech2ConformerTokenizer loading and bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel are unrelated." } ] }, @@ -9349,7 +9496,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:27957", "issue:29127", @@ -9360,10 +9507,12 @@ "issue:42175", "issue:42490", "issue:42915", + "issue:43232", "issue:43262", "issue:43316", "issue:43334", "issue:43381", + "issue:43388", "issue:43452", "issue:43525", "issue:43526", @@ -9371,7 +9520,6 @@ "issue:43575", "issue:43611", "issue:43646", - "issue:43701", "issue:43720", "issue:43723", "issue:43761", @@ -9390,26 +9538,25 @@ "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:44442|issue:45072", - "issue:43575|issue:43701", + "issue:43761|issue:45859", "issue:42915|issue:43931", "issue:42490|issue:43381", "issue:43575|issue:44387", "issue:43526|issue:44857", "issue:43334|issue:43531", "issue:44977|issue:45245", - "issue:43262|issue:44206", "issue:44464|issue:45491", + "issue:43262|issue:44206", "issue:27957|issue:29127", "issue:43452|issue:45588", "issue:44743|issue:44811", @@ -9428,222 +9575,211 @@ "issue:43316|issue:43525", "issue:42175|issue:46234", "issue:34689|issue:43611", - "issue:43720|issue:44442" + "issue:43720|issue:44442", + "issue:43232|issue:43388" ], - "prepared_review_unit_hash": "e0fd499cf4db813c151933d2712c5647c57a2a623de7505ae0864b2b3e23e002", + "prepared_review_unit_hash": "117e02343e060122143bce5c461caeb75962850f7654602043f0975d3f4b169b", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 35762, - "estimated_input_tokens": 8941, - "estimated_eval_tokens": 18138 + "serialized_chars": 36424, + "estimated_input_tokens": 9106, + "estimated_eval_tokens": 18468 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a mixed-topic similarity cluster, not a true duplicate set: it spans tokenizer/config API regressions, OOM/quantization issues, processor bugs, and unrelated model-loading failures. None of the soft pairs look like the same underlying bug/change closely enough to merge as duplicates. If one issue must serve as the anchor, the broad, high-activity quantization/offloading regression is the best representative.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 43873 is the broadest and most active open thread in this set, with multiple inbound references and a large discussion around quantization/offloading memory behavior, so it is the best anchor if a single issue must represent the cluster.", + "summary": "This is a heterogeneous set of Transformers issues spanning unrelated bugs (loading, generation, quantization, processors, config/API regressions). I don\u2019t see any convincing duplicate pairs; the soft edges are mostly same-subsystem, not same underlying defect.", + "confidence": 0.98, + "canonical_issue_reason": "Issue #43873 is the most discussion-heavy and broadly user-impacting item here, so it\u2019s the best representative if one issue must stand in for the set.", "canonical_pr_reason": null, - "best_issue_reason": "43873 is the most suitable representative issue because it is current, high-signal, and covers the widest user-facing regression surface in the cluster; however, it is not a duplicate of the other items.", + "best_issue_reason": "#43873 has the highest activity and inbound references, and its quantization/offloading regression is a prominent, high-impact user problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44442", - "right": "issue:45072", - "accept": false, - "reason": "Tokenizer loading failure vs dtype-mismatch inference bug; different code paths and symptoms." - }, - { - "left": "issue:43575", - "right": "issue:43701", + "left": "issue:43761", + "right": "issue:45859", "accept": false, - "reason": "OOM during tensor-parallel model load vs resume_from_checkpoint key mismatch; unrelated bugs." + "reason": "Both are model-related regressions, but one is CLIP hidden-states behavior and the other is CPU-offload compatibility in Qwen3_5Moe; different code paths and fixes." }, { "left": "issue:42915", "right": "issue:43931", "accept": false, - "reason": "FP8 config failure in Qwen3Moe vs Qwen3-VL weight-shape mismatch; different model families and failure modes." + "reason": "Same Qwen3 family, but one is a FineGrainedFP8Config failure and the other is a model shape-mismatch loading error for Qwen3-VL; not the same bug." }, { "left": "issue:42490", "right": "issue:43381", "accept": false, - "reason": "Local-path import warning vs gradient-checkpointing eval-mode restriction; no shared code-path." + "reason": "One is a local-path import warning, the other is a gradient-checkpointing eval-mode limitation; unrelated behaviors." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Both mention OOM, but one is TP load OOM and the other is int4 quantization memory growth; not the same bug." + "reason": "Both mention OOM, but the causes differ: tensor-parallel loading of a specific model vs increased reserved memory under int4 quantization." }, { "left": "issue:43526", "right": "issue:44857", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug vs LwDetrImageLoss AMP crash; unrelated subsystems." + "reason": "Different image-processing/model-loss bugs with no shared concrete failure mode." }, { "left": "issue:43334", "right": "issue:43531", "accept": false, - "reason": "Missing pad_token_id on Qwen3-VL config vs sliding_window recurrent-state reset in Qwen3-MoE; different defects." + "reason": "Both involve Qwen3 variants, but one is a missing config attribute during load and the other is a sliding-window state-reset bug." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "Flash-attention generation issue vs category-cardinality limit; not the same concrete bug." + "reason": "Completely different failures: flash-attention generation bug vs category-count overflow." }, { - "left": "issue:43262", - "right": "issue:44206", + "left": "issue:44464", + "right": "issue:45491", "accept": false, - "reason": "Audio chat-template sampling-rate default vs unsupported 'center' arg in feature extractor; different processor paths." + "reason": "Different generation/attention issues in different models; no shared code-path problem." }, { - "left": "issue:44464", - "right": "issue:45491", + "left": "issue:43262", + "right": "issue:44206", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency vs NaNs from mixed-length sliding-window padding; different failure mechanisms." + "reason": "Both are audio processor regressions, but one is a chat-template sampling-rate default and the other is an unsupported `center` argument in a feature extractor." }, { "left": "issue:27957", "right": "issue:29127", "accept": false, - "reason": "FlashAttention 2 support question vs LayoutLMv3 error-message clarity; unrelated issues." + "reason": "Unrelated: XLM-RoBERTa flash-attention support vs a LayoutLMv3 error-message clarity issue." }, { "left": "issue:43452", "right": "issue:45588", "accept": false, - "reason": "gguf_file from_pretrained breakage vs flash_attention s_aux=None crash; different loaders and code paths." + "reason": "Tokenizer/model loading with gguf is unrelated to a flash-attention `s_aux=None` crash." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Cache/recurrent-state reset bug vs Whisper batch_decode skip_special_tokens bug; unrelated functionality." + "reason": "Different subsystems and symptoms: Qwen3.5 cache state reset vs Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:27957", "right": "issue:33290", "accept": false, - "reason": "Model/attention support request vs deepspeed Adafactor OOM; no shared underlying bug." + "reason": "No concrete overlap beyond being user-facing bugs; flash-attention support and deepspeed Adafactor OOM are unrelated." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Broad v5 custom-init regression vs patchtsmixer-specific post_init policy change; too different and too model-specific." + "reason": "Both touch model initialization, but one is a broad v5 custom-model breakage and the other is a specific `patchtsmixer` post_init policy issue." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "Optimizer OOM in deepspeed vs Llama 3.2 Vision loading breakage; unrelated failures." + "reason": "Different problems and models; shared OOM/loading wording is too broad to merge." }, { "left": "issue:44062", "right": "issue:45907", "accept": false, - "reason": "AddedToken special-arg duplication vs list_repo_files revision forwarding; unrelated APIs." + "reason": "Tokenization `AddedToken` argument duplication is unrelated to `list_repo_files` revision forwarding." }, { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "OOM in training/optimizer stack vs trainer token-count tracking not updating; different code paths." + "reason": "OOM under deepspeed Adafactor vs trainer token-count bookkeeping; no shared bug." }, { "left": "issue:33290", "right": "issue:35141", "accept": false, - "reason": "OOM vs embedding reinitialization after resize; unrelated bugs." + "reason": "Both mention training/model internals, but one is optimizer OOM and the other is embedding reinitialization in `post_init`." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction vs quantization/offloading memory regression; different concrete problems." + "reason": "Both are training/runtime behavior issues, but gradient-checkpointing-in-eval and quantization offloading are distinct code paths." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "Tokenizer load failure vs Perceiver interpolation/non-default resolution failure; different model classes and symptoms." + "reason": "Tokenizer loading failure vs image-classification resolution/interpolation failure; unrelated." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe sys.modules access vs DeepSeekV3 implementation divergence; unrelated implementation issues." + "reason": "Unsafe `sys.modules` access and DeepSeekV3 implementation divergence are different defects." }, { "left": "issue:43723", "right": "issue:45479", "accept": false, - "reason": "AutoTokenizer loading issue in v5 vs zero-loss sequence-classification regression; different failures." + "reason": "Tokenizer-loading regression and single-label classification zero-loss bug are unrelated." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "CLIP hidden_states regression vs ModelOutput key assignment bug; not the same code-path." - }, - { - "left": "issue:42175", - "right": "issue:46234", - "accept": false, - "reason": "Packaging/backends selection issue vs missing import_utils symbol; unrelated." + "reason": "CLIP hidden-states output regression is unrelated to `ModelOutput` key assignment semantics." }, { - "left": "issue:34689", - "right": "issue:43611", + "left": "issue:42490", + "right": "issue:44743", "accept": false, - "reason": "Specific model-loading failure vs base_model_prefix loading regression; same broad area, but not the same bug." + "reason": "Import-path warning and Qwen recurrent-state reset are unrelated." }, { - "left": "issue:43720", - "right": "issue:44442", + "left": "issue:43316", + "right": "issue:43525", "accept": false, - "reason": "Accelerate loading/unpacking bug for BitNet vs tokenizer loading failure; unrelated." + "reason": "Both concern config API differences, but they are different model families and different missing attributes." }, { - "left": "issue:44743", - "right": "issue:44811", + "left": "issue:42175", + "right": "issue:46234", "accept": false, - "reason": "Recurrent-state reset bug vs Whisper decoding option bug; unrelated." + "reason": "Packaging/backend dependency issue vs missing utility export; not the same underlying change." }, { - "left": "issue:43316", - "right": "issue:43525", + "left": "issue:34689", + "right": "issue:43611", "accept": false, - "reason": "Both touch config API shape, but one is a broad Gemma3TextConfig discrepancy and the other is a specific Llama4 pad_token_id AttributeError; not clearly the same underlying bug." + "reason": "Both are model-loading regressions, but they affect different versions and different failure mechanisms." }, { - "left": "issue:44077", - "right": "issue:44079", + "left": "issue:43720", + "right": "issue:44442", "accept": false, - "reason": "PatchTSMixer post_init policy issue vs ModelOutput key handling bug; unrelated." + "reason": "BitNet packed-weight loading under accelerate is unrelated to tokenizer loading for FastSpeech2ConformerTokenizer." }, { - "left": "issue:42175", - "right": "issue:46234", + "left": "issue:43232", + "right": "issue:43388", "accept": false, - "reason": "Package install backend omission vs missing utility export; different issues." + "reason": "Generation kwargs synchronization and metric-gathering label truncation are different pipeline bugs." } ] }, @@ -9665,17 +9801,19 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ + "issue:16998", + "issue:22355", "issue:25251", + "issue:27957", "issue:28282", + "issue:32090", "issue:42617", "issue:42907", "issue:43065", - "issue:43232", "issue:43295", "issue:43316", - "issue:43388", "issue:43493", "issue:43550", "issue:43575", @@ -9711,222 +9849,215 @@ "issue:45803", "issue:46178", "issue:46202", - "issue:46365", "issue:46421", "issue:46424" ], "soft_pairs": [ - "issue:43232|issue:43388", - "issue:43761|issue:44188", "issue:43931|issue:45216", + "issue:43761|issue:44188", "issue:44077|issue:44479", "issue:44987|issue:46178", - "issue:44464|issue:46421", "issue:43493|issue:43827", + "issue:44464|issue:46421", "issue:44442|issue:45803", - "issue:44479|issue:45290", "issue:43873|issue:44871", "issue:42907|issue:45698", "issue:43749|issue:43931", "issue:43643|issue:44913", + "issue:44479|issue:45290", "issue:43316|issue:45759", "issue:44855|issue:45137", "issue:44190|issue:44496", "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:43065|issue:44617", - "issue:43575|issue:43742" + "issue:27957|issue:32090" ], - "prepared_review_unit_hash": "28cac491a22c01e0e042c293f48473138f2105430c60a11e9db35ace58732b1a", + "prepared_review_unit_hash": "2cc4cc5c15341c228229963ce08a4c0f1edff785f61851a42f560018888c91c1", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 34877, - "estimated_input_tokens": 8720, - "estimated_eval_tokens": 17696 + "serialized_chars": 35305, + "estimated_input_tokens": 8827, + "estimated_eval_tokens": 17910 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is heterogeneous: the soft pairs mostly connect unrelated issues that only share broad Transformers vocabulary (generation, loading, config, caching, offload, etc.). I found no true duplicate pair, so the safest interpretation is that this is not a real duplicate cluster.", - "confidence": 0.67, - "canonical_issue_reason": "Issue 45698 is the strongest standalone issue in this set: it is recent, active, open, and describes a concrete custom-module save/load regression with broad user impact. That said, the cluster itself is too mixed to indicate a true canonical duplicate.", + "summary": "This cluster is highly heterogeneous: the paired items span unrelated model-loading, configuration, pipeline, cache, compilation, and docs issues. I do not see any true duplicate groups here; all soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No true canonical duplicate target stands out. If forced to pick a representative, issue 45698 has the most discussion/inbound references, but it is not a duplicate hub for the rest of this cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45698 is the best overall issue candidate for triage because it is the most actionable and currently active among the listed issues; however, it should not be treated as representative of the whole cluster beyond that.", + "best_issue_reason": "Issue 45698 is the strongest representative only by activity and recency, not because it matches the others. It is the most visible issue in an otherwise unrelated set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43232", - "right": "issue:43388", + "left": "issue:43931", + "right": "issue:45216", "accept": false, - "reason": "Generation kwargs handling after sync_gpus and gather_for_metrics label truncation are different bugs in different code paths." + "reason": "Different failures: Qwen3-VL weight-shape mismatch on load vs a Qwen3.5 save_pretrained regression." }, { "left": "issue:43761", "right": "issue:44188", "accept": false, - "reason": "CLIP hidden_states regression and compile-time attention-kernel divergence are unrelated model/runtime problems." - }, - { - "left": "issue:43931", - "right": "issue:45216", - "accept": false, - "reason": "A Qwen3-VL shape-mismatch load failure is not the same as a Qwen3.5 save_pretrained checkpoint regression." + "reason": "Different code paths: CLIPVisionModel hidden_states bug vs attention-kernel divergence under torch.compile." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "PatchTSMixer post_init validation and Qwen video-input regression are different model-level issues." + "reason": "Unrelated: patchtsmixer post_init validation vs a v5.3.0 video-input regression in Qwen/VL models." }, { "left": "issue:44987", "right": "issue:46178", "accept": false, - "reason": "Loading a specific model failing under transformers>=5.1 and a backend mapping error for tensorflow_text are unrelated." - }, - { - "left": "issue:44464", - "right": "issue:46421", - "accept": false, - "reason": "Chunked generation inconsistency with compiled forward and cache lazy-init recompilation are only loosely related by compilation." + "reason": "Different problems: loading a specific model on transformers>=5.1.0 vs an unsupported tensorflow_text backend mapping error." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation drift and v5 docs still referencing pipeline() are different product areas." + "reason": "One is a SigLIP2 implementation mismatch; the other is stale docs referencing pipeline() after removals." }, { - "left": "issue:44442", - "right": "issue:45803", + "left": "issue:44464", + "right": "issue:46421", "accept": false, - "reason": "AutoTokenizer failing on a custom tokenizer and a bare-except bug in Fuyu tensor conversion are distinct issues." + "reason": "Both touch generation/compile, but one is incorrect outputs and the other is spurious recompilation from cache initialization." }, { - "left": "issue:44479", - "right": "issue:45290", + "left": "issue:44442", + "right": "issue:45803", "accept": false, - "reason": "A video-input regression and a chat-template crash on tool-call messages do not share the same root cause." + "reason": "Tokenizer loading failure vs tensor-conversion exception handling; not the same bug." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantization/offloading behavior and Gemma eos_token_id config mismatch are unrelated." + "reason": "Different features: quantization/offloading cache behavior vs Gemma eos_token_id config mismatch." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Saving dequantized model artifacts and loading the wrong custom module after save_pretrained are separate save/load failures." + "reason": "Saving a dequantized model is different from loading the wrong custom module after save_pretrained." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "Broken FSDP CPU RAM efficient loading and a Qwen3-VL weight-shape mismatch are different loading paths and failures." + "reason": "FSDP CPU-RAM-efficient loading failure vs Qwen3-VL weight-shape mismatch." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Missing fields from trust_remote_code AutoConfig and rotary_pct not persisting on reload are different config serialization bugs." + "reason": "Missing fields with trust_remote_code on AutoConfig is not the same as rotary_pct resetting on reload." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video-input regression across models vs chat-template crashes on assistant tool-call messages." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Gemma3TextConfig API discrepancy and AutoModelForCausalLM.from_config not unwrapping text_config are related only at a broad API level, not the same bug." + "reason": "Both involve config APIs, but Gemma3TextConfig discrepancy and Qwen multimodal text_config unwrapping are distinct bugs." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "A Python 3.13 import parsing error and a DeepSpeed ZeRO3 deque IndexError are completely unrelated." - }, - { - "left": "issue:44190", - "right": "issue:44496", - "accept": false, - "reason": "Local dataset loading in an example script and an unrecognized model/config error are different failures." + "reason": "Python 3.13 import parsing error vs a DeepSpeed ZeRO3 deque IndexError; unrelated failures." }, { "left": "issue:43650", "right": "issue:46202", "accept": false, - "reason": "A placeholder-data issue and a Gemma pipeline roundtrip failure are not duplicates." + "reason": "Unrelated; one is a vague data request, the other is a Gemma pipeline roundtrip failure." }, { "left": "issue:45632", "right": "issue:46424", "accept": false, - "reason": "trust_remote_code cache-path collision and max_cache_length only applying to Quantized Cache are different cache-related behaviors." - }, - { - "left": "issue:45803", - "right": "issue:46365", - "accept": false, - "reason": "A bare except swallowing errors and an import conflict involving Trainer/datasets are unrelated." + "reason": "Cache-path collision for trust_remote_code is unrelated to max_cache_length behavior in quantized cache." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "Pipeline top_k output nesting and an AutoModel ImportError due to missing PyTorch are not the same underlying problem." + "reason": "Pipeline top_k output-shape change is unrelated to a missing PyTorch dependency import error." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "A processor/tokenizer regression and a pipeline v5 error-message issue are different regression symptoms." + "reason": "Processor/tokenizer regression is different from an incorrect translation-support error message after pipeline removals." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "Bamba torch.compile SDPA failure and resume_from_checkpoint key mismatch share no concrete code path." + "reason": "torch.compile + SDPA failure in Bamba is unrelated to a resume_from_checkpoint key mismatch." }, { "left": "issue:42617", "right": "issue:43575", "accept": false, - "reason": "3d_parallel.py runtime failure and Qwen2-57B tensor-parallel OOM are different large-model issues." + "reason": "3d_parallel.py runtime failure and Qwen2 TP OOM are different loading/execution problems." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU-offload device mismatch and checkpoint key mismatch are unrelated." + "reason": "CPU offload device mismatch is unrelated to checkpoint key mismatch." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "TP OOM while loading Qwen2-57B is not the same as a key error loading MobileLLM-125M." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Model_max_length question for DeBERTa-V3 is unrelated to a transformers.onnx import error." }, { "left": "issue:43065", "right": "issue:44617", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and Sam3Video CUDA OOM are not the same bug." + "reason": "Dummy Conv2d in Sam3PixelDecoder is a different bug from Sam3Video CUDA OOM." }, { - "left": "issue:43575", - "right": "issue:43742", + "left": "issue:27957", + "right": "issue:32090", "accept": false, - "reason": "Tensor-parallel OOM for Qwen2-57B and a MobileLLM load KeyError are different model-loading failures." + "reason": "XLMRoberta flash-attention support and a Trainer _gpu_broadcast_one NoneType error are unrelated." } ] }, @@ -9948,12 +10079,8 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ - "issue:16998", - "issue:22355", - "issue:27957", - "issue:32090", "issue:41950", "issue:43295", "issue:43316", @@ -9963,6 +10090,8 @@ "issue:43452", "issue:43475", "issue:43504", + "issue:43525", + "issue:43531", "issue:43540", "issue:43572", "issue:43646", @@ -9972,6 +10101,7 @@ "issue:43761", "issue:44077", "issue:44162", + "issue:44242", "issue:44479", "issue:44496", "issue:44514", @@ -9984,6 +10114,7 @@ "issue:44871", "issue:44877", "issue:44913", + "issue:44977", "issue:45127", "issue:45200", "issue:45406", @@ -9995,27 +10126,27 @@ "issue:46170", "issue:46178", "issue:46234", + "issue:46365", "issue:46369", "issue:46421" ], "soft_pairs": [ - "issue:16998|issue:22355", - "issue:43540|issue:44514", - "issue:27957|issue:32090", "issue:43316|issue:43716", - "issue:44610|issue:46369", + "issue:43540|issue:44514", "issue:43295|issue:44871", "issue:44871|issue:45584", "issue:43316|issue:44877", + "issue:44610|issue:46369", "issue:44162|issue:44792", - "issue:43404|issue:43441", "issue:44077|issue:44534", + "issue:43404|issue:43441", + "issue:45803|issue:46365", "issue:45803|issue:46170", "issue:44871|issue:46050", "issue:41950|issue:43475", "issue:43452|issue:44496", - "issue:43761|issue:44479", "issue:43646|issue:44913", + "issue:43761|issue:44479", "issue:43388|issue:43688", "issue:45127|issue:45200", "issue:43673|issue:44743", @@ -10024,199 +10155,217 @@ "issue:43504|issue:45698", "issue:43673|issue:46421", "issue:44734|issue:45406", - "issue:45542|issue:46234" + "issue:45542|issue:46234", + "issue:43525|issue:43531", + "issue:44792|issue:44977", + "issue:43761|issue:44242" ], - "prepared_review_unit_hash": "5df06c03705811b50b6e6ece28636d1d163983b827357881ebd8113345f94ea7", + "prepared_review_unit_hash": "edd3bfea7e291fa14a077a231fa7d0d0f4e0a1b684b4c754056bccc052cacaa8", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 34984, - "estimated_input_tokens": 8746, - "estimated_eval_tokens": 17748 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36170, + "estimated_input_tokens": 9043, + "estimated_eval_tokens": 18342 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose semantic grouping of Transformers regressions and model-specific bugs rather than true duplicates. The only safe canonical choice is a broad v5 regression issue; the soft pairs are generally too different in model, code path, or symptom to merge.", - "confidence": 0.47, - "canonical_issue_reason": "issue:43646 is the broadest representative: a general Transformers v5 breakage around custom model initialization, which matches the cluster\u2019s overall regression theme better than the narrower model-specific reports.", + "summary": "Most candidates are related by broad Transformers/multimodal regression themes, but they describe distinct model- or API-specific bugs. The only strong duplicate-like pair is the chunked-prefill cache regression (43673/46421).", + "confidence": 0.86, + "canonical_issue_reason": "issue:46421 is the clearest root-cause-focused description of the shared chunked-prefill cache lazy-initialization regression; issue:43673 reads like an earlier symptom report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best umbrella issue for this set because it is broad, version-level, and likely to attract related reports; however, the cluster is heterogeneous and not a strong duplicate set.", + "best_issue_reason": "issue:46421 is the best representative because it names the underlying cache-lazy-init problem and the triggering path more precisely than the sibling report.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:16998", - "right": "issue:22355", - "accept": false, - "reason": "Unrelated topics: DeBERTa model_max_length behavior vs an import error for transformers.onnx." - }, - { - "left": "issue:43540", - "right": "issue:44514", - "accept": false, - "reason": "Both involve multimodal Qwen processing, but one is video-input processing in Qwen3OmniMoe and the other is batched chat template padding in Qwen2.5-VL; different code paths and bugs." - }, - { - "left": "issue:27957", - "right": "issue:32090", - "accept": false, - "reason": "No shared underlying bug: FlashAttention support for XLM-Roberta vs a Trainer broadcast TypeError." - }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Different model families and failures: Gemma3TextConfig API mismatch vs Mistral-3 image preprocessor dtype mismatch." + "reason": "Related config/model mismatch theme, but different models and different failure modes." }, { - "left": "issue:44610", - "right": "issue:46369", + "left": "issue:43540", + "right": "issue:44514", "accept": false, - "reason": "Both are model-specific vision issues, but one is an input-size mismatch in OmDet-Turbo and the other is OOM in sam3-litetext; not the same fix." + "reason": "Both are Qwen video/processor bugs, but the concrete failures differ." }, { "left": "issue:43295", "right": "issue:44871", "accept": false, - "reason": "Processor/tokenizer regression in one issue vs eos_token_id config inconsistency in another; only loosely related to processors/configs." + "reason": "Processor-tokenizer regression vs Gemma eos-token-id config mismatch; not the same bug." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Different bugs: Gemma-3 eos token configuration vs Whisper empty-transcription generation after special-token alignment." + "reason": "Unrelated model/config issue vs Whisper empty-transcription failure." }, { "left": "issue:43316", "right": "issue:44877", "accept": false, - "reason": "Both concern config loading, but Gemma3TextConfig API discrepancy and granite_speech strict-config loading failure are distinct issues." + "reason": "Both involve config handling, but they affect different model families and code paths." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "Different subsystems: image-size mismatch vs SAM3 litetext memory overflow." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "Completely different models and symptoms: ESM2 breakage vs a failed janus image-generation test." + "reason": "Different models and symptoms; one is ESM2 breakage, the other a Janus image-generation test failure." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Unrelated bugs: optional post_init enforcement vs non-persistent buffer junk." }, { "left": "issue:43404", "right": "issue:43441", "accept": false, - "reason": "Same broad model family, but different concrete failures: untied lm_head weights vs FlashAttention failure in Ministral-3." + "reason": "Different Mistral/Ministral issues and different failure mechanisms." }, { - "left": "issue:44077", - "right": "issue:44534", + "left": "issue:45803", + "right": "issue:46365", "accept": false, - "reason": "Optional post_init misuse in patchtsmixer is unrelated to non-persistent buffer junk values." + "reason": "Both affect imports, but one is Fuyu tensor conversion and the other is Trainer/datasets import interference." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "Bare except swallowing interrupts in Fuyu conversion is unrelated to a Windows/ROCm torch.distributed import/type error." + "reason": "Bare except swallowing errors is unrelated to the Windows/ROCm torch.distributed import problem." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Different save/load regressions: Gemma-3 eos token mismatch vs sinq quantized model breaking after save_pretrained/from_pretrained." + "reason": "Gemma token-id config mismatch vs quantized-model save/load corruption." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Both are video-adjacent, but one is pipeline processor selection and the other is a missing attribute in Sam3 video encoder output." + "reason": "Both are video/SAM3-adjacent, but one is processor selection and the other is a missing encoder attribute." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "Both are from_pretrained/config loading problems, but one is gguf/custom class loading and the other is an unrecognized model_type in config.json." + "reason": "Different load-time failures: gguf path handling vs missing model_type in config.json." }, { - "left": "issue:43761", - "right": "issue:44479", + "left": "issue:43646", + "right": "issue:44913", "accept": false, - "reason": "Different regressions: CLIPVisionModel hidden_states handling vs Qwen video-input regression across multiple model variants." + "reason": "Broad v5 initialization breakage vs a specific GPTNeoX config reload bug." }, { - "left": "issue:43646", - "right": "issue:44913", + "left": "issue:43761", + "right": "issue:44479", "accept": false, - "reason": "General custom model initialization breakage vs a specific GPTNeoXConfig rotary_pct persistence bug; not the same underlying defect." + "reason": "Different regression surfaces: CLIP hidden_states output vs Qwen video-input generation." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "Training metric gathering dropping tuple labels is unrelated to auxiliary-loss normalization." + "reason": "Different training bugs: metric gathering truncation vs auxiliary-loss normalization." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge with extended vocab collapsing models is unrelated to Gemma 4 mm_token_type_ids defaults." + "reason": "Both involve Gemma-family fine-tuning, but the underlying bugs are unrelated." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both mention cache/recurrent state, but one is missing cache in chunked_prefill and the other is recurrent states resetting in a specific modular Qwen3.5 file; different bugs." + "reason": "Both mention cache/recurrent-state behavior, but different models and distinct code paths." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is unrelated to tensorflow_text backend mapping errors." + "reason": "StableLm pad-token config regression is unrelated to the tensorflow_text backend registration error." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "CUDA OOM in Sam3Video is unrelated to Whisper empty-transcription generation behavior." + "reason": "CUDA OOM in SAM3 video is not the same as Whisper empty-output generation failure." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Legacy field blocks Beit preset loading vs wrong custom module loaded after save_pretrained; both are loading issues but not the same code path." + "reason": "Pretrained preset loading with a legacy field is different from loading the wrong custom module after save_pretrained." }, { "left": "issue:43673", "right": "issue:46421", - "accept": false, - "reason": "Both involve chunked prefill caching, but one is a missing cache in v5.0.0 and the other is lazy-init causing spurious recompilation; related area, different bug." + "accept": true, + "reason": "Both describe the same chunked-prefill cache initialization regression; one reports missing cache, the other recompilation from lazy init." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "KV-cache continuation tensor indexing crash in serve is unrelated to Gemma4Processor missing _tokenizer." + "reason": "Serve KV-cache indexing bug and Gemma4Processor missing _tokenizer are different failures." }, { "left": "issue:45542", "right": "issue:46234", "accept": false, - "reason": "TensorBoard/TensorFlow backend installation issue is unrelated to a missing import utility helper for falcon mamba ssm." + "reason": "TensorBoard/TensorFlow backend setup issue vs missing import_utils availability helper." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Llama4 pad_token_id absence and Qwen3-MoE sliding_window behavior are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure vs Qwen3.5 flash-attention generation issue." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden_states regression is unrelated to load-balancing loss logic." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the soft-edge verdicts stay conservative overall. The 43673/46421 pair is the only clearly duplicate-like match and is supported by the issue titles, while the other rejects are appropriately cautious given the available evidence." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -10231,7 +10380,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:25251", "issue:27957", @@ -10239,20 +10388,19 @@ "issue:41720", "issue:42490", "issue:43065", - "issue:43525", - "issue:43531", + "issue:43116", + "issue:43335", "issue:43540", "issue:43550", "issue:43644", - "issue:43761", "issue:43827", "issue:43931", "issue:43976", "issue:43994", "issue:44060", "issue:44077", + "issue:44112", "issue:44190", - "issue:44242", "issue:44462", "issue:44464", "issue:44479", @@ -10261,13 +10409,13 @@ "issue:44610", "issue:44617", "issue:44734", + "issue:44737", "issue:44792", "issue:44855", "issue:44877", "issue:44913", "issue:44933", "issue:44964", - "issue:44977", "issue:45127", "issue:45245", "issue:45357", @@ -10278,236 +10426,244 @@ "issue:45588", "issue:45600", "issue:45632", + "issue:45803", "issue:45907", "issue:46234", "issue:46369" ], "soft_pairs": [ - "issue:43525|issue:43531", - "issue:43761|issue:44242", "issue:45381|issue:46369", - "issue:44792|issue:44977", "issue:45563|issue:45584", - "issue:43931|issue:44060", "issue:44792|issue:45600", + "issue:43931|issue:44060", "issue:44534|issue:44913", "issue:44190|issue:44964", - "issue:44479|issue:44734", "issue:35141|issue:42490", "issue:44855|issue:45588", - "issue:44617|issue:46369", + "issue:44479|issue:44734", "issue:43827|issue:43994", + "issue:44617|issue:46369", "issue:41720|issue:45245", "issue:43065|issue:44610", - "issue:44521|issue:45381", "issue:44464|issue:45563", + "issue:44521|issue:45381", "issue:44933|issue:45362", "issue:43644|issue:44913", "issue:43976|issue:44855", - "issue:43550|issue:44190", "issue:44077|issue:44877", - "issue:43540|issue:44734", + "issue:43550|issue:44190", "issue:45127|issue:45357", "issue:44462|issue:45907", + "issue:43540|issue:44734", "issue:25251|issue:27957", - "issue:45632|issue:46234" + "issue:45632|issue:46234", + "issue:44112|issue:45600", + "issue:43335|issue:44190", + "issue:44737|issue:45803", + "issue:43116|issue:44190" ], - "prepared_review_unit_hash": "805253767e368899cb6e8d0a4ce2db46a9761e97a7d4eef18d4aba13ee2aee88", + "prepared_review_unit_hash": "2ab07b41f6262870ff8e13c79b76bc2dbc0b745cd1313cf5c4a45d3ca28cf724", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 28, - "serialized_chars": 35893, - "estimated_input_tokens": 8974, - "estimated_eval_tokens": 18204 + "soft_pair_count": 29, + "serialized_chars": 36351, + "estimated_input_tokens": 9088, + "estimated_eval_tokens": 18432 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly unrelated Transformers bug reports rather than one duplicate cluster. None of the soft pairs describe the same underlying bug or change closely enough to merge; they usually differ by model, code path, or bug type.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 44521 is the best representative issue in this set: it is open, has the strongest cross-references, and describes a concrete multimodal regression with clear downstream impact.", + "summary": "All candidate soft edges are false positives: the items span unrelated bugs, configs, model families, and CI/doc issues rather than one duplicate cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the set spans unrelated model bugs, config problems, docs, and CI failures, and none of the soft pairs look like the same underlying defect.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44521 stands out as the strongest standalone bug report in the group, with multiple inbound references and active discussion, making it the most useful anchor issue here.", + "best_issue_reason": "No global best issue; there is no coherent duplicate cluster here, so no issue is a strong canonical representative.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43525", - "right": "issue:43531", - "accept": false, - "reason": "Different models and different failure modes: missing pad_token_id on Llama4Config vs sliding_window behavior in Qwen3-MoE." - }, - { - "left": "issue:43761", - "right": "issue:44242", - "accept": false, - "reason": "Unrelated regressions: CLIPVision hidden_states handling vs MoE load-balancing loss when router logits are disabled." - }, { "left": "issue:45381", "right": "issue:46369", "accept": false, - "reason": "Both are model-specific multimodal bugs, but one is a qwen2.5-vl video position-id regression and the other is Sam3-litetext memory overflow." - }, - { - "left": "issue:44792", - "right": "issue:44977", - "accept": false, - "reason": "Different code paths and symptoms: Janus image-generation test failure vs Qwen3.5 flash-attention generation issue." + "reason": "Different models and failures: Qwen2.5-VL video position ids vs Sam3LiteText memory overflow." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "A stale warning about num_return_sequences is not the same as Whisper failing on empty transcription after token alignment." + "reason": "Both involve generation, but one is a stale warning in paged generate and the other is Whisper failing on empty transcription." }, { - "left": "issue:43931", - "right": "issue:44060", + "left": "issue:44792", + "right": "issue:45600", "accept": false, - "reason": "One is a weight-shape mismatch on load, the other is an incorrect tied-weights warning involving a different Qwen3-Next parameter." + "reason": "Janus image-generation test failure and Sam3LiteText auto_mappings CI break are unrelated code paths." }, { - "left": "issue:44792", - "right": "issue:45600", + "left": "issue:43931", + "right": "issue:44060", "accept": false, - "reason": "A failing Janus generation test and a broken auto_mappings reference to removed Sam3LiteText configs are distinct issues." + "reason": "Different Qwen variants with different symptoms: weight-shape mismatch vs an incorrect tied-weights warning." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Generic non-persistent buffer initialization bug vs a GPTNeoX config field not persisting on reload." + "reason": "Non-persistent buffer junk fill is unrelated to GPTNeoX rotary_pct reverting on reload." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Local dataset loading in a training script is unrelated to loading a specific multimodal model with latest transformers." - }, - { - "left": "issue:44479", - "right": "issue:44734", - "accept": false, - "reason": "Video-input regression in Qwen multimodal models vs KV-cache continuation crash in transformers serve are different failures." + "reason": "Local dataset loading script failure and Phi-4 multimodal model loading error are distinct problems." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Embedding reinitialization during post_init is unrelated to the import warning caused by a trailing slash in a local model path." + "reason": "Embedding reinitialization after resize is unrelated to the import warning from a local path ending with a slash." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Python 3.13 import parsing failure from torch.jit.script formatting is unrelated to flash_attention.py crashing on s_aux=None." + "reason": "Python 3.13 JIT parsing/IndentationError is unrelated to a flash_attention AttributeError on s_aux=None." }, { - "left": "issue:44617", - "right": "issue:46369", + "left": "issue:44479", + "right": "issue:44734", "accept": false, - "reason": "Both mention memory pressure, but they affect different models and likely different root causes: Sam3Video OOM vs Sam3-litetext memory overflow." + "reason": "Qwen VL video regression and serve KV-cache continuation indexing are different bugs in different flows." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs still referencing pipeline() after removals is not the same as a SigLIP2 pipeline producing nonsensical outputs." + "reason": "Docs still referencing pipeline after removals is not the same as nonsensical SigLIP2 outputs with pipeline/AutoModel." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Both mention memory, but Sam3Video OOM and Sam3LiteText memory overflow are different model-specific issues." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "CUDA assert during Qwen3 auto device mapping is unrelated to the 2^24 category limit error." + "reason": "Qwen3 device-mapping cudaErrorAssert and a categories-limit RuntimeError are unrelated." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "A dummy Conv2d in Sam3PixelDecoder is a model-internal placeholder issue, not the OmDet-Turbo processor/model input-size mismatch." + "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo processor size mismatch are different model/component bugs." }, { - "left": "issue:44521", - "right": "issue:45381", + "left": "issue:44464", + "right": "issue:45563", "accept": false, - "reason": "Both are multimodal, but one is all-zero assistant masks in apply_chat_template and the other is incorrect vision_position_ids in qwen2.5-vl video input processing." + "reason": "Compiled-forward chunked generation inconsistency is unrelated to a stale warning about num_return_sequences." }, { - "left": "issue:44464", - "right": "issue:45563", + "left": "issue:44521", + "right": "issue:45381", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is a different bug from a stale generation warning." + "reason": "All-zero assistant masks in multimodal chat templates is a different bug from wrong Qwen2.5-VL video position ids." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "A missing import in image_utils is unrelated to Qwen3.5-35B crashing in chat generation." + "reason": "A nonexistent import from image_utils is unrelated to Qwen3.5 chat crashes." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "A broad v5 buffer initialization regression is not the same as GPTNeoX rotary_pct being lost on config reload." + "reason": "Non-persistent buffer junk fill and GPTNeoX rotary_pct reload behavior are separate config/runtime issues." }, { "left": "issue:43976", "right": "issue:44855", "accept": false, - "reason": "General Python version incompatibility does not match a specific Python 3.13 IndentationError caused by a decorator/comment parsing edge case." - }, - { - "left": "issue:43550", - "right": "issue:44190", - "accept": false, - "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to local dataset loading in an image-classification training script." + "reason": "These are different Python-version failures with different root causes and models." }, { "left": "issue:44077", "right": "issue:44877", "accept": false, - "reason": "Optional post_init in patchtsmixer and strict config blocking granite_speech are different model/config validation issues." + "reason": "PatchTSTMixer post_init policy and strict granite_speech config loading are unrelated config issues." }, { - "left": "issue:43540", - "right": "issue:44734", + "left": "issue:43550", + "right": "issue:44190", "accept": false, - "reason": "Processing video inputs in Qwen3OmniMoe is a different problem from server-side KV-cache continuation indexing." + "reason": "Bamba torch.compile/SDPA failure is unrelated to loading a local dataset in an example script." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "LoRA merge collapse with extended vocab and incorrect visual encoder keys in save_pretrained are separate serialization/training bugs." + "reason": "LoRA merge with extended vocab causing collapse is not the same as Qwen3.5 save_pretrained writing wrong visual encoder keys." }, { "left": "issue:44462", "right": "issue:45907", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to list_repo_files failing to forward revision in CI." + "reason": "AutoTokenizer ignoring tokenizer.json and list_repo_files not forwarding revision are unrelated hub/tokenizer behaviors." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and serve KV-cache tensor indexing are different failures." }, { "left": "issue:25251", "right": "issue:27957", "accept": false, - "reason": "Pipeline top_k nesting behavior and XLMRoberta Flash Attention 2 support are unrelated." + "reason": "Pipeline top_k output nesting and XLMRoberta Flash Attention 2 are unrelated bugs." }, { "left": "issue:45632", "right": "issue:46234", "accept": false, - "reason": "A cache-path collision for trust_remote_code is unrelated to a missing import-utils helper function." + "reason": "Local trust_remote_code cache collisions and a missing import_utils helper are unrelated." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "A stale GraniteSpeech test and Sam3LiteText auto_mappings references are different CI breakages." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformersConfig sparse-layer creation and local dataset loading are unrelated." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet device placement in torch.arange and a bare except swallowing errors are unrelated code defects." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Multi-label classification returning empty results and local dataset loading are different example-script issues." } ] }, @@ -10529,7 +10685,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:27957", "issue:35141", @@ -10538,27 +10694,28 @@ "issue:41628", "issue:41950", "issue:42175", - "issue:43116", + "issue:42907", "issue:43295", - "issue:43335", "issue:43404", + "issue:43441", "issue:43479", + "issue:43526", + "issue:43531", "issue:43643", "issue:43650", "issue:43673", "issue:43854", + "issue:43901", "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", @@ -10572,7 +10729,6 @@ "issue:45584", "issue:45600", "issue:45632", - "issue:45803", "issue:46082", "issue:46121", "issue:46123", @@ -10581,10 +10737,6 @@ "issue:46424" ], "soft_pairs": [ - "issue:44112|issue:45600", - "issue:43335|issue:44190", - "issue:44737|issue:45803", - "issue:43116|issue:44190", "issue:45276|issue:45584", "issue:43295|issue:45276", "issue:44938|issue:45542", @@ -10598,8 +10750,8 @@ "issue:44246|issue:45003", "issue:36246|issue:38617", "issue:41950|issue:44016", - "issue:44186|issue:44734", "issue:43673|issue:46424", + "issue:44186|issue:44734", "issue:43479|issue:45375", "issue:44683|issue:45446", "issue:43643|issue:44496", @@ -10609,210 +10761,207 @@ "issue:44112|issue:44964", "issue:46123|issue:46369", "issue:27957|issue:35141", - "issue:43976|issue:44987" + "issue:43976|issue:44987", + "issue:43531|issue:45600", + "issue:42907|issue:43441", + "issue:43526|issue:43901" ], - "prepared_review_unit_hash": "03a358bc17d297e717c933020a0bac8d2adf2eacf61b079c5f96688abbc1ccd0", + "prepared_review_unit_hash": "687c135be617a9cb4509f5c1829160c40ff3eb313049d13bf2ee2ed6811a5456", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36512, - "estimated_input_tokens": 9128, - "estimated_eval_tokens": 18512 + "soft_pair_count": 28, + "serialized_chars": 36064, + "estimated_input_tokens": 9016, + "estimated_eval_tokens": 18288 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous: a few weakly similar import/cache/config issues, but no pair looks like the same underlying bug or change. I\u2019d use the resize-token-embeddings issue as the best representative of the only clearly coherent bug family here.", - "confidence": 0.9, - "canonical_issue_reason": "issue:35141 is the clearest, most general statement of a concrete bug in the resize_token_embeddings/post_init path; it best represents the only strong bug family in this set.", + "summary": "These items are mostly unrelated Transformers issues with only superficial overlap around model loading, config handling, cache, or attention subsystems. I don\u2019t see a true duplicate cluster here, so there is no good canonical issue/PR anchor.", + "confidence": 0.93, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:35141 is the most broadly applicable and well-scoped issue among the candidates, with a concrete reproducible failure and a model-agnostic fix path.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44112", - "right": "issue:45600", - "accept": false, - "reason": "Different failures: GraniteSpeech CI device-override test vs removed Sam3LiteText auto-mapping references." - }, - { - "left": "issue:43335", - "right": "issue:44190", - "accept": false, - "reason": "Unrelated code paths: SwitchTransformers sparse-layer config vs image-classification script local dataset loading." - }, - { - "left": "issue:44737", - "right": "issue:45803", - "accept": false, - "reason": "Different bugs: XLNet CPU tensor allocation vs FuyuBatchFeature bare except swallowing errors." - }, - { - "left": "issue:43116", - "right": "issue:44190", - "accept": false, - "reason": "Both touch example/pipeline scripts, but one is multilabel classification output handling and the other is local dataset loading for image classification." - }, { "left": "issue:45276", "right": "issue:45584", "accept": false, - "reason": "Unrelated model behaviors: Gemma4 embedding resize vs Whisper empty-transcription generation." + "reason": "Different bugs: Gemma4 resize_token_embeddings vs Whisper empty-transcription generation after align_special_tokens." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Different regressions in different paths: processor/tokenizer image handling vs resize_token_embeddings output-embedding updates." + "reason": "Different code paths: processor.tokenizer/image-input regression vs Gemma4 embedding resizing/output embedding behavior." }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Both are environment/import issues, but one is Python 3.14 load failure and the other is missing TensorFlow backend wiring." + "reason": "Unrelated failures: Python 3.14 import/load regression vs missing TensorFlow backend when installing torch extras." }, { "left": "issue:46082", "right": "issue:46121", "accept": false, - "reason": "Distinct config-validation bugs: Llama head_dim constraints vs rope-validation list typing." + "reason": "Different config validation problems: Llama head_dim rules vs JSON-loaded rope ignore-keys type handling." }, { "left": "issue:45632", "right": "issue:46421", "accept": false, - "reason": "Both mention cache, but one is trust_remote_code cache-path collision and the other is chunked-prefill recompilation behavior." + "reason": "Both mention cache, but one is a trust_remote_code cache-path collision and the other is lazy cache init causing recompilation." }, { "left": "issue:45245", "right": "issue:45584", "accept": false, - "reason": "Completely different symptoms and subsystems: category-limit runtime error vs Whisper empty transcription handling." + "reason": "Different issues: category-count limit runtime error vs Whisper empty transcription generation failure." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but for different missing symbols and different modules." + "reason": "Both are import errors, but for different symbols and modules: Qwen2_5_VLImageProcessor vs AutoImageProcessor." }, { "left": "issue:42175", "right": "issue:44279", "accept": false, - "reason": "Dependency/backend packaging issue vs generic dependency issue; too broad and not the same defect." + "reason": "One is an extras/backend packaging complaint; the other is a vague dependency issue without the same concrete failure." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Different model-saving/training bugs: tied lm_head in Mistral3 vs Qwen3.5 checkpoint serialization." + "reason": "Different bugs: untied lm_head weights in Mistral3 vs incorrect checkpoint saving for Qwen3.5." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "Malformed issue vs audio feature extraction bug; no shared underlying defect." + "reason": "No relation: placeholder 'ADD THE DATA' issue vs _torch_extract_fbank_features implementation bug." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import-time slowness vs unsafe sys.modules access are different problems in different code paths." + "reason": "Different concerns: import slowdowns vs unsafe sys.modules access in modeling_utils." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Both are missing-export import errors, but they concern different symbols and modules." + "reason": "Both are import errors, but they refer to different removed/missing names and separate fixes." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Video pipeline processor-selection bug vs notebook syntax error; unrelated." + "reason": "Pipeline bug vs notebook syntax error; no shared underlying code path." }, { - "left": "issue:44186", - "right": "issue:44734", + "left": "issue:43673", + "right": "issue:46424", "accept": false, - "reason": "Tokenizer padding/NER crash vs serve KV-cache continuation indexing bug; different subsystems." + "reason": "Both are cache-related, but one is missing GenerationMixin cache in chunked prefill and the other is max_cache_length handling for quantized cache." }, { - "left": "issue:43673", - "right": "issue:46424", + "left": "issue:44186", + "right": "issue:44734", "accept": false, - "reason": "Both cache-related, but one is missing GenerationMixin cache in chunked prefill and the other is quantized-cache max length handling." + "reason": "Tokenizer padding/truncation crash vs serve KV-cache continuation tensor indexing bug; unrelated paths." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Both are config-field issues, but they affect different models and different fields/serialization behavior." + "reason": "Different model-config omissions: Phi4 multimodal defaults vs Qwen3_5MoeVisionConfig missing field." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Related to flex_attention, but one is a compile failure on torch>=2.9 and the other is a version-check import guard bug." + "reason": "Both touch flex_attention, but one is a torch>=2.9 compiled failure and the other is a version-check/import guard bug." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "Different auto-config problems: missing fields with trust_remote_code vs unrecognized model missing model_type." + "reason": "Different load failures: missing fields with trust_remote_code vs unrecognized model/config for Olmo-Hybrid." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Distinct model-loading failures: Phi-4-multimodal regression vs stale Sam3LiteText auto-mapping references." + "reason": "Different model-loading regressions: Phi-4 multimodal loading vs stale auto_mappings reference to removed Sam3LiteText configs." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Different models and failure modes, both load issues but not the same bug." + "reason": "Different models and failures; both loading-related but not the same underlying bug." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "Notebook syntax error vs audio feature extraction bug; unrelated." + "reason": "Notebook syntax error and audio feature extraction bug are unrelated." }, { "left": "issue:44112", "right": "issue:44964", "accept": false, - "reason": "CI test failure vs Phi-4-multimodal model loading regression; unrelated." + "reason": "CI test flake in GraniteSpeech vs Phi-4 multimodal model-loading issue; no common code path." }, { "left": "issue:46123", "right": "issue:46369", "accept": false, - "reason": "MaskGenerationPipeline partial-batch bug vs sam3-litetext memory overflow; different symptoms and code paths." + "reason": "Different problems: final partial batch dropping results in MaskGenerationPipeline vs memory overflow in sam3-litetext." }, { "left": "issue:27957", "right": "issue:35141", "accept": false, - "reason": "Different areas: Flash Attention support for XLM-Roberta vs resize-token-embeddings reinitialization." + "reason": "Unrelated: XLMRoberta FlashAttention2 support vs output embedding reinitialization after resizing token embeddings." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python-version compatibility bug vs specific model load failure on 5.1.0; not the same underlying issue." + "reason": "Environment/version compatibility issue vs a specific model loading failure; not the same bug." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Different models and symptoms: Qwen3-MoE sliding_window vs Sam3LiteText auto_mappings cleanup." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Same model family only at a broad level, but one is save_pretrained on dequantized Ministrals/Devstrals and the other is FlashAttention failure in Ministral-3." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "Completely different areas: BeitImageProcessorFast label reduction vs TextClassificationPipeline docs mismatch." } ] }, @@ -10834,18 +10983,16 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:28282", "issue:35141", "issue:38175", - "issue:42907", + "issue:41950", "issue:43295", "issue:43299", "issue:43381", "issue:43441", - "issue:43526", - "issue:43531", "issue:43577", "issue:43650", "issue:43673", @@ -10858,17 +11005,19 @@ "issue:44242", "issue:44279", "issue:44336", + "issue:44351", "issue:44368", "issue:44496", - "issue:44521", "issue:44617", + "issue:44779", "issue:44792", "issue:44933", + "issue:45020", "issue:45042", "issue:45103", "issue:45200", "issue:45216", - "issue:45325", + "issue:45278", "issue:45335", "issue:45357", "issue:45372", @@ -10876,8 +11025,8 @@ "issue:45584", "issue:45588", "issue:45593", - "issue:45600", "issue:45759", + "issue:45812", "issue:46082", "issue:46209", "issue:46234", @@ -10885,231 +11034,231 @@ "issue:46424" ], "soft_pairs": [ - "issue:43531|issue:45600", - "issue:42907|issue:43441", - "issue:43526|issue:43901", "issue:43295|issue:45335", "issue:44792|issue:46424", "issue:28282|issue:35141", "issue:44496|issue:45042", "issue:46082|issue:46399", - "issue:43441|issue:45588", "issue:45372|issue:46234", + "issue:43441|issue:45588", "issue:44016|issue:44062", "issue:43299|issue:45216", "issue:43901|issue:44496", - "issue:43824|issue:44933", "issue:43673|issue:44336", + "issue:43824|issue:44933", "issue:43381|issue:45593", "issue:43976|issue:44279", - "issue:44075|issue:44368", "issue:44496|issue:45759", + "issue:44075|issue:44368", "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:44279|issue:45405", - "issue:43299|issue:45357" + "issue:43299|issue:45357", + "issue:41950|issue:45020", + "issue:43295|issue:44779", + "issue:44779|issue:45812", + "issue:44351|issue:45278" ], - "prepared_review_unit_hash": "c41aab92e72b94ff3a60708eb49be9d7bc1c8bb43e2835b9fa4c031fc5e1d436", + "prepared_review_unit_hash": "bb3d9240456b978427133d49c400dfcf933abfae65aae03791035c2613023802", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35662, - "estimated_input_tokens": 8916, - "estimated_eval_tokens": 18088 + "serialized_chars": 35586, + "estimated_input_tokens": 8897, + "estimated_eval_tokens": 18050 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated bug reports across different models, loaders, and utilities. None of the soft pairs look like the same underlying bug/change, so I rejected them all. No PRs are present in this cluster.", - "confidence": 0.93, - "canonical_issue_reason": "issue:44521 is the most active/open item in the set and has the strongest external signal (multiple inbound references and discussion), so it is the best representative issue for the cluster even though the cluster is not a true duplicate set.", + "summary": "This cluster is mostly a loose collection of unrelated Transformers v5 regression reports, centered on model loading, tokenizers, processors, and import/export breakages. I do not see any true duplicate pairs among the soft edges.", + "confidence": 0.41, + "canonical_issue_reason": "issue:45020 is the broadest umbrella-style report in the set: it explicitly describes recent-version breakage affecting models that use `remote_code`, which best matches the overall theme of the cluster better than the more specialized reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the best standalone issue to keep as the cluster representative because it is current, well-scoped, and has the most linkage/discussion among the listed issues.", + "best_issue_reason": "issue:45020 is the most representative issue because it frames the widest user-visible regression class and could serve as the closest cluster anchor, even though the items are not duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43531", - "right": "issue:45600", - "accept": false, - "reason": "Different bugs: Qwen3-MoE sliding window behavior vs stale auto_mappings references. Same area of model support, but not the same code path." - }, - { - "left": "issue:42907", - "right": "issue:43441", - "accept": false, - "reason": "Both mention Ministral, but one is about saving dequantized weights and the other is a FlashAttention runtime failure. Not the same defect." - }, - { - "left": "issue:43526", - "right": "issue:43901", - "accept": false, - "reason": "BEIT processor label reduction bug vs docs mismatch for TextClassificationPipeline. Different subsystem and failure mode." - }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Both involve embeddings/tokenizer-related behavior, but one is a processor/tokenizer regression and the other is t5gemma decoder embedding resize. Not one underlying bug." + "reason": "Different failures: processor/tokenizer image handling regression vs token embedding resize not updating decoder weights." }, { "left": "issue:44792", "right": "issue:46424", "accept": false, - "reason": "Janus image-generation test failure vs cache max length handling for quantized cache. Unrelated behaviors." + "reason": "Unrelated bugs: janus image-generation test failure vs cache length handling for quantized cache." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "AutoModel PyTorch import error vs post_init reinitializing output embeddings after resize. Different issue class and code path." + "reason": "ImportError due to missing PyTorch vs embedding reinitialization after resize; different code paths." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "Olmo model config import failure vs PIL backend processors requiring torchvision. Separate problems." + "reason": "Unrecognized model config vs PIL image processor requiring torchvision; unrelated issues." }, { "left": "issue:46082", "right": "issue:46399", "accept": false, - "reason": "LlamaConfig head_dim validation vs Qwen3.5 rotary kernel registration error. Different models and logic." + "reason": "LlamaConfig head_dim validation vs Qwen3.5 kernel registration error; not the same bug." }, { - "left": "issue:43441", - "right": "issue:45588", + "left": "issue:45372", + "right": "issue:46234", "accept": false, - "reason": "Both touch FlashAttention, but one is a Ministral-specific failure and the other is an AttributeError for sink-less models in integration code. Not the same bug." + "reason": "Both are missing-import reports, but for different symbols and subsystems; not a duplicate bug." }, { - "left": "issue:45372", - "right": "issue:46234", + "left": "issue:43441", + "right": "issue:45588", "accept": false, - "reason": "Gemma 4 processor import breakage vs missing falcon_mamba availability helper. Unrelated imports." + "reason": "Both mention flash attention, but one is a model-specific failure and the other is a sink-less-model AttributeError in a helper module." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error vs AddedToken keyword collision. No shared underlying code path." + "reason": "Notebook syntax error vs AddedToken keyword argument collision; unrelated." }, { "left": "issue:43299", "right": "issue:45216", "accept": false, - "reason": "Qwen3.5 loading failure vs incorrect save_pretrained checkpoint contents. Same family, but different operations and bugs." + "reason": "Qwen3VL MoE loading breakage vs incorrect Qwen3.5 save_pretrained output; different failure modes." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "Pipeline docs/behavior mismatch vs model config import error. Different issue type and scope." + "reason": "Documentation mismatch for `return_all_scores` vs model config import error; not the same underlying problem." }, { - "left": "issue:43824", - "right": "issue:44933", + "left": "issue:43673", + "right": "issue:44336", "accept": false, - "reason": "Two import errors, but for different symbols/modules and different root causes." + "reason": "Generation cache regression vs ANSI color output in loading_report; unrelated." }, { - "left": "issue:43673", - "right": "issue:44336", + "left": "issue:43824", + "right": "issue:44933", "accept": false, - "reason": "Generation cache regression vs ANSI color codes in loading_report. Completely unrelated." + "reason": "Different import errors involving different missing symbols and modules." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing eval-mode behavior vs D-FINE auxiliary loss omission. Different training semantics." + "reason": "Gradient checkpointing eval-mode behavior vs D-FINE auxiliary-loss logic; unrelated." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "Python version compatibility issue vs generic dependency problem. Too broad and not the same bug." + "reason": "Python version compatibility report vs vague dependency issue; not the same concrete bug." }, { - "left": "issue:44075", - "right": "issue:44368", + "left": "issue:44496", + "right": "issue:45759", "accept": false, - "reason": "SGD optimizer args ignored vs tie_word_embeddings warning in Qwen fine-tuning. Different feature areas." + "reason": "Generic unrecognized model error vs composite Qwen config unwrapping bug; different code paths." }, { - "left": "issue:44496", - "right": "issue:45759", + "left": "issue:44075", + "right": "issue:44368", "accept": false, - "reason": "Unrecognized model config vs AutoModelForCausalLM not unwrapping text_config for composite Qwen multimodal configs. Related to loading, but not the same defect." + "reason": "SGD optimizer args ignored vs tie_word_embeddings warning during LoRA fine-tuning; unrelated." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "t5gemma embedding resize issue vs Whisper empty-transcription generation failure. Unrelated models and code paths." + "reason": "Token embedding resize bug vs Whisper generation failure after align_special_tokens; different models and paths." }, { "left": "issue:43650", "right": "issue:46209", "accept": false, - "reason": "Placeholder issue title vs FP8 inference failure on B200. No overlap in bug report content." + "reason": "Placeholder/data request vs FP8 inference failure on B200; unrelated." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "BLIP2 dtype propagation bug vs Qwen3.5 save_pretrained regression. Different models and operations." + "reason": "BLIP2 dtype propagation bug vs Qwen3.5 checkpoint save regression; not a duplicate." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "auto_docstring crash with future annotations vs Gemma 4 processor import failure. Different tooling and root cause." + "reason": "Auto-docstring crash vs missing mistral_common symbol; different tooling and failure modes." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "MoE load-balancing loss omission vs Gemma 4 mm_token_type_ids defaulting. Different model behavior." + "reason": "MoE load-balancing loss bug vs Gemma 4 mm_token_type_ids defaulting issue; unrelated." }, { "left": "issue:38175", "right": "issue:44617", "accept": false, - "reason": "SigLIP2 zero probabilities vs Sam3Video CUDA OOM. Not the same bug or subsystem." - }, - { - "left": "issue:44521", - "right": "issue:45325", - "accept": false, - "reason": "Both multimodal, but assistant-mask generation and Qwen2.5-VL rope-position scaling are distinct code paths and symptoms." + "reason": "Zero-probability model output vs CUDA OOM; not the same bug." }, { "left": "issue:44279", "right": "issue:45405", "accept": false, - "reason": "Generic transformer dependency issue vs unreleased MIN_PEFT_VERSION bump. Different release/dependency concerns." + "reason": "General dependency issue vs minimum PEFT version bump; unrelated." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Both Qwen3.5 regressions, but loading breaks vs save_pretrained stores wrong visual encoder keys. Not the same concrete bug." + "reason": "Both involve Qwen3.5, but one is loading failure and the other is incorrect visual-encoder keys on save; different bugs." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification pipeline processor lookup bug vs broad remote_code regression report; not the same concrete issue." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Processor/tokenizer image handling regression vs Deepseek tokenizer output regression; different affected paths." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer correctness regressions in different model families; too broad to treat as the same bug." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "Single missing import for HybridCache vs broad 'many import errors' report; related theme but not a single identical bug." } ] }, @@ -11131,13 +11280,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:33290", "issue:39692", "issue:41950", + "issue:42175", "issue:42994", - "issue:43295", "issue:43299", "issue:43317", "issue:43441", @@ -11146,28 +11295,30 @@ "issue:43525", "issue:43650", "issue:43653", + "issue:43723", "issue:43854", "issue:43901", "issue:43976", + "issue:43994", "issue:44016", + "issue:44162", "issue:44261", "issue:44263", "issue:44279", "issue:44291", "issue:44297", - "issue:44351", "issue:44361", "issue:44462", "issue:44466", + "issue:44521", "issue:44623", "issue:44671", "issue:44683", "issue:44749", "issue:44779", - "issue:45020", "issue:45081", "issue:45276", - "issue:45278", + "issue:45325", "issue:45397", "issue:45405", "issue:45542", @@ -11175,20 +11326,16 @@ "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:44521|issue:45325", "issue:39692|issue:43901", - "issue:41950|issue:45823", "issue:43479|issue:43525", + "issue:41950|issue:45823", "issue:44683|issue:45542", "issue:43976|issue:45405", "issue:44462|issue:45081", @@ -11209,204 +11356,214 @@ "issue:44261|issue:44263", "issue:44361|issue:45593", "issue:44671|issue:44779", - "issue:44016|issue:45823" + "issue:44016|issue:45823", + "issue:42175|issue:43976", + "issue:43723|issue:44162", + "issue:43441|issue:44683", + "issue:43901|issue:43994" ], - "prepared_review_unit_hash": "d3b7d7781452c321599ef6d5c57ebe4a5b6e1b237a0f182defe7bf1e2a607c62", + "prepared_review_unit_hash": "b6eea5077641f64c4eab9484265128502509fdf786476ee64ee23e7ec3ade8d3", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35380, - "estimated_input_tokens": 8845, - "estimated_eval_tokens": 17946 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36297, + "estimated_input_tokens": 9075, + "estimated_eval_tokens": 18406 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes unrelated Transformers regressions across tokenizers, processors, configs, CI, docs, quantization, and deepspeed. None of the soft pairs look like true duplicates under a strict bug-level definition.", - "confidence": 0.91, - "canonical_issue_reason": "No clear canonical duplicate target; the items are mostly distinct regressions with only broad topical overlap.", + "summary": "This set is highly heterogeneous overall; most soft pairs are only thematically similar at a broad subsystem level, not true duplicates. The strongest hub among the issues is the tokenizer-loading regression around AutoTokenizer/tokenizer.json, but even that does not unify the cluster cleanly.", + "confidence": 0.22, + "canonical_issue_reason": "issue:44462 is the clearest concrete regression in the mostly tokenizer/loading-heavy subset, with a specific failure mode and enough breadth to be a plausible anchor if a single issue must represent the group.", "canonical_pr_reason": null, - "best_issue_reason": "If one issue must anchor the cluster, issue 45278 is the broadest umbrella-style regression report about imports breaking after an upgrade, but it still does not actually subsume the rest.", + "best_issue_reason": "issue:44462 is the best global representative because it is specific, user-facing, and closer than the docs/config-only or unrelated model/task bugs; it also sits near other tokenizer-related reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41950", - "right": "issue:45020", - "accept": false, - "reason": "Different failures: a video-classification processor lookup bug vs a broad remote_code regression." - }, - { - "left": "issue:44351", - "right": "issue:45278", - "accept": false, - "reason": "One is a specific missing HybridCache import; the other is a generic report of many import errors after upgrade." - }, - { - "left": "issue:44779", - "right": "issue:45812", - "accept": false, - "reason": "Both are tokenizer regressions, but on different models and different concrete symptoms." - }, - { - "left": "issue:43295", - "right": "issue:44779", + "left": "issue:44521", + "right": "issue:45325", "accept": false, - "reason": "Processor.tokenizer/custom image handling regression is not the same bug as Deepseek tokenizer misbehavior." + "reason": "Both are multimodal-related, but one is about assistant masks in chat templating and the other is about temporal position IDs in Qwen2.5-VL; different code paths and failures." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "Both are documentation-related, but one is a SigLIP2 example error and the other is stale pipeline docs for return_all_scores." + "reason": "One is a SigLIP2 docs example with model/processor mismatch and quantization failure; the other is a TextClassificationPipeline docs mismatch. Both are documentation issues, but not the same bug." }, { - "left": "issue:41950", - "right": "issue:45823", + "left": "issue:43479", + "right": "issue:43525", "accept": false, - "reason": "Video-classification image-processor lookup and Gemma4 device-mismatch are unrelated bugs." + "reason": "Different model configs and different failures: Phi4Multimodal default sub-config initialization vs Llama4Config missing pad_token_id." }, { - "left": "issue:43479", - "right": "issue:43525", + "left": "issue:41950", + "right": "issue:45823", "accept": false, - "reason": "Different config bugs: default multimodal sub-config initialization vs missing pad_token_id on Llama4Config." + "reason": "Both touch video/multimodal areas, but the bugs are unrelated: video-classification choosing image processors vs Gemma4 device mismatch in forward." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Unrelated code paths: flex_attention compilation vs tensorboard/TensorFlow backend setup." + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to the TensorFlow backend install issue." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python version compatibility failure is not the same as the PEFT minimum-version bump issue." + "reason": "One is a Python version compatibility bug; the other is a dependency version bump to unreleased peft. No shared code path." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Both involve tokenizer loading, but one ignores tokenizer.json while the other crashes in a Mistral regex patch." + "reason": "Both concern tokenizer loading, but one is AutoTokenizer ignoring tokenizer.json and the other is Mistral regex patching crashing on backend_tokenizer access; different defects." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Different model-loading failures for different models and different underlying causes." + "reason": "Both are model-loading regressions, but for different models and causes: Qwen3VL MoE loading vs GLM-4.7-Flash unit-test loading." }, { "left": "issue:44297", "right": "issue:44462", "accept": false, - "reason": "Related serialization/loading area, but tokenizer_class mismatch is not the same concrete bug as ignoring tokenizer.json." + "reason": "Both are tokenizer-related, but save_pretrained writing the wrong tokenizer_class is not the same bug as AutoTokenizer ignoring tokenizer.json." }, { "left": "issue:43450", "right": "issue:44279", "accept": false, - "reason": "Video processor shape bug and a dependency issue with transformers are not the same defect." + "reason": "Video processor batch-shape bug is unrelated to a generic dependency issue with transformers." }, { "left": "issue:43650", "right": "issue:46132", "accept": false, - "reason": "Completely unrelated issues: placeholder text vs AttentionInterface.register behavior change." + "reason": "A vague data request is unrelated to AttentionInterface.register changing function behavior." }, { "left": "issue:45081", "right": "issue:45907", "accept": false, - "reason": "Tokenizer regex crash and list_repo_files revision forwarding are unrelated." + "reason": "Tokenizer regex crash and list_repo_files not forwarding revision are unrelated." }, { "left": "issue:44466", "right": "issue:45276", "accept": false, - "reason": "lm_head.weight serialization inconsistency is different from Gemma4 resize_token_embeddings behavior." + "reason": "Both involve serialization/model-shape mechanics, but one is inconsistent lm_head.weight saving and the other is resize_token_embeddings not propagating in Gemma4." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error and init_empty_weights argument handling are unrelated." + "reason": "Notebook syntax error and init_empty_weights TypeError are unrelated failures." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "FlashAttention failure in Ministral-3 and long-ocr-input index error in Kosmos2.5 are different bugs." + "reason": "Ministral-3 FlashAttention failure and Kosmos2.5 OCR long-input index error are different model bugs." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Both touch quantization/offload, but one is model saving and the other is loading dequantized models with device_map=auto." + "reason": "Quantized model saving failure is not the same bug as device_map=auto failing to load a dequantized model with GPU+CPU offload." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "BigBirdTokenizer special-token registration and D-FINE auxiliary-loss behavior are unrelated." + "reason": "BigBirdTokenizer special-token registration bug is unrelated to D-FINE auxiliary losses being skipped." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "processor.save_pretrained missing files is not the same bug as gemma-4 zero3 from_pretrained failure." + "reason": "Processor save_pretrained missing files and Gemma4 zero3 from_pretrained are different code paths and symptoms." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "A save_pretrained file issue is unrelated to a post-upgrade data filtering slowdown report." + "reason": "Processor save_pretrained missing files is unrelated to the performance regression after upgrading Transformers." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Both are cleanup/CI issues, but they affect different files and different failure modes." + "reason": "Broken auto_mappings references and leftover dummy_pt_objects classes are separate repository-maintenance issues." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "Deepspeed OOM with Adafactor and optimizer.step being called twice are related only at a high level, not the same concrete bug." + "reason": "Both involve DeepSpeed and optimizers, but one is an OOM with Adafactor and the other is optimizer.step being called twice; not the same bug." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "A precision/config issue in MLA is unrelated to torch.split return values in GlmMoeDsaIndexer." + "reason": "Different GLM model internals: missing rms_norm_eps in config vs torch.split return values in GlmMoeDsaIndexer." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "MLukeTokenizer AttributeError and D-FINE auxiliary-loss handling are unrelated." + "reason": "MLukeTokenizer AttributeError on tasks is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "Different model-specific tokenizer/masked-LM regressions with no shared concrete code path." + "reason": "Both are output regressions in v5, but CamemBERT masked LM predictions and Deepseek tokenizer results are different models and failures." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, "reason": "Notebook syntax error and Gemma4 device mismatch are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "TensorFlow backend installation and Python version compatibility are different dependency issues." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading in v5 and ESM2 being broken are both broad regressions, but they do not point to the same underlying defect." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "FlashAttention failure in Ministral-3 and compiled flex_attention failure on torch>=2.9 are distinct attention-path bugs." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch for return_all_scores and incorrect SigLIP2 outputs with AutoModel/pipeline are unrelated." } ] }, @@ -11428,32 +11585,27 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36246", "issue:41628", "issue:41950", - "issue:42175", "issue:42491", "issue:42757", "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", @@ -11461,14 +11613,18 @@ "issue:44295", "issue:44297", "issue:44336", + "issue:44351", "issue:44373", + "issue:44485", "issue:44496", "issue:44521", "issue:44589", - "issue:44683", + "issue:44734", + "issue:44779", "issue:44871", "issue:45003", "issue:45103", + "issue:45292", "issue:45362", "issue:45397", "issue:45542", @@ -11476,28 +11632,25 @@ "issue:45563", "issue:45593", "issue:46132", - "issue:46399" + "issue:46399", + "issue:46424" ], "soft_pairs": [ - "issue:43723|issue:44162", - "issue:42175|issue:43976", - "issue:43441|issue:44683", - "issue:43901|issue:43994", - "issue:43262|issue:44521", "issue:43673|issue:43761", + "issue:43262|issue:44521", "issue:44297|issue:45397", "issue:43824|issue:43976", "issue:43525|issue:43881", "issue:43901|issue:44246", "issue:43525|issue:43643", "issue:44589|issue:45362", - "issue:41628|issue:43976", "issue:43784|issue:45542", + "issue:41628|issue:43976", "issue:44188|issue:44336", "issue:45103|issue:46399", "issue:45561|issue:45563", - "issue:44246|issue:44496", "issue:44336|issue:44373", + "issue:44246|issue:44496", "issue:44295|issue:45397", "issue:44242|issue:45593", "issue:42491|issue:43299", @@ -11506,198 +11659,222 @@ "issue:41950|issue:44220", "issue:44246|issue:44871", "issue:42886|issue:46132", - "issue:36246|issue:42757" + "issue:44734|issue:46424", + "issue:36246|issue:42757", + "issue:44336|issue:44485", + "issue:36246|issue:44351", + "issue:45003|issue:45561", + "issue:44779|issue:45292", + "issue:44336|issue:46424" ], - "prepared_review_unit_hash": "55689a5cf6bb13384d540b27b661672356698774438dce0d070242fd509f7c94", + "prepared_review_unit_hash": "16670fed138d61d2210802fd79a806f14883d4749ca3aff70367b24f394449f2", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35454, - "estimated_input_tokens": 8864, - "estimated_eval_tokens": 17984 + "soft_pair_count": 30, + "serialized_chars": 36041, + "estimated_input_tokens": 9011, + "estimated_eval_tokens": 18278 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a broad cluster of mostly unrelated Transformers v5-era regression reports, import failures, and model-specific loading bugs. None of the soft pairs look like the same concrete underlying bug/change under a strict duplicate standard, so all are rejected. The strongest representative issue is the Qwen3 MoE/v5 loading regression report.", - "confidence": 0.33, - "canonical_issue_reason": "issue:43299 is the clearest, most concrete representative of the cluster\u2019s recurring v5 regression theme: a version-bounded model-loading failure with a detailed Qwen3 MoE repro and explicit hf4.x vs hf5.x breakage.", + "summary": "This set is heterogeneous: it mixes import/export errors, model-loading regressions, cache/generation bugs, and unrelated docs/tests issues. None of the soft pairs look like the same underlying bug or a PR that could plausibly merge into one change set, so I would not accept any duplicate edges.", + "confidence": 0.95, + "canonical_issue_reason": "issue:43824 is the clearest representative of the cluster\u2019s concrete import/export regression pattern: a user-facing top-level import failure with multiple related references.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43299 best captures the dominant signal in this cluster: a specific, actionable Transformers v5 loading regression affecting Qwen3 MoE models, and it is closer to the cluster\u2019s main recurring pattern than the other, more isolated import/pipeline bugs.", + "best_issue_reason": "issue:43824 is the most actionable and recognizable failure mode in the set, but only as a rough representative; the cluster is too mixed to treat as one duplicate bucket.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43723", - "right": "issue:44162", - "accept": false, - "reason": "Both are load-time failures, but one is a generic AutoTokenizer regression and the other is an ESM2 model-specific breakage; different code paths." - }, - { - "left": "issue:42175", - "right": "issue:43976", - "accept": false, - "reason": "Packaging extras vs Python-version compatibility; unrelated issues." - }, - { - "left": "issue:43441", - "right": "issue:44683", - "accept": false, - "reason": "Both touch attention kernels, but one is Ministral FlashAttention in Transformers RC and the other is compiled flex_attention on torch>=2.9; not the same bug." - }, - { - "left": "issue:43901", - "right": "issue:43994", + "left": "issue:43673", + "right": "issue:43761", "accept": false, - "reason": "Docs mismatch vs incorrect model outputs; different underlying problems." + "reason": "Different regressions: chunked-prefill cache behavior vs CLIPVisionModel hidden_states output." }, { "left": "issue:43262", "right": "issue:44521", "accept": false, - "reason": "Different apply_chat_template bugs: audio sampling-rate default vs multimodal assistant-mask generation." - }, - { - "left": "issue:43673", - "right": "issue:43761", - "accept": false, - "reason": "Cache/chunked-prefill regression vs CLIP hidden_states regression; unrelated." + "reason": "Both involve chat templates, but one is an audio sampling-rate default and the other is multimodal assistant masks." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer save metadata mismatch vs DeepSpeed/from_pretrained loading failure; different paths." + "reason": "Tokenizer save metadata mismatch is unrelated to gemma-4 zero3 from_pretrained loading." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "Specific Qwen2.5-VL import error vs Python 3.9+ support issue; unrelated." + "reason": "Importing Qwen2.5-VL symbols is a different failure than Python version compatibility for 5.1.0." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Both are model-load failures, but they affect different model families and have different missing-field symptoms." + "reason": "Different model families and different loading failures: Llama4Config pad_token_id vs glm-4v-9b load failure." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Documentation issue vs intermittent import slowness; unrelated." + "reason": "Docs for return_all_scores are unrelated to intermittent import performance." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Different config-loading problems: missing pad_token_id for Llama4 vs missing fields with trust_remote_code." + "reason": "Both are config-related, but missing pad_token_id and trust_remote_code stripping fields are distinct bugs." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Different error modes and model paths; not the same underlying bug." + "reason": "Float8 storage lookup error is unrelated to Qwen3.5 chat crashes." }, { - "left": "issue:41628", - "right": "issue:43976", + "left": "issue:43784", + "right": "issue:45542", "accept": false, - "reason": "ImportError for AutoImageProcessor vs Python-version compatibility regression; unrelated." + "reason": "Different backend/import problems: sentence-transformers nn import vs tensorboard/tf backend handling." }, { - "left": "issue:43784", - "right": "issue:45542", + "left": "issue:41628", + "right": "issue:43976", "accept": false, - "reason": "Both are import/backend issues, but they involve different packages and code paths." + "reason": "AutoImageProcessor import failure is not the same as Python 3.9/3.10 support mismatch." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Attention-kernel divergence vs ANSI formatting in loading_report; unrelated." + "reason": "Attention kernel divergence on torch.compile is unrelated to ANSI codes in loading_report." }, { "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "auto_docstring crash with future annotations vs Qwen3.5 kernel-registration ValueError; unrelated." + "reason": "Future-annotations docstring crash and Qwen3.5 kernel registration error are different code paths." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "pytest-xdist file race vs stale warning in paged generate; different bugs." + "reason": "xdist race on captured_info.txt is unrelated to a stale num_return_sequences warning." }, { - "left": "issue:44246", - "right": "issue:44871", + "left": "issue:44336", + "right": "issue:44373", "accept": false, - "reason": "Import-performance issue vs Gemma eos_token_id config mismatch; unrelated." + "reason": "Terminal ANSI output bug is unrelated to a wrong docstring for position_ids." }, { - "left": "issue:44336", - "right": "issue:44373", + "left": "issue:44242", + "right": "issue:44496", "accept": false, - "reason": "ANSI code emission in loading_report vs a docstring typo; not the same issue." + "reason": "Missing load-balancing loss and unrecognized-model config errors do not share a code path." }, { "left": "issue:44295", "right": "issue:45397", "accept": false, - "reason": "Buffer read error vs model loading with zero3; different code paths." + "reason": "position_ids buffer access bug is unrelated to gemma-4 zero3 loading." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "Both concern optional-loss branches, but they are different models and different mechanisms; not the same concrete bug." + "reason": "Different training-loss issues: router-logits gating vs D-FINE auxiliary losses." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both mention Qwen3 MoE/v5 compatibility, but one is a LoRA portability problem and the other is a Qwen3VL MoE loading regression." + "reason": "Both mention Qwen MoE, but one is LoRA compatibility across branches and the other is a Qwen3VL-MoE loading regression." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Both are Qwen-related, but sliding_window behavior and chat crashes are distinct failures." + "reason": "Sliding_window behavior in Qwen3-MoE is not the same as Qwen3.5 chat crashes." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BeitImageProcessorFast label reduction vs unsafe sys.modules access; unrelated." + "reason": "BeitImageProcessorFast label reduction and unsafe sys.modules access are unrelated." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification pipeline misrouting to image processors vs audio fbank extraction issue; different subsystems." + "reason": "Video-classification pipeline processor selection is unrelated to _torch_extract_fbank_features()." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowness and Gemma-3 eos_token_id inconsistency are different subsystems." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Offline tokenizer cache loading vs AttentionInterface.register semantics; unrelated." + "reason": "Offline cache loading bug is unrelated to AttentionInterface.register changing function behavior." + }, + { + "left": "issue:44734", + "right": "issue:46424", + "accept": false, + "reason": "KV cache continuation crash in serve is not the same as max_cache_length being honored only for Quantized Cache." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Two different import errors involving unrelated APIs/packages; not a duplicate." + "reason": "Both are import errors, but they involve different symbols in different packages and different causes." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI output formatting and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Missing Qwen2_5_VLImageProcessor and missing HybridCache are separate export/import regressions." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access and xdist file races are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "DeepSeek tokenizer regression and resize_token_embeddings/output_embeddings mismatch are different bugs." + }, + { + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "ANSI code emission and cache length handling are unrelated." } ] }, @@ -11719,13 +11896,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36246", "issue:43257", "issue:43262", "issue:43352", "issue:43475", + "issue:43540", "issue:43596", "issue:43673", "issue:43901", @@ -11733,16 +11911,11 @@ "issue:44242", "issue:44246", "issue:44295", - "issue:44336", - "issue:44351", "issue:44466", "issue:44479", - "issue:44485", "issue:44496", "issue:44561", - "issue:44734", "issue:44749", - "issue:44779", "issue:44857", "issue:44877", "issue:44987", @@ -11757,259 +11930,242 @@ "issue:45397", "issue:45412", "issue:45459", - "issue:45561", "issue:45584", + "issue:45600", "issue:45656", + "issue:45685", "issue:45701", "issue:45759", "issue:45859", + "issue:45987", + "issue:46082", "issue:46123", "issue:46170", "issue:46291", "issue:46369", "issue:46399", - "issue:46424" + "issue:46431" ], "soft_pairs": [ - "issue:44734|issue:46424", - "issue:36246|issue:44351", - "issue:44336|issue:44485", - "issue:45003|issue:45561", - "issue:44779|issue:45292", - "issue:44336|issue:46424", - "issue:44242|issue:44857", "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:43901|issue:45003", "issue:43352|issue:45250", "issue:43673|issue:44561", + "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:45412|issue:46123", "issue:45292|issue:45701", "issue:44561|issue:46399", "issue:44877|issue:45375", + "issue:45685|issue:46431", "issue:44987|issue:45459", "issue:44995|issue:45200", "issue:44466|issue:45335", "issue:44479|issue:46170", - "issue:45397|issue:45656" + "issue:45397|issue:45656", + "issue:45137|issue:45600", + "issue:45987|issue:46082", + "issue:43475|issue:43540" ], - "prepared_review_unit_hash": "5be94fde938ffc8d2b68019fb6dc065aa6efebcc6a498e2c29624f86199411bc", + "prepared_review_unit_hash": "d408a70f253fb6ad906b14c38a9a0caa99aa7f57e53035b1f9b0e0b4c3be4d16", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36727, - "estimated_input_tokens": 9182, - "estimated_eval_tokens": 18620 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35608, + "estimated_input_tokens": 8902, + "estimated_eval_tokens": 18060 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is highly heterogeneous and does not look like a true duplicate cluster. The only plausible representative issue is the broad Qwen video regression report, but even that is just a weak canonical choice.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 44479 is the broadest and most concrete report here: a v5.3.0 regression affecting several Qwen VL/video variants with a clear symptom and scope.", + "summary": "This is a heterogeneous cluster of mostly unrelated Transformers issues. The soft pairs share broad keywords like Qwen, deepspeed, config, or generation, but they describe different models, symptoms, and code paths, so none of the candidate merges look like true duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "issue:45003 is the closest thing to a cluster representative because it is a core library/runtime bug in modeling_utils and sits near several import/config-related regressions, even though the overall cluster is not a single coherent bug.", "canonical_pr_reason": null, - "best_issue_reason": "44479 is the best single issue to represent the set because it is specific, user-facing, and spans multiple related model families, unlike the other narrower or vaguer reports.", + "best_issue_reason": "issue:45003 is the strongest global anchor: it touches core library behavior, is broadly impactful, and is more representative than the model-specific or docs-only reports in this mixed set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44734", - "right": "issue:46424", - "accept": false, - "reason": "Different bugs: KV-cache continuation crash in serve responses vs cache-length config handling." - }, - { - "left": "issue:36246", - "right": "issue:44351", - "accept": false, - "reason": "Both are import errors, but for unrelated symbols and model paths." - }, - { - "left": "issue:44336", - "right": "issue:44485", - "accept": false, - "reason": "ANSI-color output bug in loading_report vs GLM-5 RoPE implementation discussion." - }, - { - "left": "issue:45003", - "right": "issue:45561", - "accept": false, - "reason": "sys.modules access bug is unrelated to the xdist captured_info.txt race." - }, - { - "left": "issue:44779", - "right": "issue:45292", - "accept": false, - "reason": "Tokenization regression and resize_token_embeddings propagation are different code paths." - }, - { - "left": "issue:44336", - "right": "issue:46424", + "left": "issue:43475", + "right": "issue:46369", "accept": false, - "reason": "Console coloring in loading_report is unrelated to cache-length behavior." + "reason": "Different SAM 3 video failures: missing attribute on vision encoder output vs memory overflow in sam3-litetext." }, { "left": "issue:44242", "right": "issue:44857", "accept": false, - "reason": "MoE load-balancing loss omission vs an AMP/CUDA crash in LwDetrImageLoss." - }, - { - "left": "issue:43475", - "right": "issue:46369", - "accept": false, - "reason": "SAM3 vision encoder attribute error vs memory overflow in sam3-litetext are different failures." + "reason": "MoE load-balancing loss omission and RT-DETR AMP crash are unrelated bugs in different code paths." }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "Position_ids buffer access error vs a slow-data-processing regression after upgrade." + "reason": "Position_ids buffer access error is not the same as the reported slowdown after upgrading Transformers." }, { "left": "issue:45859", "right": "issue:46399", "accept": false, - "reason": "CPU-offload incompatibility in Qwen3_5MoeTextRotaryEmbedding is distinct from a kernels decorator registration crash." + "reason": "CPU offload incompatibility in rotary embeddings and a missing decorator on apply_rotary_pos_emb are distinct Qwen3.5 issues." }, { "left": "issue:46291", "right": "issue:46399", "accept": false, - "reason": "Kernels import/version requirement failure is not the same as the Qwen3.5 rotary-embedding registration error." + "reason": "Kernel version/revision requirements for evaluate are unrelated to the Qwen3.5 rotary registration crash." }, { "left": "issue:36246", "right": "issue:43596", "accept": false, - "reason": "Qwen2.5-VL import error and DeepSpeed ZeRO3/BertModel index error are unrelated." - }, - { - "left": "issue:44162", - "right": "issue:45200", - "accept": false, - "reason": "ESM2 model breakage and Gemma 4 token-type defaulting are different model-specific bugs." + "reason": "Qwen2.5 VL image processor import failure and BertModel Zero3 index error are different failures with no shared code path." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for return_all_scores is unrelated to unsafe sys.modules access." + "reason": "One is a documentation mismatch; the other is a runtime bug in sys.modules access." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "Nemotron Flash Attention 2 unsupported is not the same as the vague generic Flash Attention 2 issue." + "reason": "Nemotron flash-attn unsupported error is model-specific; the other is just a generic flash attention topic." }, { "left": "issue:43673", "right": "issue:44561", "accept": false, - "reason": "Chunked-prefill cache bug and removal of is_torch_fx_available affect different paths." + "reason": "Chunked-prefill cache handling and removal of is_torch_fx_available are separate incompatibilities." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are unrelated model-specific bugs." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Unknown-model config loading failure is unrelated to sys.modules indexing." + "reason": "Unrecognized model/config loading and unsafe sys.modules access are different loader failures." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MOE weight conversion issue and ZeRO3 deque underflow are different DeepSpeed failures." + "reason": "Qwen3 MoE weights not converted under accelerate+deepspeed is different from the ZeRO3 deque/index error." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio processor sampling-rate default bug is unrelated to import slowdown." + "reason": "Audio chat-template sampling-rate default and import slowdowns are unrelated issues." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "Embedding resize propagation and Whisper empty-transcription failure are different behaviors." + "reason": "resize_token_embeddings output mismatch and Whisper empty-transcription generation failure do not share the same root cause." }, { - "left": "issue:45412", - "right": "issue:46123", + "left": "issue:45375", + "right": "issue:45759", "accept": false, - "reason": "RT-DETR memory leak and MaskGenerationPipeline batch-finalization bug are unrelated." + "reason": "Strict config dropping a field and from_config not unwrapping text_config are both config-handling bugs, but they affect different fields and paths." }, { - "left": "issue:45375", - "right": "issue:45759", + "left": "issue:45412", + "right": "issue:46123", "accept": false, - "reason": "Missing deepstack_visual_indexes field and text_config unwrapping are separate Qwen config issues." + "reason": "RT-DETR memory leak and MaskGenerationPipeline dropping final partial batches are unrelated." }, { "left": "issue:45292", "right": "issue:45701", "accept": false, - "reason": "Embedding resize bug and version-dependent tokenization change are not the same defect." + "reason": "Embedding resize behavior and version-dependent tokenization changes are different problems." }, { "left": "issue:44561", "right": "issue:46399", "accept": false, - "reason": "is_torch_fx_available removal breakage is unrelated to the Qwen3.5 apply_rotary_pos_emb registration crash." + "reason": "Removed helper breaking trust_remote_code is unrelated to the Qwen3.5 apply_rotary_pos_emb registration crash." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "granite_speech strict-config loading and Qwen3_5MoeVisionConfig field loss are different config problems." + "reason": "Both involve config strictness, but they concern different models and different missing fields." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Loading a specific model package failing is unrelated to tokenizer errors hidden by protobuf exception handling." + "reason": "Kernel revision requirements and tokenizer error masking when protobuf is absent are distinct issues." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "GlmMoeDsa stale cache on second forward and Gemma 4 mm_token_type_ids defaults are unrelated." + "reason": "GlmMoeDsa stale cache on second forward and Gemma 4 zero defaults are unrelated." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "lm_head.weight serialization inconsistency is distinct from resize_token_embeddings not updating decoder.embed_tokens." + "reason": "Tied-weight serialization inconsistency and decoder.embed_tokens not updating after resize are related only loosely by embeddings, not the same bug." }, { "left": "issue:44479", "right": "issue:46170", "accept": false, - "reason": "Qwen video-input regression and Windows ROCm distributed import failure are unrelated." + "reason": "Qwen video input regression and Windows/ROCm torch.distributed import failure are unrelated." }, { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "Gemma-4 ZeRO3 loading and duplicate optimizer-step behavior with DeepSpeed are different bugs." + "reason": "Gemma-4 Zero3 from_pretrained and optimizer.step being called twice under deepspeed are different deepspeed problems." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 index error and stale auto_mappings references are not the same underlying bug." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache seq-length type mismatch and LlamaConfig head_dim validation are unrelated." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MPS histogram failure in MoE and missing chat template error are different features and code paths." } ] }, @@ -12031,8 +12187,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ + "issue:28282", + "issue:30990", + "issue:41628", "issue:42886", "issue:42994", "issue:43065", @@ -12040,7 +12199,6 @@ "issue:43408", "issue:43475", "issue:43482", - "issue:43540", "issue:43643", "issue:43646", "issue:43653", @@ -12061,20 +12219,18 @@ "issue:44869", "issue:44871", "issue:44877", + "issue:44933", "issue:44977", "issue:44995", - "issue:45137", "issue:45229", "issue:45230", "issue:45356", + "issue:45381", "issue:45397", "issue:45563", "issue:45584", "issue:45588", - "issue:45600", "issue:45656", - "issue:45987", - "issue:46082", "issue:46123", "issue:46132", "issue:46170", @@ -12082,17 +12238,14 @@ "issue:46421" ], "soft_pairs": [ - "issue:45137|issue:45600", - "issue:45987|issue:46082", - "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:42886|issue:45229", "issue:44368|issue:44977", - "issue:43065|issue:43408", + "issue:44995|issue:45588", "issue:44368|issue:45563", + "issue:43065|issue:43408", "issue:44188|issue:46202", "issue:44877|issue:46170", "issue:43931|issue:44368", @@ -12111,216 +12264,219 @@ "issue:45230|issue:45397", "issue:42994|issue:43482", "issue:44038|issue:44623", - "issue:43475|issue:43643" + "issue:43475|issue:43643", + "issue:28282|issue:30990", + "issue:43475|issue:45381", + "issue:41628|issue:44933" ], - "prepared_review_unit_hash": "89f0df0ec15f372b06c5b64fb02374d5c62db277ad340f276f5df4416c451507", + "prepared_review_unit_hash": "d19114d88199de63fbcc0ee69bbd2a227997372927e62fa30b19a4d081771000", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 35928, - "estimated_input_tokens": 8982, - "estimated_eval_tokens": 18220 + "serialized_chars": 35881, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, - "split": true, - "analyst_result": { - "summary": "No soft pairs look like true duplicates; they mostly share a broad subsystem or model family but describe different concrete failures, so all soft edges are rejected. For the issue set, issue 46421 is the strongest canonical tracking candidate because it is specific, open, and directly describes a cache-lazy-init recompilation regression in chunked prefill.", - "confidence": 0.86, - "canonical_issue_reason": "Issue 46421 is the clearest representative of the cache/chunked-prefill regression family: it states a concrete symptom, a narrow code path, and an active bug surface.", - "canonical_pr_reason": null, - "best_issue_reason": "Issue 46421 is the most actionable issue in the set: it is open, technically specific, and tied to a reproducible performance regression rather than a broad or ambiguous report.", - "best_pr_reason": null, - "soft_edge_verdicts": [ - { - "left": "issue:45137", - "right": "issue:45600", - "accept": false, - "reason": "DeepSpeed ZeRO3 deque exhaustion and removed auto_mappings config refs are unrelated failures in different code paths." - }, - { - "left": "issue:45987", - "right": "issue:46082", - "accept": false, - "reason": "StaticCache return-type mismatch and LlamaConfig head_dim validation are different bugs." - }, - { - "left": "issue:43475", - "right": "issue:43540", - "accept": false, - "reason": "SAM3 VisionEncoder attribute loss and Qwen3OmniMoe video-input ValueError are different model paths and symptoms." - }, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the items span unrelated import errors, tokenizer/processor save issues, model-specific regressions, generation/cache bugs, and SAM/Qwen/Whisper-specific failures. I did not find any soft-edge pair that clearly describes the same underlying bug or change, so all soft links are rejected.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 45229 is the most prominent single report in the set: it has the highest discussion activity, an inbound reference, and a concrete high-impact runtime failure. If one issue must represent the cluster, this is the best anchor despite the cluster\u2019s overall heterogeneity.", + "canonical_pr_reason": "No PRs are present in this set.", + "best_issue_reason": "45229 is the strongest standalone issue candidate because it is concrete, heavily discussed, and clearly actionable. None of the other issues are close enough duplicates to displace it.", + "best_pr_reason": "No PRs are present in this set.", + "soft_edge_verdicts": [ { "left": "issue:44483", "right": "issue:46202", "accept": false, - "reason": "API request rejection and Gemma tokenizer roundtrip failure are unrelated." + "reason": "Both involve generation/output behavior, but one is an API acceptance regression and the other is a Gemma pipeline roundtrip/tokenization issue; they are not the same bug." }, { "left": "issue:43653", "right": "issue:44162", "accept": false, - "reason": "BigBirdTokenizer special-token registration and ESM2 breakage do not describe the same failure." + "reason": "Both are model/tokenizer-related, but BigBird tokenizer decode behavior and ESM2 breakage are distinct failures with no shared code-path evidence." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Different models and failures: one is a tied-weights warning, the other is a post_init policy issue for patchtsmixer." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline tokenizer cache loading and Gemma4 multi-GPU CUDA OOM are unrelated bugs." + "reason": "Offline tokenizer cache loading and Gemma multi-GPU CUDA OOM are unrelated problems." }, { - "left": "issue:44060", - "right": "issue:44077", + "left": "issue:44368", + "right": "issue:44977", "accept": false, - "reason": "Tied-weights warning behavior and patchtsmixer post_init policy are different issues." + "reason": "Both mention Qwen3.5, but one is a weight-tying warning while the other is a flash-attention generation bug; not the same underlying defect." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "Stale indexer cache on second forward and flash_attention s_aux=None AttributeError are different code paths." + "reason": "A stale indexer cache on second forward and a flash-attention None-attribute crash are different code paths and symptoms." }, { "left": "issue:44368", - "right": "issue:44977", + "right": "issue:45563", "accept": false, - "reason": "Tied embeddings warning and flash-attention generation failure are not the same concrete bug." + "reason": "One is a config warning about tied embeddings; the other is a stale warning in paged generate for num_return_sequences." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "Both are SAM3-related, but one concerns a dummy Conv2d in the pixel decoder while the other is a model-type mismatch warning." - }, - { - "left": "issue:44368", - "right": "issue:45563", - "accept": false, - "reason": "A tied-embeddings warning is unrelated to a stale num_return_sequences warning in paged generate()." + "reason": "Both are SAM3-related, but one concerns a dummy Conv2d in pixel decoding and the other a model-type mismatch warning during instantiation." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "Torch.compile kernel divergence and Gemma tokenizer roundtrip failure are unrelated." + "reason": "torch.compile attention-kernel divergence and Gemma pipeline roundtrip failure are unrelated bugs." }, { "left": "issue:44877", "right": "issue:46170", "accept": false, - "reason": "Strict granite_speech config loading and Windows ROCm import/type errors are different problems." + "reason": "Strict config loading for granite_speech and a Windows ROCm torch.distributed import/dependency problem are unrelated." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and tied-embeddings warnings are unrelated." + "reason": "A Qwen3-VL weight-shape mismatch and a Qwen3.5 tie_word_embeddings warning are different loading/config issues." }, { "left": "issue:45584", "right": "issue:46170", "accept": false, - "reason": "Whisper empty-transcription generation failure and Windows ROCm import issues do not share the same code path." + "reason": "Whisper empty-transcription generation failure and a Windows ROCm import error do not share the same failure mode." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Both touch cache/compiled generation, but one is about diverging attention kernels under torch.compile and the other about lazy cache initialization causing recompilation." }, { "left": "issue:43065", "right": "issue:46123", "accept": false, - "reason": "SAM3 pixel-decoder setup and MaskGenerationPipeline final-batch handling are unrelated." + "reason": "SAM3 pixel-decoder configuration and MaskGenerationPipeline partial-batch result dropping are unrelated." }, { "left": "issue:44297", "right": "issue:44623", "accept": false, - "reason": "Tokenizer save_pretrained class mismatch and processor save_pretrained missing files are both save-pretrained issues, but not the same concrete failure." + "reason": "Both are save_pretrained-related, but tokenizer_class mismatch and missing processor files are different bugs and likely different save paths." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer AttributeError and a performance regression after upgrade are different bugs." + "reason": "MLukeTokenizer AttributeError and a regression causing slower filtering after upgrade are unrelated." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Whisper decode crash on trailing replacement characters and Kimi-K2.5 codec regression are different tokenizer issues." + "reason": "Both involve tokenization/decoding regressions, but Whisper timestamp decoding and Kimi-K2.5 codec handling are distinct issues." }, { "left": "issue:44623", "right": "issue:45656", "accept": false, - "reason": "Missing processor files and double optimizer stepping under deepspeed are unrelated." + "reason": "Missing processor save files and a deepspeed optimizer double-step bug are entirely different subsystems." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Both mention broad v5 breakage, but custom model initialization and ESM2 are different code paths and symptoms." + "reason": "A generic Transformers 5.0 custom-model initialization breakage is not the same as the ESM2-specific breakage report." }, { "left": "issue:44749", "right": "issue:45656", "accept": false, - "reason": "A slowdown regression and optimizer.step being called twice are not the same bug." + "reason": "Performance regression in filtering after upgrade and double optimizer stepping with deepspeed are unrelated." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "Checkpoint key mismatch and data-processing slowdown are unrelated." + "reason": "resume_from_checkpoint key mismatch and slower filtering after upgrade do not describe the same defect." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Generation kwargs update after sync_gpus and missing cache during chunked_prefill are both generation/cache topics, but they are distinct failure modes." + "reason": "Both concern generation internals, but one is an _update_model_kwargs_for_generation/sync_gpus bug and the other is missing cache in chunked_prefill." }, { "left": "issue:44871", "right": "issue:46132", "accept": false, - "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior change are unrelated." + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior changes are unrelated." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "A generic bug report is not evidence of the same issue as gemma-4 zero3 from_pretrained." + "reason": "The first item is too generic to match the specific gemma-4 zero3 from_pretrained problem." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are separate model I/O problems." + "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are different quantization/loading paths." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe/v5 breakage and processor save_pretrained missing files are different bugs." + "reason": "Qwen3-VL-Moe breakage and missing processor save files are unrelated." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "SAM3 video encoder missing fpn_position_embeddings and trust_remote_code config-field loss are unrelated." + "reason": "A missing field in Sam3VisionEncoderOutput is not the same as a trust_remote_code AutoConfig missing-fields issue." }, { - "left": "issue:44188", - "right": "issue:46421", + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch-missing ImportError for AutoModel and Sentence Transformers loading hang are different problems." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal vision issues, but they affect different models and different attributes/position-id logic." + }, + { + "left": "issue:41628", + "right": "issue:44933", "accept": false, - "reason": "Torch.compile branching causing divergent attention kernels and cache lazy init causing recompilation are related broadly, but not the same concrete code-path problem." + "reason": "Both are import-related, but one is a missing top-level AutoImageProcessor import and the other is a non-existent image_utils import; insufficient evidence they stem from the same change." } ] }, @@ -12342,34 +12498,35 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ - "issue:28282", - "issue:30990", "issue:36010", "issue:41628", "issue:42175", "issue:42371", "issue:42673", "issue:42831", + "issue:42898", "issue:42994", "issue:43232", + "issue:43278", "issue:43377", - "issue:43475", "issue:43504", "issue:43525", "issue:43526", "issue:43704", "issue:43723", "issue:43784", - "issue:43824", "issue:43873", + "issue:43976", "issue:44038", + "issue:44164", "issue:44188", "issue:44222", "issue:44242", "issue:44351", "issue:44496", + "issue:44568", "issue:44792", "issue:44933", "issue:44945", @@ -12380,7 +12537,6 @@ "issue:45137", "issue:45310", "issue:45375", - "issue:45381", "issue:45406", "issue:45478", "issue:45561", @@ -12393,26 +12549,23 @@ "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:45406|issue:46202", "issue:45375|issue:45478", - "issue:43723|issue:46326", "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:46077|issue:46123", "issue:41628|issue:44351", - "issue:43232|issue:44188", "issue:44964|issue:45137", + "issue:43232|issue:44188", "issue:44222|issue:45310", "issue:44945|issue:45753", - "issue:42994|issue:43824", + "issue:42994|issue:44164", "issue:44038|issue:45310", "issue:36010|issue:41628", "issue:43232|issue:44792", @@ -12422,226 +12575,226 @@ "issue:43526|issue:45561", "issue:43504|issue:45561", "issue:42175|issue:44933", - "issue:43377|issue:44242" + "issue:43377|issue:44242", + "issue:42898|issue:44568", + "issue:43976|issue:44933", + "issue:43278|issue:45137" ], - "prepared_review_unit_hash": "ced8dfcfe10754e79adb075424f0f1d34271a47c3d6205814059aa003cdb56bf", + "prepared_review_unit_hash": "e58a46f4e8f4b179d359ac3a179e12a52a1f41acd727052a9a788db60a848092", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 36060, - "estimated_input_tokens": 9015, - "estimated_eval_tokens": 18286 + "serialized_chars": 36023, + "estimated_input_tokens": 9006, + "estimated_eval_tokens": 18268 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous; only the Qwen3ForCausalLM VRAM-leak reports (42673/43704) are a clear duplicate pair, while the rest are distinct import/config/runtime bugs.", - "confidence": 0.94, - "canonical_issue_reason": "issue:42673 is the best duplicate anchor because issue:43704 repeats the exact same VRAM-leak title, and 42673 has the fuller report with more discussion/references.", + "summary": "This cluster is mostly a loose bundle of unrelated Transformers bug reports, with one clear duplicate pair around the Qwen3.5 MoE from_pretrained failure. Most other soft pairs share only broad subsystem terms (imports, saving/loading, generation, Qwen models) but not the same concrete bug.", + "confidence": 0.95, + "canonical_issue_reason": "issue:45310 is the better canonical issue: it cleanly represents the Qwen3.5 MoE from_pretrained regression, and issue:45478 is an obvious duplicate with the same title and bug scope.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42673 is the strongest representative issue overall: it describes a concrete reproducible bug, has the clearest duplicate sibling, and is better documented than the other candidates.", + "best_issue_reason": "issue:45310 is the strongest global representative because it has an exact duplicate in the set and the problem description is specific and central to its subcluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:28282", - "right": "issue:30990", - "accept": false, - "reason": "Both involve loading/runtime failures, but one is a missing PyTorch dependency error and the other is Sentence Transformers hanging; different bugs." - }, - { - "left": "issue:43475", - "right": "issue:45381", + "left": "issue:42371", + "right": "issue:45092", "accept": false, - "reason": "Both are multimodal/VL issues, but they fail on different model classes and different missing fields/position-id behavior." + "reason": "Different problems: TF32 backend configuration vs. an old InternVL2 meta-initialization incompatibility." }, { "left": "issue:45406", "right": "issue:46202", "accept": false, - "reason": "Gemma Processor attribute error and Gemma pipeline roundtrip tokenization failure are different code paths and symptoms." - }, - { - "left": "issue:41628", - "right": "issue:44933", - "accept": false, - "reason": "Both are import problems, but they concern different symbols and different modules; not the same underlying defect." - }, - { - "left": "issue:42371", - "right": "issue:45092", - "accept": false, - "reason": "TF32 API guidance and InternVL2 meta-init incompatibility are unrelated issues." + "reason": "Both involve multimodal/image-ish behavior, but one is a processor `_tokenizer` attribute error and the other is a Gemma roundtrip/tokenization issue." }, { "left": "issue:45375", "right": "issue:45478", "accept": false, - "reason": "These are both Qwen3.5/MoE-related, but one is a missing vision-config field and the other is a generic from_pretrained error; not enough to merge." + "reason": "Related Qwen3.5 family, but one is a missing config field silently dropped by strict, the other is a from_pretrained loading error." }, { - "left": "issue:43723", - "right": "issue:46326", + "left": "issue:42994", + "right": "issue:43525", "accept": false, - "reason": "Tokenizer loading in v5 and mistral_common continue_final_message are different tokenizer bugs." + "reason": "Different code paths: quantized model save failure vs. missing `pad_token_id` on `Llama4Config`." }, { - "left": "issue:42994", - "right": "issue:43525", + "left": "issue:43723", + "right": "issue:46326", "accept": false, - "reason": "Quantized-model save failure and Llama4Config missing pad_token_id are unrelated failures." + "reason": "Tokenizer loading regression vs. `continue_final_message` behavior for mistral_common tokenizers; not the same bug." }, { "left": "issue:43784", "right": "issue:46399", "accept": false, - "reason": "Import-time NameError in sentence-transformers and Qwen3.5 kernel-registration crash are different problems." + "reason": "Import-time `nn` NameError in sentence-transformers integration is unrelated to Qwen3.5 kernel registration." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "One is a Qwen3.5-MoE from_pretrained error; the other is a missing deepstack_visual_indexes field in Qwen3_5MoeVisionConfig." + "reason": "Same model family, but different failures: from_pretrained error vs. config field dropping." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Tokenizer regex patch crash and auto-docstring annotation crash are unrelated attribute errors." + "reason": "Both are AttributeErrors, but one is Mistral tokenizer regex patching and the other is an auto-docstring kwargs processing crash." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading behavior and pipeline-parallel output corruption are different bugs." + "reason": "Quantization/offloading bug is unrelated to pipeline-parallelism output corruption." }, { - "left": "issue:46077", - "right": "issue:46123", + "left": "issue:42673", + "right": "issue:43704", "accept": false, - "reason": "Gemma4 attention-mask construction and MaskGenerationPipeline partial-batch handling are separate issues." + "reason": "Same Qwen3 model label, but the reports describe different VRAM leak contexts and are not clearly the same code-path." }, { - "left": "issue:42673", - "right": "issue:43704", - "accept": true, - "reason": "Exact same Qwen3ForCausalLM VRAM-leak bug title and same failure mode; clearly duplicate reports." + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Both touch generation attention/pipeline logic, but one is bidirectional attention masking and the other is partial-batch result dropping." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they are for different missing exports and different underlying causes." + "reason": "Both are import errors, but for different symbols (`AutoImageProcessor` vs `HybridCache`) and different failure modes." }, { - "left": "issue:43232", - "right": "issue:44188", + "left": "issue:44964", + "right": "issue:45137", "accept": false, - "reason": "Generation kwarg sync behavior and torch.compile attention-kernel branching are different generation bugs." + "reason": "Model loading failure for Phi-4 multimodal is unrelated to the DeepSpeed ZeRO3 deque error." }, { - "left": "issue:44964", - "right": "issue:45137", + "left": "issue:43232", + "right": "issue:44188", "accept": false, - "reason": "Phi-4 multimodal loading and DeepSpeed ZeRO3 deque underflow are unrelated." + "reason": "Different generation/attention issues: kwargs update after `sync_gpus` vs. diverging attention kernels under `torch.compile`." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "FP8 save_pretrained on MoE and Qwen3.5 MoE from_pretrained failure are distinct operations and symptoms." + "reason": "Both mention MoE/Qwen, but one is FP8 save_pretrained and the other is a from_pretrained loading error; not the same bug." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Pipeline-parallel output errors and Qwen3_5 image infinite-loop behavior are not the same bug." + "reason": "Pipeline parallelism incorrect output is unrelated to the Qwen3.5 specific-image infinite loop." }, { "left": "issue:42994", - "right": "issue:43824", + "right": "issue:44164", "accept": false, - "reason": "Quantized model saving and missing Qwen2_5_VLForConditionalGeneration import are unrelated." + "reason": "Both involve save/load, but one is quantized model saving and the other is extra_state handling; different concrete failures." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Qwen3-VL-Moe compatibility and Qwen3.5-MoE from_pretrained failure are different model/version problems." + "reason": "Same broad Qwen MoE area, but different model variants and generations of transformers; insufficient evidence of the same underlying bug." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Missing GenerationMixin and missing AutoImageProcessor are separate import regressions." + "reason": "Different missing exports: `GenerationMixin` vs `AutoImageProcessor`." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation kwargs after sync_gpus and janus image-generation test failure do not look like the same bug." + "reason": "Generation kwargs handling is unrelated to the janus image-generation test failure." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FineGrainedFP8 accuracy drift and FP8 save_pretrained failure are different FP8-related issues." + "reason": "FP8 accuracy regression is not the same as FP8 save_pretrained failure, even if both mention FP8." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "TensorFlow backend packaging and leftover dummy PT classes are unrelated packaging problems." + "reason": "Backend dependency packaging issue is unrelated to leftover dummy classes leaking into `dir(transformers)`." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Unrecognized model/config handling and pytest-xdist file-race in testing utils are unrelated." + "reason": "Model config loading failure is unrelated to the pytest-xdist captured_info.txt race." }, { "left": "issue:43526", "right": "issue:45561", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug and xdist race on captured_info.txt are unrelated." + "reason": "BeitImageProcessorFast label reduction bug is unrelated to the xdist file race." }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Legacy-field pretrained loading for Beit and xdist testing race are unrelated." + "reason": "Pretrained preset loading with a legacy field is unrelated to the testing-utils race condition." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "TensorFlow backend installation and image_utils import issue are not the same bug." + "reason": "Package backend selection is unrelated to missing image_utils imports." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "MIMI padding-mask batching bug and MoE load-balancing loss omission are different model behaviors." + "reason": "MIMI padding-mask batching bug is unrelated to load-balancing loss not being added." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer cleanup behavior change is unrelated to mdeberta BOS/EOS token insertion regression." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python-version compatibility issue is unrelated to the image_utils import problem." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift during evaluation is unrelated to the ZeRO3 deque underflow." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft edge, issue:42673 vs issue:43704, is well supported by the exact matching title and similar report content. The other pairwise rejections read appropriately cautious rather than overstated. No mergeability concerns apply because there are no PR pairs." - }, - "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 }, @@ -12656,7 +12809,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:22355", "issue:27957", @@ -12670,14 +12823,12 @@ "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", @@ -12692,10 +12843,10 @@ "issue:44779", "issue:44869", "issue:44933", + "issue:45071", "issue:45092", "issue:45103", "issue:45125", - "issue:45137", "issue:45200", "issue:45310", "issue:45356", @@ -12703,31 +12854,31 @@ "issue:45656", "issue:45712", "issue:45812", + "issue:46082", "issue:46132", - "issue:46231" + "issue:46209", + "issue:46231", + "issue:46431" ], "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:42222|issue:45200", "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:42898|issue:45356", "issue:43874|issue:45092", "issue:44295|issue:45656", "issue:44075|issue:46231", - "issue:42898|issue:44448", "issue:44060|issue:45125", + "issue:42898|issue:44448", "issue:45405|issue:45712", "issue:44222|issue:44297", "issue:44933|issue:45712", @@ -12738,218 +12889,232 @@ "issue:30333|issue:34689", "issue:43479|issue:43644", "issue:42831|issue:44623", + "issue:46209|issue:46431", "issue:30333|issue:36331", "issue:30333|issue:35141", - "issue:22355|issue:28282" + "issue:22355|issue:28282", + "issue:45071|issue:46082" ], - "prepared_review_unit_hash": "63045dc476d79f5ef3d39f042f4afdbcd9115f3d56351fd0f6a0a4ff27da2289", + "prepared_review_unit_hash": "a7dffda18a604c305c18b8d02b6655c61d626629a4be5703c00975060d8b6217", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 34, - "serialized_chars": 37355, - "estimated_input_tokens": 9339, - "estimated_eval_tokens": 18934 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37640, + "estimated_input_tokens": 9410, + "estimated_eval_tokens": 19076 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is highly heterogeneous: most soft-similarity pairs share broad vocabulary or the same subsystem, but not the same concrete bug. I would not deduplicate any of the candidate pairs.", - "confidence": 0.64, - "canonical_issue_reason": "issue:42898 is the best cluster anchor because it is the clearest, most general v5 tokenizer-regression report and sits near several other tokenizer-related candidates, even though the set as a whole is not a true duplicate cluster.", + "summary": "These items are mostly unrelated bug reports spanning different subsystems, models, and release regressions. I found no soft-edge pair that clearly describes the same underlying issue, so I would keep them separate.", + "confidence": 0.61, + "canonical_issue_reason": "issue:45812 is a broad, concrete regression report with clear user impact: AutoTokenizer returning wrong token IDs across an entire model family. It is specific enough to anchor a cluster while still representing a common v5 tokenizer regression theme.", "canonical_pr_reason": null, - "best_issue_reason": "issue:42898 is the most representative issue among these entries: it is specific, well-scoped, and broadly phrased compared with the more model-specific or unrelated reports.", + "best_issue_reason": "issue:45812 is the strongest standalone issue here because it is severe, clearly described, and affects multiple Granite models with a silent correctness regression.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42898", - "right": "issue:44568", - "accept": false, - "reason": "Both are tokenizer regressions in v5, but they describe different behaviors and different models; not the same bug." - }, - { - "left": "issue:43976", - "right": "issue:44933", - "accept": false, - "reason": "Unrelated failures: Python compatibility vs a missing image_utils import." - }, - { - "left": "issue:43278", - "right": "issue:45137", - "accept": false, - "reason": "Different code paths: dtype mismatch in evaluation vs a DeepSpeed ZeRO3 deque error." - }, { "left": "issue:30990", "right": "issue:35141", "accept": false, - "reason": "Stuck loading Sentence Transformers is unrelated to embedding reinitialization after resize." + "reason": "Different problems: loading/stall behavior in Sentence Transformers vs embedding reinitialization after resize. No shared code path." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Save-pretrained missing files and Qwen3.5 MoE from_pretrained errors are different problems." + "reason": "Different artifacts and symptoms: processor.save_pretrained file omission vs Qwen3.5 MoE from_pretrained failure." }, { - "left": "issue:42222", - "right": "issue:45200", + "left": "issue:30333", + "right": "issue:33290", "accept": false, - "reason": "VitPose breakage and Gemma 4 mm_token_type_ids behavior are unrelated multimodal issues." + "reason": "MLflow reporting on training failure is unrelated to a deepspeed OOM during Adafactor optimization." }, { - "left": "issue:30333", - "right": "issue:33290", + "left": "issue:42222", + "right": "issue:45200", "accept": false, - "reason": "MLFlow reporting on failed jobs is unrelated to an Adafactor/DeepSpeed OOM." + "reason": "Different model families and failure modes: broken VitPose models vs Gemma 4 multimodal token type defaults." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Different model components: multimodal config defaults vs Whisper timestamp decoding." + "reason": "Phi4MultimodalConfig default-init bug is unrelated to Whisper word timestamp decode crashing on trailing replacement chars." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Config initialization bug is unrelated to ModelOutput key assignment." + "reason": "Config default initialization and ModelOutput key assignment are distinct code paths and bugs." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but they affect different tokenization behaviors and different models." + "reason": "Both are v5 tokenizer regressions, but they affect different behaviors and model families; no evidence of one underlying bug." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Sentence-transformers import failure and missing Qwen2.5-VL export are different import errors." + "reason": "Importing sentence-transformers with missing nn is unrelated to a missing Qwen2.5 VL symbol export." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "General tokenizer loading issue and a specific BOS/EOS regression are not the same underlying bug." + "reason": "Generic tokenizer loading issue vs a specific add_special_tokens BOS/EOS regression; not the same bug." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "Both mention v5 tokenizer regressions, but the affected tokenizer and failure mode differ." - }, - { - "left": "issue:42898", - "right": "issue:45356", - "accept": false, - "reason": "Different tokenizer regressions for different models; not one concrete duplicate." + "reason": "Different tokenizer regressions for different models; similar theme but not the same underlying defect." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error and documentation parameter mismatch are unrelated." + "reason": "Notebook syntax error is unrelated to a documentation parameter-name mismatch." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "KV-cache continuation crash and auto-docstring annotation handling are unrelated." + "reason": "KV-cache continuation crash in serve is unrelated to auto_docstring failing under future annotations." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "Different multimodal bugs: missing image patch method vs remote-code checkpoint meta-init incompatibility." + "reason": "Different multimodal model compatibility issues: missing image patch API vs remote-code checkpoint/meta-init incompatibility." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Position_ids buffer read error is unrelated to duplicate optimizer stepping under DeepSpeed." + "reason": "Reading a registered buffer and double optimizer stepping in deepspeed are unrelated." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "SGD argument handling and a docs-only parameter name mismatch are unrelated." + "reason": "Optimizer argument handling is unrelated to a documentation typo." }, { - "left": "issue:42898", - "right": "issue:44448", + "left": "issue:44060", + "right": "issue:45125", "accept": false, - "reason": "Both are v4-to-v5 regression reports, but for different models and different outputs." + "reason": "Both involve Qwen models, but one is a tied-weights warning bug and the other is missing tensor-parallel plan; not clearly the same change." }, { - "left": "issue:44060", - "right": "issue:45125", + "left": "issue:42898", + "right": "issue:44448", "accept": false, - "reason": "Incorrect tied-weights warning and missing tensor-parallel plan are separate Qwen3 issues." + "reason": "Tokenizer cleanup behavior change and Pegasus output drift are different regressions with different likely causes." }, { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "PEFT version pinning and leftover dummy classes are unrelated repo-maintenance issues." + "reason": "A dependency version bump is unrelated to leftover dummy class cleanup." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "Both involve save_pretrained, but one is FP8 MoE serialization and the other is tokenizer config metadata." + "reason": "FP8 save_pretrained behavior and tokenizer save_pretrained metadata mismatch affect different save paths." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing image_utils import and leftover dummy classes are both export/namespace issues, but not the same concrete bug." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer config metadata and deepspeed optimizer stepping are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Pegasus v4/v5 output mismatch is unrelated to AttentionInterface.register behavior changes." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both touch save_pretrained, but one is FP8 model serialization and the other is processor file saving; different code paths." }, { "left": "issue:27957", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 on XLM-Roberta is unrelated to MLFlow reporting on failed training jobs." + "reason": "XLMRoberta Flash Attention 2 support is unrelated to MLflow reporting after training failures." }, { "left": "issue:30333", "right": "issue:34689", "accept": false, - "reason": "Training-job reporting and Llama 3.2 Vision model loading are unrelated." + "reason": "Failed-job reporting in MLflow is unrelated to a Llama 3.2 vision model loading regression." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Default sub-config initialization and non-persistent buffer filling are different issues." + "reason": "Defaulting multimodal configs to None and junk-filling non-persistent buffers are different initialization bugs." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FP8 accuracy concerns and processor save_pretrained missing files are unrelated." + "reason": "FP8 accuracy concerns are unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 inference failure on B200 and missing chat template on a processor are different issues." }, { "left": "issue:30333", "right": "issue:36331", "accept": false, - "reason": "MLFlow failure reporting is unrelated to the CustomTrainer signature breakage." + "reason": "MLflow reporting on failure is unrelated to a CustomTrainer signature change." }, { "left": "issue:30333", "right": "issue:35141", "accept": false, - "reason": "Failed-job MLFlow reporting is unrelated to token-embedding resizing behavior." + "reason": "Training-job reporting and output embedding reinitialization are unrelated bugs." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Missing transformers.onnx module and missing PyTorch dependency are different import-time failures." + "reason": "Missing transformers.onnx module and missing PyTorch dependency errors are different import failures." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "PretrainedConfig type-checking regression and LlamaConfig head_dim validation are distinct config issues." } ] }, @@ -12971,12 +13136,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:30990", "issue:38175", "issue:41628", "issue:42757", + "issue:42831", "issue:42907", "issue:43122", "issue:43232", @@ -13003,7 +13169,6 @@ "issue:44871", "issue:44977", "issue:44991", - "issue:45071", "issue:45092", "issue:45103", "issue:45245", @@ -13023,12 +13188,11 @@ "issue:46424" ], "soft_pairs": [ - "issue:45071|issue:46082", "issue:45245|issue:46132", - "issue:43122|issue:46209", "issue:44297|issue:45310", - "issue:41628|issue:42757", + "issue:43122|issue:46209", "issue:43475|issue:46305", + "issue:41628|issue:42757", "issue:44190|issue:46082", "issue:43723|issue:44977", "issue:44206|issue:45103", @@ -13044,193 +13208,188 @@ "issue:44483|issue:45406", "issue:44279|issue:46328", "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:43644|issue:46305", "issue:44829|issue:46002", - "issue:44261|issue:45092" + "issue:44261|issue:45092", + "issue:42831|issue:45310" ], - "prepared_review_unit_hash": "cc33674823a871156e0cb2d8800fb182ade56d541d5f6029a1583574eb53bf7a", + "prepared_review_unit_hash": "2c18871f235082c312ffb436c7e5e5a34a49af8a15f0b6d5383327af19336602", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36136, - "estimated_input_tokens": 9034, - "estimated_eval_tokens": 18324 + "serialized_chars": 36130, + "estimated_input_tokens": 9033, + "estimated_eval_tokens": 18322 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most items are unrelated Transformers v5 regressions or model-specific bugs that only share broad themes like loading, saving, generation, or config handling. None of the soft pairs look like the same underlying issue, so I would not merge any of them.", - "confidence": 0.37, - "canonical_issue_reason": "If one issue must serve as the representative, issue 43723 is the broadest and most central v5 loading regression in this set, but the cluster does not really have a true duplicate canonical.", + "summary": "This set is mostly a grab bag of unrelated Transformers issues. Several are broadly about v5 loading/saving regressions, but the soft pairs are not the same concrete bug, so none should be merged as duplicates.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43723 is the best cluster anchor: it is a clear, user-facing v5 tokenizer-loading regression in a generic API path, making it more representative than the model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43723 is the best overall representative because it is a concrete, user-facing tokenizer-loading regression on the common pretrained loading path; the other issues are mostly narrower, model-specific, or unrelated.", + "best_issue_reason": "issue:43723 is the most broadly applicable and easiest-to-track issue among this set; it captures a common v5 loading failure without being tied to one model family.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45071", - "right": "issue:46082", + "left": "issue:45245", + "right": "issue:46132", "accept": false, - "reason": "Different bugs: PreTrainedConfig type-checking regression vs LlamaConfig head_dim validation error." + "reason": "Different bugs: category-count runtime error vs attention registration behavior change." }, { - "left": "issue:45245", - "right": "issue:46132", + "left": "issue:44297", + "right": "issue:45310", "accept": false, - "reason": "Unrelated failures: category-count limit in tabular encoding vs AttentionInterface.register behavior change." + "reason": "Both involve model/tokenizer loading, but one is tokenizer config serialization and the other is a Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Different code paths and symptoms: tokenizer output changes vs FP8 inference failure on B200." + "reason": "Different failure modes: tokenizer output changes vs FP8 inference failure on B200." }, { - "left": "issue:44297", - "right": "issue:45310", + "left": "issue:43475", + "right": "issue:46305", "accept": false, - "reason": "Both involve loading/saving, but the concrete bugs differ: tokenizer_class mismatch vs Qwen3.5 MoE from_pretrained failure." + "reason": "SAM3 video encoder attribute error is unrelated to MoonshineStreaming window masking off-by-one." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different symbols and different dependencies; not the same underlying breakage." - }, - { - "left": "issue:43475", - "right": "issue:46305", - "accept": false, - "reason": "Different model/components: SAM3 video encoder attribute error vs MoonshineStreaming sliding-window off-by-one." + "reason": "Both are import errors, but they concern different symbols and different dependency paths." }, { "left": "issue:44190", "right": "issue:46082", "accept": false, - "reason": "Unrelated: local dataset loading script issue vs LlamaConfig configuration validation." + "reason": "Dataset loading in an example script is unrelated to LlamaConfig head_dim validation." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Different problems: tokenizer loading regression vs Qwen3.5 flash-attention generation behavior." + "reason": "Both are tokenizer/model loading complaints, but one is a generic AutoTokenizer issue and the other is Qwen3.5 generation with flash-attention." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "Different failure modes: unsupported center arg in feature extractor vs docstring processing crash with future annotations." + "reason": "One is a LasrFeatureExtractor argument regression; the other is an auto_docstring crash with future annotations." }, { "left": "issue:45381", "right": "issue:46305", "accept": false, - "reason": "Both are index/position-related, but they affect different pipelines and tensors; not the same bug." + "reason": "Different sequence bugs: Qwen2.5-VL vision position ids vs Moonshine streaming mask off-by-one." }, { "left": "issue:43232", "right": "issue:46424", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to cache max_length honoring." + "reason": "Generation kwargs update after sync_gpus is unrelated to cache max length honoring." }, { "left": "issue:44734", "right": "issue:46421", "accept": false, - "reason": "Both mention cache, but one is wrong tensor indexing during continuation and the other is lazy cache init causing recompilation." + "reason": "KV-cache continuation tensor indexing crash is a different issue from lazy cache init causing recompilation." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "Different save_pretrained concerns: FP8 moe serialization vs shard size default question." + "reason": "FP8 MoE save_pretrained bug is unrelated to the 50GB shard-size default question." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Generation kwarg update bug is unrelated to Kosmos2.5 long OCR input indexing." + "reason": "Generation kwargs handling and Kosmos2.5 OCR index errors are different code paths." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Position_ids buffer read error vs processor.save_pretrained missing files are different save/load issues." + "reason": "Position_ids buffer access bug is unrelated to processor.save_pretrained missing files." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "torch.split return handling in a MoE indexer is unrelated to meta initialization compatibility for old InternVL2 checkpoints." + "reason": "torch.split return handling in a DSA indexer is not the same as remote-code meta-init incompatibility." }, { "left": "issue:44373", "right": "issue:44485", "accept": false, - "reason": "Wrong docstring is documentation-only; GLM-5 RoPE is a runtime implementation issue." + "reason": "Docstring mismatch for position_ids is unrelated to GLM-5 RoPE implementation correctness." }, { "left": "issue:43723", "right": "issue:44871", "accept": false, - "reason": "Tokenizer loading regression vs eos_token_id config mismatch; related area, but not the same underlying bug." + "reason": "Tokenizer loading regression is different from eos_token_id config inconsistency in Gemma-3." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "Different server-side failures: request rejection in /v1/chat/completions vs missing _tokenizer on Gemma4Processor." + "reason": "Chat/completions request rejection is unrelated to Gemma4Processor missing _tokenizer." }, { "left": "issue:44279", "right": "issue:46328", "accept": false, - "reason": "Dependency issue with transformers is unrelated to the numpy-to-list padding inefficiency." + "reason": "Generic dependency trouble is not the same as a performance bug in SequenceFeatureExtractor.pad." }, { "left": "issue:43723", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading reports, but they do not identify the same concrete failure mode or fix." + "reason": "Same broad area, but not the same bug: one is generic tokenizer loading, the other is a specific EMBEDDIA/est-roberta failure." }, { - "left": "issue:42907", - "right": "issue:46050", + "left": "issue:44263", + "right": "issue:44360", "accept": false, - "reason": "Different quantization paths and symptoms: dequantized Ministral/Devstral save failure vs sinq save/load breakage." + "reason": "Both mention DSA indexer, but one is split return values and the other is a missing ReLU." }, { - "left": "issue:30990", - "right": "issue:38175", + "left": "issue:42907", + "right": "issue:46050", "accept": false, - "reason": "Completely different models and bugs: Sentence Transformers load hang vs SigLIP2 zero probabilities." + "reason": "Saving dequantized Ministrals/Devstrals is unrelated to sinq quantization breaking after save/load." }, - { - "left": "issue:44263", - "right": "issue:44360", + { + "left": "issue:30990", + "right": "issue:38175", "accept": false, - "reason": "Same broad subsystem name, but different defects in the DSA indexer; not enough to merge." + "reason": "Sentence Transformers loading hang and zero probabilities in siglip2 are different model/runtime issues." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Old InternVL2 meta-init incompatibility is separate from Qwen3.5 save_pretrained visual key serialization." + "reason": "Both involve Qwen/InternVL save/load regressions, but the concrete failures and code paths differ." }, { "left": "issue:43761", @@ -13242,19 +13401,25 @@ "left": "issue:43644", "right": "issue:46305", "accept": false, - "reason": "Non-persistent buffer junk fill is unrelated to streaming mask off-by-one behavior." + "reason": "Non-persistent buffer junk in v5 is unrelated to a streaming attention window off-by-one." }, { "left": "issue:44829", "right": "issue:46002", "accept": false, - "reason": "Different training stability problems with different triggers and model paths." + "reason": "FlashAttention-3 training degeneration is different from NaNs during checkout/init." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "MLA layernorm epsilon precision issue is unrelated to InternVL2 meta-init checkpoint compatibility." + "reason": "MLA RMS norm epsilon precision issue is unrelated to remote-code checkpoint meta-init incompatibility." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FP8 accuracy issues on one side and Qwen3.5 MoE from_pretrained errors on the other are different bugs." } ] }, @@ -13276,7 +13441,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:22355", "issue:27957", @@ -13311,13 +13476,13 @@ "issue:44297", "issue:44368", "issue:44373", + "issue:44488", "issue:44521", "issue:44568", "issue:44704", "issue:44913", "issue:44945", "issue:45278", - "issue:45310", "issue:45593", "issue:45753", "issue:46002", @@ -13328,11 +13493,10 @@ "issue:46402" ], "soft_pairs": [ - "issue:42831|issue:45310", "issue:43479|issue:44913", "issue:41628|issue:43824", - "issue:43408|issue:46402", "issue:42831|issue:44038", + "issue:43408|issue:46402", "issue:44263|issue:44373", "issue:39401|issue:44568", "issue:22355|issue:27957", @@ -13351,222 +13515,223 @@ "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:30990|issue:33357", "issue:44188|issue:44373", + "issue:42907|issue:44945", + "issue:43122|issue:44521", "issue:43122|issue:46326", - "issue:41628|issue:45278" + "issue:41628|issue:45278", + "issue:39401|issue:44488" ], - "prepared_review_unit_hash": "505cb83f8d686abd8587ff79dddcba2f005c10348a54c751585cacad16978f3e", + "prepared_review_unit_hash": "946934485050c01537b0267472f10ff20faadd5e51ed33ee8825c836e596f00b", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36196, - "estimated_input_tokens": 9049, - "estimated_eval_tokens": 18354 + "serialized_chars": 36187, + "estimated_input_tokens": 9047, + "estimated_eval_tokens": 18350 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab bag of distinct Transformers regressions and API issues (imports, tokenizers, config reloads, model-specific bugs). The apparent similarities are broad subsystem overlaps, not the same concrete defect, so no soft-edge duplicate merges are justified.", - "confidence": 0.9, - "canonical_issue_reason": "issue:45278 is the broadest umbrella in the set for version-upgrade import regressions, but the cluster is too heterogeneous for a true duplicate canonical.", + "summary": "This is not a true duplicate cluster: it mixes unrelated tokenizer, import, quantization, generation, attention, and model-specific bugs. I rejected every soft edge; no single issue cleanly represents the rest.", + "confidence": 0.98, + "canonical_issue_reason": "No reliable canonical issue: the set is heterogeneous, so picking one issue as the duplicate anchor would be arbitrary.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the closest cluster anchor because it is the most general and representative of the import/API-regression subset, even though it does not actually subsume the other reports.", + "best_issue_reason": "No global best issue: none of these reports safely covers the rest of the cluster, so there is no strong representative anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42831", - "right": "issue:45310", - "accept": false, - "reason": "Different bugs: FP8 accuracy drift vs a Qwen3.5-Moe from_pretrained loading error." - }, { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Different config problems: multimodal default sub-config initialization vs GPTNeoX rotary_pct not persisting on reload." + "reason": "Different models and failures: Phi4MultimodalConfig default-init bug vs GPTNeoX rotary_pct reload regression." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Both are import failures, but for different symbols/models and different code paths." + "reason": "Both are import errors, but they involve different missing symbols and likely different breakpoints." }, { - "left": "issue:43408", - "right": "issue:46402", + "left": "issue:42831", + "right": "issue:44038", "accept": false, - "reason": "Different SAM3 issues: a model-type warning vs missing text-encoder weight loading." + "reason": "Accuracy issue in FineGrainedFP8 is unrelated to the Qwen3-VL-Moe/transformers 5.0 bug." }, { - "left": "issue:42831", - "right": "issue:44038", + "left": "issue:43408", + "right": "issue:46402", "accept": false, - "reason": "Unrelated: FP8 accuracy issue vs a Qwen3-VL-Moe transformer's from_pretrained bug." + "reason": "Same SAM3 family, but one is a warning mismatch and the other is a text-encoder weight-loading failure." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "Different scope: torch.split behavior in an indexer vs a docstring mistake for position_ids." + "reason": "A torch.split/indexing bug and a docstring bug are not the same underlying issue." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Both are tokenization-related, but one is wrong offset mappings and the other is missing BOS/EOS insertion." + "reason": "Both are tokenizer regressions, but offset_mapping and add_special_tokens/BOS-EOS behavior are different code paths." }, { "left": "issue:22355", "right": "issue:27957", "accept": false, - "reason": "Different areas: missing transformers.onnx module vs XLMRoberta Flash Attention 2 support." + "reason": "Module import failure for transformers.onnx vs XLMRoberta Flash Attention 2 support; unrelated problems." }, { "left": "issue:42757", "right": "issue:45278", "accept": false, - "reason": "45278 is a broad import-regression umbrella, but this is a specific huggingface_hub symbol import breakage." + "reason": "Specific huggingface_hub symbol import breakage vs a broad 'many import errors' report; not the same bug." }, { "left": "issue:27957", "right": "issue:30990", "accept": false, - "reason": "XLMRoberta Flash Attention 2 support is unrelated to Sentence Transformers hanging on load." + "reason": "Flash Attention 2 support and Sentence Transformers loading hangs are different issues." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Different tokenizer bugs: offline cache loading vs wrong decoder type in v5." + "reason": "Offline cache loading failure is different from wrong tokenizer decoder type in Transformers v5." }, { "left": "issue:44368", "right": "issue:46305", "accept": false, - "reason": "A warning about tie_word_embeddings is unrelated to the MoonshineStreaming off-by-one mask bug." + "reason": "A tied-embeddings warning and a sliding-window mask off-by-one are unrelated." }, { "left": "issue:22355", "right": "issue:28218", "accept": false, - "reason": "onnx import failure and tokenizer added-space behavior are unrelated bugs." + "reason": "Importing transformers.onnx and tokenizer added-token spacing are different bugs." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Offline tokenizer cache loading and quantized model saving failure are different code paths." + "reason": "Offline tokenizer cache loading and quantized model saving are separate code paths." }, { "left": "issue:42831", "right": "issue:45753", "accept": false, - "reason": "Different failures: accuracy regression in FineGrainedFP8 vs Qwen3_5 infinite loop on a specific image." + "reason": "FP8 accuracy regression and a Qwen3_5 image infinite loop are not the same underlying problem." }, { "left": "issue:43408", "right": "issue:46369", "accept": false, - "reason": "Both mention SAM3, but one is a warning about model type and the other is a memory overflow in a different model variant." + "reason": "SAM3 warning/instantiation mismatch vs memory overflow in sam3-litetext; different failure modes." }, { "left": "issue:43232", "right": "issue:45593", "accept": false, - "reason": "Generation kwargs after sync_gpus is unrelated to D-FINE auxiliary-loss behavior." + "reason": "Generation kwargs handling and D-FINE auxiliary-loss behavior are unrelated." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "A broken transformers env command is unrelated to a sentence-transformers import NameError." + "reason": "Broken transformers env command vs import-time NameError in sentence-transformers are different issues." }, { "left": "issue:43646", "right": "issue:46002", "accept": false, - "reason": "Custom model initialization breakage and NaNs in classification heads are different symptoms and likely different fixes." + "reason": "Custom model initialization breakage and NaNs in classification heads are distinct bugs." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Different model-specific regressions: Smollm3 RoPE layer handling vs Qwen3-Next tied-weights warning." + "reason": "Smollm3 RoPE-layer mismatch and Qwen3-Next tied-weights warning do not share the same code path." }, { "left": "issue:43550", "right": "issue:46132", "accept": false, - "reason": "torch.compile SDPA failure is not the same as AttentionInterface.register changing behavior." + "reason": "torch.compile/SDPA failure and AttentionInterface.register behavior change are different attention-related regressions." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "These are separate tokenizer/position-id issues: buffer access error vs tokenizer_class mismatch on save." + "reason": "Reading a registered position_ids buffer and tokenizer_class mismatch on save are unrelated." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None is a different failure from AutoProcessor not forwarding kwargs." + "reason": "TOKENIZER_MAPPING_NAMES None handling and AutoProcessor kwargs propagation are different loader bugs." }, { - "left": "issue:30990", - "right": "issue:33357", + "left": "issue:42371", + "right": "issue:44263", "accept": false, - "reason": "Sentence Transformers loading hangs and a MacOS bus error with CLIP are distinct runtime failures." + "reason": "TF32 settings guidance and torch.split return values are unrelated." }, { - "left": "issue:42371", - "right": "issue:44263", + "left": "issue:30990", + "right": "issue:33357", "accept": false, - "reason": "TF32 settings guidance and torch.split return values in GlmMoeDsaIndexer are unrelated." + "reason": "Loading hang and MacOS bus error are different runtime failures." }, { - "left": "issue:43122", - "right": "issue:44521", + "left": "issue:44188", + "right": "issue:44373", "accept": false, - "reason": "Different tokenizer behavior: version-dependent tokenization drift vs multimodal assistant_masks all-zero." + "reason": "torch.compile attention-kernel divergence and a wrong docstring are not the same issue." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Quantized model saving failure and incorrect pipeline-parallel LLM output are separate bugs." + "reason": "Quantized model saving failure and pipeline-parallelism output corruption are unrelated." }, { - "left": "issue:44188", - "right": "issue:44373", + "left": "issue:43122", + "right": "issue:44521", "accept": false, - "reason": "torch.compile attention-kernel divergence is unrelated to a docstring error." + "reason": "Tokenization differs across versions, but that is not the same as all-zero assistant_masks for multimodal chat templates." }, { "left": "issue:43122", "right": "issue:46326", "accept": false, - "reason": "Tokenization drift across versions and continue_final_message for mistral_common tokenizers are different defects." + "reason": "Both involve tokenizer behavior, but different_final_message handling and version-to-version tokenization drift are different bugs." }, { "left": "issue:41628", "right": "issue:45278", "accept": false, - "reason": "Both are import problems, but 45278 is a broad report of many import errors while 41628 is a specific missing AutoImageProcessor symbol." + "reason": "A specific missing export is not the same as a broad import-error regression report." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Wrong offset_mapping and failure to load a specific model are unrelated tokenizer/model-loading problems." } ] }, @@ -13588,7 +13753,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:22355", "issue:29127", @@ -13621,7 +13786,6 @@ "issue:44360", "issue:44373", "issue:44485", - "issue:44488", "issue:44514", "issue:44521", "issue:44625", @@ -13640,29 +13804,28 @@ "issue:45753" ], "soft_pairs": [ - "issue:42907|issue:43937", - "issue:39401|issue:44488", "issue:44261|issue:44373", + "issue:42907|issue:43937", "issue:43976|issue:44945", "issue:43784|issue:45600", "issue:44704|issue:45290", "issue:42907|issue:44964", "issue:39401|issue:44521", - "issue:43116|issue:45200", "issue:44038|issue:44297", - "issue:43122|issue:44514", + "issue:43116|issue:45200", "issue:44625|issue:44829", "issue:44360|issue:44485", "issue:44038|issue:44295", + "issue:43122|issue:44514", "issue:43408|issue:43475", - "issue:42907|issue:43452", "issue:43122|issue:44779", + "issue:42907|issue:43452", "issue:44038|issue:44222", "issue:45216|issue:45478", - "issue:43824|issue:44351", "issue:36296|issue:43116", - "issue:42757|issue:44351", + "issue:43824|issue:44351", "issue:22355|issue:29159", + "issue:42757|issue:44351", "issue:43278|issue:44945", "issue:42175|issue:44938", "issue:43493|issue:45753", @@ -13671,214 +13834,208 @@ "issue:22355|issue:29127", "issue:43992|issue:44704" ], - "prepared_review_unit_hash": "61f86d457a6a738d8d5600be048587a42e89c5ef077b34a937c47a9d1741eb7f", + "prepared_review_unit_hash": "88c21cfb32505dcd25a46bf357bf45caeb8e22e0b28ece628b0c8de81a5539b5", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36310, - "estimated_input_tokens": 9078, - "estimated_eval_tokens": 18412 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35462, + "estimated_input_tokens": 8866, + "estimated_eval_tokens": 17988 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most candidate edges are not true duplicates; they connect issues from different subsystems, models, or failure modes. I reject all soft edges. If one issue has to represent the cluster, issue 43122 is the broadest tokenizer-regression report, but it is not a duplicate of the others.", - "confidence": 0.66, - "canonical_issue_reason": "Issue 43122 is the broadest, most representative regression report in this set (tokenization changes across versions), while most other issues are narrower, model-specific, or unrelated.", + "summary": "The set is mostly heterogeneous bug reports across unrelated models, tokenizers, import paths, and training/runtime paths. None of the soft-similarity pairs look like the same underlying issue, so I would not merge any of them as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue fits the whole set because the items span unrelated subsystems and symptoms.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43122 best serves as the cluster\u2019s representative because it captures a generic tokenizer-regression pattern and has more cross-references than the other candidates.", + "best_issue_reason": "issue:44038 is the strongest standalone bug report candidate: it is model-specific, regression-shaped, and concrete enough to anchor follow-up triage.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42907", - "right": "issue:43937", - "accept": false, - "reason": "Different failure modes: dequantized save/load behavior vs GenerationConfig validation." - }, - { - "left": "issue:39401", - "right": "issue:44488", + "left": "issue:44261", + "right": "issue:44373", "accept": false, - "reason": "Qwen3 tokenizer offset mapping bug and a separate model-loading failure are unrelated." + "reason": "One is a config/precision bug in MLA q_a_layernorm; the other is a docstring problem for position_ids. Different symptoms and fixes." }, { - "left": "issue:44261", - "right": "issue:44373", + "left": "issue:42907", + "right": "issue:43937", "accept": false, - "reason": "LayerNorm epsilon precision issue and a docstring problem are not the same bug." + "reason": "Dequantized save_pretrained failure and invalid GenerationConfig are unrelated code paths." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python-version compatibility and pipeline-parallelism output corruption are distinct issues." + "reason": "Python-version support breakage is not the same as incorrect output under pipeline parallelism." }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Import-time NameError in sentence-transformers is unrelated to stale auto_mappings references." + "reason": "Both are import/compatibility related, but they reference different missing pieces in different integration paths." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "cached_file kwargs forwarding and chat-template tool-call crashes are different code paths." + "reason": "cached_file kwargs propagation and chat-template tool-call crashes are distinct API bugs." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "Dequantized save_pretrained failure and Phi-4 multimodal loading failure are unrelated." + "reason": "Both involve loading/saving, but they are different models and different failure modes." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Offset mapping correctness and multimodal assistant_masks generation are different tokenizer behaviors." - }, - { - "left": "issue:43116", - "right": "issue:45200", - "accept": false, - "reason": "Example-script empty outputs and mm_token_type_ids defaults are not the same underlying bug." + "reason": "Wrong offset_mapping and all-zero assistant masks are separate tokenizer/template behaviors." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe issue and tokenizer_class serialization mismatch are separate problems." + "reason": "A Qwen3-VL-Moe bug and tokenizer_class metadata mismatch are not the same underlying issue." }, { - "left": "issue:43122", - "right": "issue:44514", + "left": "issue:43116", + "right": "issue:45200", "accept": false, - "reason": "Both are tokenizer-related, but one is a version-wide regression and the other is a batched chat-template crash." + "reason": "Classification example output failure and Gemma 4 mm_token_type_ids defaulting are unrelated." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "num_labels propagation and flash_attention_3 training degeneration are unrelated." + "reason": "num_labels propagation and flash_attention_3 degenerate training affect different paths." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "DSA indexer activation bug and GLM-5 RoPE implementation issue are not duplicates." + "reason": "A missing ReLU in one component is not the same as a GLM-5 RoPE implementation discussion." }, { "left": "issue:44038", "right": "issue:44295", "accept": false, - "reason": "Qwen3-VL-Moe loading behavior and position_ids buffer access are unrelated." + "reason": "Qwen3-VL-Moe bug vs position_ids buffer access error are different code paths." }, { - "left": "issue:43408", - "right": "issue:43475", + "left": "issue:43122", + "right": "issue:44514", "accept": false, - "reason": "Model-type warning and missing fpn_position_embeddings attribute are separate SAM3 issues." + "reason": "Tokenizer-version output differences and batched chat-template crashes are different bugs." }, { - "left": "issue:42907", - "right": "issue:43452", + "left": "issue:43408", + "right": "issue:43475", "accept": false, - "reason": "Dequantized save/load behavior and gguf_file loading failures do not share the same bug." + "reason": "Model-type warning and missing fpn_position_embeddings are distinct SAM 3 Video problems." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both mention tokenizer regressions, but they affect different models and likely different code paths." + "reason": "Both are tokenization-related, but they describe different models and different regressions." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "save_pretrained for dequantized models and gguf_file loading are unrelated failure modes." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "Qwen3-VL-Moe load issue and FP8 save_pretrained moe issue are different concrete failures." + "reason": "Same broad model family, but one is a general transformers/Qwen3-VL-Moe bug and the other is FP8 save_pretrained behavior." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Related model/version, but save_pretrained corruption and from_pretrained load error are not clearly the same bug." - }, - { - "left": "issue:43824", - "right": "issue:44351", - "accept": false, - "reason": "Missing Qwen2_5_VLForConditionalGeneration import and missing HybridCache import are unrelated import errors." + "reason": "Checkpoint save correctness and from_pretrained load errors are separate operations." }, { "left": "issue:36296", "right": "issue:43116", "accept": false, - "reason": "Tensor parallel training bug and multi-label classification example behavior are unrelated." + "reason": "Tensor parallel training bug and example script multi-label output bug are not the same issue." }, { - "left": "issue:42757", + "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "huggingface_hub import failure and transformers import failure are different missing-symbol problems." + "reason": "Missing exports for Qwen2_5_VLForConditionalGeneration and HybridCache are different missing symbols." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "Missing transformers.onnx module and slow/fast tokenizer inconsistency are unrelated." + "reason": "Missing transformers.onnx import and slow/fast tokenizer inconsistency are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different packages and different missing names." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "Embedding dtype drift in evaluate and pipeline-parallelism output errors are not the same issue." + "reason": "dtype drift between training and evaluate is not the same as incorrect LLM output in pipeline parallelism." }, { "left": "issue:42175", "right": "issue:44938", "accept": false, - "reason": "Package-extras backend omission and Python 3.14 load failure are unrelated." + "reason": "Packaging backend omission and Python 3.14 load failure are distinct environment/setup issues." }, { "left": "issue:43493", "right": "issue:45753", "accept": false, - "reason": "SigLIP2 implementation discrepancy and Qwen3_5 infinite-loop behavior are different problems." + "reason": "SigLIP2 HF-vs-JAX discrepancy and Qwen3_5 infinite loop are different model behaviors." }, { "left": "issue:43066", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer regressions in v5, but the reported symptoms and models differ too much to treat as one bug." + "reason": "Wrong tokenizer decoder type in v5 and Deepseek tokenizer output regression are different tokenizer bugs." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leak in dataloader-thread usage and bfloat16 dtype mismatches are unrelated." + "reason": "VRAM leak in Qwen3ForCausalLM and dtype mismatches in CI inference are unrelated." }, { "left": "issue:22355", "right": "issue:29127", "accept": false, - "reason": "Missing module import and LayoutLMv3 error-message clarity are not duplicates." + "reason": "Import failure for transformers.onnx and LayoutLMv3 error-message clarity are not the same bug." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Missing embed_tokens.weight on load and cached_file kwarg forwarding are different issues." + "reason": "Missing embed_tokens.weight in UMT5Encoder loading and cached_file kwarg forwarding are unrelated." } ] }, @@ -13900,7 +14057,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:33453", "issue:38617", @@ -13910,7 +14067,6 @@ "issue:42673", "issue:43064", "issue:43116", - "issue:43122", "issue:43441", "issue:43450", "issue:43493", @@ -13930,6 +14086,7 @@ "issue:44291", "issue:44355", "issue:44380", + "issue:44410", "issue:44466", "issue:44485", "issue:44779", @@ -13939,6 +14096,7 @@ "issue:45081", "issue:45083", "issue:45278", + "issue:45310", "issue:45357", "issue:45405", "issue:45446", @@ -13948,7 +14106,6 @@ "issue:46002", "issue:46077", "issue:46231", - "issue:46298", "issue:46328" ], "soft_pairs": [ @@ -13958,11 +14115,11 @@ "issue:44188|issue:46328", "issue:44355|issue:46231", "issue:43116|issue:43493", - "issue:42222|issue:43441", "issue:42371|issue:43874", + "issue:42222|issue:43441", "issue:44075|issue:44936", - "issue:45357|issue:45478", "issue:44038|issue:45478", + "issue:45357|issue:45478", "issue:43976|issue:45446", "issue:43116|issue:43825", "issue:45083|issue:45446", @@ -13976,221 +14133,227 @@ "issue:43827|issue:45446", "issue:43450|issue:43986", "issue:43673|issue:46328", - "issue:44291|issue:45071", "issue:44466|issue:46002", + "issue:44291|issue:45071", "issue:43937|issue:44977", "issue:45405|issue:45446", "issue:44117|issue:45081", - "issue:43122|issue:46298" + "issue:44410|issue:45310" ], - "prepared_review_unit_hash": "c3334f1916fd2b48bc42a1bc07591c99501a68ccad829ca96ec1978705874c1d", + "prepared_review_unit_hash": "c9a51276c1137ef49e0445648a55bae0c8fb3aeebc576a45286b565fd66e2bb4", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36566, - "estimated_input_tokens": 9142, - "estimated_eval_tokens": 18540 + "serialized_chars": 36585, + "estimated_input_tokens": 9147, + "estimated_eval_tokens": 18550 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is a heterogeneous collection of Transformers v5 regression issues across unrelated subsystems (tokenizers, model loading, training, attention, docs, etc.). The soft links are mostly superficial title similarity rather than true duplicates. No PRs were provided.", - "confidence": 0.87, - "canonical_issue_reason": "issue:45278 is the broadest, most representative regression report in the set: it captures a general v5 upgrade breakage theme and is more suitable as an umbrella than the narrower model-specific bugs.", + "summary": "Mostly a heterogeneous mix of Transformers v5 regressions and model-specific bugs. I do not see any soft pair that is the same underlying issue; the only clear duplicate-like item is the repeated Qwen3.5 MoE from_pretrained report (45310/45478), but it is not part of the provided soft-edge list.", + "confidence": 0.77, + "canonical_issue_reason": "Issue 45310 is the best canonical anchor only for the narrow Qwen3.5 MoE from_pretrained duplicate pair, since 45478 repeats the same title later. It is not a good canonical for the full cluster because the rest of the items are unrelated.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best single issue representative because it is high-level, version-focused, and likely the easiest to route or cluster against compared with the many narrow, model-specific reports.", + "best_issue_reason": "45310 is the strongest issue anchor overall because it is an exact duplicate-style report with a clear, concrete failure mode. Even so, it should only represent that small subtopic, not the whole heterogeneous cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Both are import-related, but 45278 is a broad upgrade regression while 38617 is a specific missing symbol ImportError; not the same concrete bug." + "reason": "Both are import-related v5 regressions, but one is a specific missing symbol and the other is a broad upgrade-breakage report." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "Same model family (GLM-5) but different failures: invalid GenerationConfig vs RoPE implementation discussion." + "reason": "Both touch GLM-5, but one is a generation config validation error and the other is a RoPE implementation discussion." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Unrelated bugs: model loading with base_model_prefix vs NaNs in classification heads." + "reason": "Base-model loading breakage and NaNs in classification heads are different code paths and symptoms." }, { "left": "issue:44188", "right": "issue:46328", "accept": false, - "reason": "Different subsystems entirely: attention kernel divergence vs SequenceFeatureExtractor.pad performance." + "reason": "Attention-kernel divergence and feature-extractor padding performance are unrelated bugs." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "Compiled Python file errors are unrelated to the set_training() docs parameter mismatch." + "reason": "Compiled-file runtime errors are not the same as a documentation parameter-name mismatch." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Example script multi-label empty results and SigLIP2 implementation discrepancy are unrelated." + "reason": "Multi-label classification output logic and SigLIP2 model discrepancy are different problems." }, { - "left": "issue:42222", - "right": "issue:43441", + "left": "issue:42371", + "right": "issue:43874", "accept": false, - "reason": "Both are model-specific issues, but vitpose broken imports and Ministral-3 flash-attention failures are different bugs." + "reason": "TF32 API guidance and a missing image-processor method are unrelated issues." }, { - "left": "issue:42371", - "right": "issue:43874", + "left": "issue:42222", + "right": "issue:43441", "accept": false, - "reason": "TF32 API settings and a missing image-patch helper in GLM46V are unrelated changes." + "reason": "VitPose import breakage and Ministral FlashAttention failure are different model-specific regressions." }, { "left": "issue:44075", "right": "issue:44936", "accept": false, - "reason": "Optimizer argument handling is unrelated to trainer.evaluate() failing after trainer.train()." + "reason": "SGD argument handling and trainer.evaluate() failing after train() are distinct trainer bugs." }, { - "left": "issue:45357", + "left": "issue:44038", "right": "issue:45478", "accept": false, - "reason": "Same model family, but save_pretrained key serialization and from_pretrained loading failure are not the same concrete bug." + "reason": "Both involve Qwen3.5/Qwen3-VL-Moe, but one is a model-loading error and the other is a different from_pretrained regression." }, { - "left": "issue:44038", + "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Both mention Qwen3/Qwen3.5 MoE, but they concern different model variants and different failure modes." + "reason": "Both are Qwen3.5-related, but one is save_pretrained key serialization and the other is a load-time error." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version compatibility and a PyTorch version check bug are unrelated." + "reason": "Python-version compatibility and a flex_attention import version check are not the same failure." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Classification example output bug and pipeline removal/error-message issue are different problems." + "reason": "Classification example output and pipeline deprecation messaging are unrelated." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "Feature-extraction length behavior in qwen3_omni_moe is unrelated to the flex_attention import/version check." + "reason": "A qwen3_omni_moe helper behavior bug is unrelated to a flex_attention dependency/version check." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Training/FSDP optimizer-state corruption and GenerationMixin cache missing in chunked_prefill are unrelated." + "reason": "Optimizer-state corruption under FSDP2/PEFT is not the same as missing GenerationMixin cache support." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Both are tokenizer regressions, but loading regression and wrong offset_mapping are distinct bugs." + "reason": "Both are tokenizer-related, but one is a generic loading regression and the other is an offset_mapping correctness bug." }, { "left": "issue:43937", "right": "issue:45563", "accept": false, - "reason": "GLM-5 invalid GenerationConfig and a stale warning for num_return_sequences are unrelated." + "reason": "GenerationConfig validation and a stale warning on num_return_sequences are different issues." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer correctness issues, but they affect different tokenizers and different failure modes." + "reason": "Both are tokenizer correctness bugs, but they affect different models and different failure modes." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES returning None is unrelated to the flash_attention crash on s_aux=None." + "reason": "TOKENIZER_MAPPING_NAMES handling and flash_attention s_aux=None crashes are unrelated subsystems." }, { "left": "issue:44380", "right": "issue:46077", "accept": false, - "reason": "Both involve attention behavior, but GPT2 backend scaling and Gemma4 bidirectional-mask construction are different bugs." + "reason": "Attention scaling being ignored and bidirectional masks being built are distinct attention bugs." }, { "left": "issue:42673", "right": "issue:43906", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM and a reproduction of another issue are not the same underlying bug." + "reason": "A VRAM leak in dataloader threads and an isolated reproduction of another issue are not enough to call the same bug." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Docs still referencing pipeline() and a flex_attention PyTorch version check are unrelated." + "reason": "Docs still referencing pipeline() and a flex_attention import version check are unrelated." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "Video processor batched-shape bug and AutoProcessor crash without torchvision are different issues." + "reason": "Batched video shape errors and missing torchvision crashes are different video-processing problems." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Generation cache handling and SequenceFeatureExtractor.pad performance are unrelated." + "reason": "Chunked-prefill cache handling and padding conversion overhead are unrelated." }, { - "left": "issue:44291", - "right": "issue:45071", + "left": "issue:44466", + "right": "issue:46002", "accept": false, - "reason": "Both are model-loading/config-related regressions, but unexpected _is_hf_initialized argument and PretrainedConfig type checking are distinct problems." + "reason": "Serialization of tied weights and NaNs in classification heads are different bugs." }, { - "left": "issue:44466", - "right": "issue:46002", + "left": "issue:44291", + "right": "issue:45071", "accept": false, - "reason": "lm_head serialization inconsistency and NaNs in classification heads are unrelated." + "reason": "Both are loading/type-checking regressions, but the error causes and code paths differ." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "GLM-5 invalid GenerationConfig and Qwen3.5 flash-attention generation failure are different bugs." + "reason": "GenerationConfig validation and Qwen3.5 flash-attention generation failure are not the same issue." }, { "left": "issue:45405", "right": "issue:45446", "accept": false, - "reason": "Package version bumping for PEFT and a PyTorch version check in flex_attention are unrelated." + "reason": "A PEFT minimum-version bump and a PyTorch version check for AuxRequest import are unrelated." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Tokenizer mapping fallback bug and Mistral regex patch crash are unrelated." + "reason": "Tokenizer mapping None handling and Mistral regex patch crashes are different tokenizer bugs." }, { - "left": "issue:43122", - "right": "issue:46298", + "left": "issue:44410", + "right": "issue:45310", "accept": false, - "reason": "Tokenizer behavior change across versions and a noisy logging message are not the same issue." + "reason": "Missing QKV/gate projections in qwen3next and a Qwen3.5 Moe from_pretrained error are different model regressions." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "These are both Qwen3-family reports, but they concern different models and different failure surfaces." } ] }, @@ -14212,7 +14375,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:38617", "issue:39401", @@ -14238,11 +14401,9 @@ "issue:44075", "issue:44220", "issue:44261", - "issue:44263", "issue:44410", "issue:44483", "issue:44484", - "issue:44485", "issue:44493", "issue:44496", "issue:44829", @@ -14261,255 +14422,242 @@ "issue:45656", "issue:46178", "issue:46231", + "issue:46298", "issue:46328" ], "soft_pairs": [ - "issue:44410|issue:45310", + "issue:43122|issue:46298", "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:42673|issue:43582", + "issue:42222|issue:46178", "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:42490|issue:42673", "issue:42371|issue:44261", "issue:44484|issue:45310", "issue:38617|issue:44220", "issue:43937|issue:44496", "issue:44933|issue:45278", "issue:43066|issue:43122", - "issue:42222|issue:45542", "issue:44410|issue:45478", - "issue:44483|issue:45464", + "issue:42222|issue:45542", "issue:42371|issue:42831", "issue:43012|issue:43577", "issue:44075|issue:45656", + "issue:44483|issue:45464", "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:41628|issue:45071" ], - "prepared_review_unit_hash": "5206a7682bbcdcc2d031ebac25621e2a3f0dcc6acf586bb8bf9e69d42b191901", + "prepared_review_unit_hash": "7a27e7f980d90c2a7ee29a70a81dae3329f77dbfdfbf8ab84308d6feddfbab75", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 36531, - "estimated_input_tokens": 9133, - "estimated_eval_tokens": 18522 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "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 is a loose thematic cluster of Transformers issues, not a true duplicate set: it mixes import/API regressions, tokenizer behavior changes, model-loading failures, backend/install problems, precision warnings, and testing utility bugs. I would not merge any of the soft edges.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue is safe to choose; the cluster is heterogeneous and the items are not the same underlying bug.", + "summary": "The soft pairs are mostly broad topical similarities, but they do not describe the same underlying bug or change. This looks like a heterogeneous issue set rather than one duplicate cluster.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "If forced, issue 45278 is the broadest representative of the import-regression subtheme, but it still does not represent most of the cluster.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44410", - "right": "issue:45310", + "left": "issue:43122", + "right": "issue:46298", "accept": false, - "reason": "Different Qwen bugs: missing projections in one model vs a from_pretrained loading error in another." + "reason": "Unrelated regressions: tokenizer output change vs a noisy logging message." }, { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Different test-utils problems: a generic bug in testing_utils.py vs a pytest-xdist file race." + "reason": "Both mention testing utilities, but one is a vague small bug and the other is a specific xdist file race; not the same bug." }, { "left": "issue:38617", "right": "issue:41628", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and different code paths." + "reason": "Different import failures for different symbols and code paths." }, { "left": "issue:45341", "right": "issue:46328", "accept": false, - "reason": "One is a correctness/race issue in test helpers; the other is a performance inefficiency in feature padding." - }, - { - "left": "issue:42222", - "right": "issue:46178", - "accept": false, - "reason": "Unrelated: vitpose model breakage vs a TensorFlow backend mapping error." + "reason": "Testing-utils bug vs SequenceFeatureExtractor.pad performance issue; no shared underlying defect." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "Different regressions: FA3 training degeneracy vs Qwen3.5 save_pretrained checkpoint corruption." + "reason": "Both relate to Qwen3.5/training behavior, but one is flash-attention training degeneration and the other is save_pretrained checkpoint corruption." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "A script execution problem is not the same bug as a testing_utils.py issue." + "reason": "Completely different areas: a script execution problem vs a testing-utils bug." }, { - "left": "issue:42673", - "right": "issue:43582", + "left": "issue:42222", + "right": "issue:46178", "accept": false, - "reason": "Different failures: VRAM leak under threaded loading vs a TypeError on Apple Silicon warmup." + "reason": "VitPose model breakage and tensorflow_text backend registration are unrelated." }, { "left": "issue:45362", "right": "issue:46328", "accept": false, - "reason": "Qwen chat crash and feature-extractor padding inefficiency are unrelated." + "reason": "Chat runtime failure vs audio feature padding performance; different subsystems." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM vs Apple Silicon TypeError in caching_allocator_warmup; not the same bug." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "A model-type warning is not the same as a key error when loading a specific checkpoint." + "reason": "Different model-loading problems for different models and causes." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Both touch tokenization, but offset_mapping is a different bug from runtime post-processor updates." + "reason": "Both are tokenizer-related, but offset_mapping is a separate bug from runtime post-processor updates." }, { - "left": "issue:42490", - "right": "issue:42673", + "left": "issue:44877", + "right": "issue:46231", "accept": false, - "reason": "A local-path import warning is unrelated to a multi-thread VRAM leak." + "reason": "Config-loading strictness issue vs documentation parameter mismatch." }, { - "left": "issue:44877", - "right": "issue:46231", + "left": "issue:42490", + "right": "issue:42673", "accept": false, - "reason": "Config loading strictness and a documentation parameter typo are different issues." + "reason": "Local-path import warning and a Qwen3 VRAM leak are unrelated." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 behavior guidance is unrelated to the MLA rms_norm_eps precision bug." + "reason": "TF32 API guidance and MLA rms_norm_eps precision error are different concerns." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "A question about save_pretrained shard size is not the same as a Qwen3.5 from_pretrained failure." + "reason": "save_pretrained shard-size question vs Qwen3.5 MoE from_pretrained regression; no shared code-path defect." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "ImportError for a missing symbol and an _torch_extract_fbank_features bug are different code paths." + "reason": "ImportError for a config symbol vs a feature-extraction bug." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "GenerationConfig validation failure is distinct from an unrecognized-model config.json error." + "reason": "GenerationConfig validation for GLM-5 vs model_type missing for a different checkpoint." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "Related theme of imports, but not the same concrete missing-symbol bug." + "reason": "A specific missing import and a broad upgrade regression are not the same bug." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Both are tokenizer regressions, but one is decoder type and the other is output tokenization differences." - }, - { - "left": "issue:42222", - "right": "issue:45542", - "accept": false, - "reason": "Different backend problems: model breakage vs TensorFlow being missing despite tensorboard being installed." + "reason": "Both are tokenizer regressions, but decoder-type mismatch and changed tokenization output are distinct issues." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Different Qwen3.5 failures with different symptoms and likely different code paths." + "reason": "Different Qwen model families and different loading failures." }, { - "left": "issue:44483", - "right": "issue:45464", + "left": "issue:42222", + "right": "issue:45542", "accept": false, - "reason": "Both involve chat/completions, but one is request validation and the other is streaming inference failure." + "reason": "Model breakage vs backend-installation problem; unrelated." }, { "left": "issue:42371", "right": "issue:42831", "accept": false, - "reason": "TF32 API guidance is unrelated to FineGrainedFP8 accuracy degradation." + "reason": "TF32 settings and FineGrainedFP8 accuracy are separate performance/precision topics." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "A compile-time float32 warning is not the same as model dtype actually remaining float32 on load." + "reason": "A compile-time warning and a model dtype-loading bug are different code paths." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "SGD argument handling and double optimizer stepping under deepspeed are different bugs." + "reason": "Both involve optimizers, but one is unused SGD args and the other is a double-step deepspeed behavior." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Different chat-completions failures with different symptoms and likely different causes." }, { "left": "issue:42175", "right": "issue:42222", "accept": false, - "reason": "Backend dependency installation and vitpose model breakage are unrelated." + "reason": "TensorFlow backend packaging issue vs VitPose model breakage." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Removed API warning and unexpected position-id keys are different regressions." + "reason": "Removed config attribute and unexpected position-id keys are not the same regression." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Chunked-prefill cache handling is unrelated to the testing_utils bug." + "reason": "Generation cache regression vs testing-utils bug; unrelated." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "An import error for AutoImageProcessor is not the same as a PreTrainedConfig type-checking regression." - }, - { - "left": "issue:44263", - "right": "issue:44485", - "accept": false, - "reason": "These are separate GLM-5 issues: torch.split return handling vs RoPE implementation details." - }, - { - "left": "issue:45071", - "right": "issue:45542", - "accept": false, - "reason": "Type-checking behavior and TensorFlow backend installation are unrelated." + "reason": "Import error for AutoImageProcessor vs PreTrainedConfig type-checking regression." } ] }, @@ -14531,7 +14679,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:27957", "issue:38617", @@ -14572,6 +14720,7 @@ "issue:45375", "issue:45440", "issue:45478", + "issue:45542", "issue:45588", "issue:45753", "issue:46050", @@ -14582,266 +14731,280 @@ "issue:46369" ], "soft_pairs": [ + "issue:44263|issue:44485", + "issue:45071|issue:45542", "issue:27957|issue:43352", "issue:43874|issue:44261", - "issue:44484|issue:45478", "issue:44263|issue:46172", - "issue:43931|issue:44410", + "issue:44484|issue:45478", "issue:44360|issue:46172", + "issue:43931|issue:44410", "issue:43441|issue:44829", "issue:43931|issue:44038", - "issue:43097|issue:43452", "issue:43335|issue:45071", + "issue:43097|issue:43452", "issue:45003|issue:45341", - "issue:43673|issue:46123", "issue:43986|issue:45341", + "issue:43673|issue:46123", "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:43531|issue:45341", "issue:45440|issue:46172", - "issue:43097|issue:46050", - "issue:43097|issue:45588", + "issue:43986|issue:45372", "issue:44261|issue:44485", - "issue:38617|issue:42757", + "issue:43097|issue:45588", + "issue:43097|issue:46050", "issue:43064|issue:44315", + "issue:38617|issue:42757", "issue:44062|issue:44117", "issue:44410|issue:45375", - "issue:43502|issue:43992", "issue:44038|issue:44410", + "issue:43502|issue:43992", "issue:42890|issue:46298", "issue:38617|issue:44351", "issue:43097|issue:43873" ], - "prepared_review_unit_hash": "a3064feab6b6b2bb8d58fba6a334efbcd8f87897c6297c714f750f37fd467f34", + "prepared_review_unit_hash": "1ae41e1e0f286754d97b9403bd0aaa7841e4e87e1c4991ec82c328a38b0eea75", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 33, - "serialized_chars": 36772, - "estimated_input_tokens": 9193, - "estimated_eval_tokens": 18642 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 37956, + "estimated_input_tokens": 9489, + "estimated_eval_tokens": 19234 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Transformers regressions: import/API breakages, model-specific loading mismatches, FlashAttention issues, and a few training/pipeline bugs. I don\u2019t see any soft pair that looks like the same concrete underlying bug, so all candidate duplicates are rejected.", - "confidence": 0.79, - "canonical_issue_reason": "No single issue is a strong canonical duplicate for the cluster; if one must be chosen, issue 43873 is the least-bad representative because it is the broadest active bug report with the most discussion/inbound references.", + "summary": "The cluster is mostly heterogeneous; none of the soft pairs looks like the same underlying bug/change. If a single anchor is needed, issue 43097 is the broadest core regression report among the set.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43097 is the most central, high-impact report here: it targets a core Trainer v5 regression, has discussion and inbound references, and is broader than the model-specific or one-off bugs around it.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43873 is the best representative by scope and activity, but only weakly: it covers a broad quantization/offloading regression and has the strongest signal of being a hub among these otherwise unrelated reports.", + "best_issue_reason": "issue:43097 is the best representative issue because it describes a concrete, cross-cutting failure in core training behavior rather than a narrow model-specific or documentation/test issue.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM, but one is about DSA indexer torch.split behavior and the other about RoPE implementation; different code paths and bugs." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "A PretrainedConfig type-checking regression is unrelated to a tensorboard-without-TF backend error." + }, { "left": "issue:27957", "right": "issue:43352", "accept": false, - "reason": "Both are FlashAttention-related, but they involve different models and different failure conditions; not the same code-path bug." + "reason": "Both involve Flash Attention 2 support, but they are model-specific requests/errors for different architectures and likely different fixes." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Different bugs in different components: image patch counting vs MLA q_a_layernorm config precision." + "reason": "Missing image-processor method and an MLA layernorm epsilon precision issue are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "These are different model families and different failures; no shared concrete bug." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "Unrelated symptoms: shard-size question vs Qwen3.5 MoE from_pretrained load error." + "reason": "A save_pretrained sharding question is not the same as a Qwen3.5 MoE from_pretrained loading error." }, { - "left": "issue:44263", + "left": "issue:44360", "right": "issue:46172", "accept": false, - "reason": "Different model internals and different failure modes; no shared underlying defect." + "reason": "A missing ReLU in the DSA indexer is unrelated to the deepseek-v4 transpose issue." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Both concern Qwen-family models, but one is a weight-shape mismatch and the other is missing projections in a different model; not the same bug." - }, - { - "left": "issue:44360", - "right": "issue:46172", - "accept": false, - "reason": "DSA ReLU concern is unrelated to the DeepSeek comb transpose issue." + "reason": "Both are Qwen3-related, but one is a weight-shape mismatch in Qwen3-VL-30B and the other is missing projections in qwen3next; not the same bug." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "Both mention FlashAttention, but one is model support/loading and the other is degenerate training with flash_attention_3; different problems." + "reason": "Both mention FlashAttention, but the model, failure mode, and code path differ." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Same general model family, but different concrete breakages: Qwen3-VL weight mismatch vs Qwen3-VL-MoE import/load failure." + "reason": "These are different Qwen3-VL/MoE loading problems, not a single concrete defect." }, { - "left": "issue:43097", - "right": "issue:43452", + "left": "issue:43335", + "right": "issue:45071", "accept": false, - "reason": "Different regressions: a removed config field vs gguf_file handling in tokenizer/model loading." + "reason": "SwitchTransformers sparse-layer construction is unrelated to PretrainedConfig type-checking." }, { - "left": "issue:43335", - "right": "issue:45071", + "left": "issue:43097", + "right": "issue:43452", "accept": false, - "reason": "SwitchTransformers layer creation bug is unrelated to PreTrainedConfig type checking." + "reason": "Tie-embeddings removal and gguf_file tokenizer/model loading are different regressions." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Unsafe sys.modules access is unrelated to the testing_utils bug." + "reason": "Unsafe sys.modules access in modeling_utils is unrelated to a small testing_utils bug." }, { - "left": "issue:43673", - "right": "issue:46123", + "left": "issue:43986", + "right": "issue:45341", "accept": false, - "reason": "Generation cache handling and MaskGenerationPipeline final-batch handling are different code paths." + "reason": "A torchvision dependency crash in AutoProcessor is unrelated to a testing utility issue." }, { - "left": "issue:43986", - "right": "issue:45341", + "left": "issue:43673", + "right": "issue:46123", "accept": false, - "reason": "Video AutoProcessor/torchvision crash is unrelated to the testing_utils issue." + "reason": "Generation cache handling and mask pipeline partial-batch handling are different subsystems and bugs." }, { "left": "issue:44933", "right": "issue:46231", "accept": false, - "reason": "Nonexistent import from image_utils is a code bug; set_training docs parameter mismatch is documentation-only." + "reason": "A missing import from image_utils is unrelated to a documentation parameter-name typo." }, { "left": "issue:45753", "right": "issue:46369", "accept": false, - "reason": "Different multimodal failures with different models and symptoms; no shared underlying bug." + "reason": "An infinite loop on a specific Qwen3_5 image and a SAM memory overflow are different model-specific failures." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "Tokenizer mapping None-handling is a different failure than loading physical-intelligence/fast." + "reason": "Tokenizer mapping returning None and a specific repo loading failure are not the same underlying issue." }, { "left": "issue:43986", "right": "issue:44484", "accept": false, - "reason": "Completely different topics: missing torchvision vs shard-size behavior." + "reason": "A processor/torchvision crash is unrelated to a save_pretrained shard-size discussion." }, { "left": "issue:43874", "right": "issue:44263", "accept": false, - "reason": "Unrelated GLM46V image patch API bug vs GlmMoeDsaIndexer split semantics." - }, - { - "left": "issue:43986", - "right": "issue:45372", - "accept": false, - "reason": "Missing torchvision crash is unrelated to the mistral_common ReasoningEffort import error." + "reason": "These are unrelated GLM46V image-processing and DSA-indexer bugs." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior has no connection to the testing_utils bug." + "reason": "A sliding_window bug in Qwen3-MoE is unrelated to a small testing utility bug." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "DeepSeekV3MoE divergence from remote implementation is a broader model parity issue, not the same as the comb transpose bug." + "reason": "DeepseekV3MoE divergence from remote implementation is not the same as the deepseek-v4 comb transpose issue." }, { - "left": "issue:43097", - "right": "issue:46050", + "left": "issue:43986", + "right": "issue:45372", "accept": false, - "reason": "Both are save/load related, but one is an API removal and the other is a quantized model persistence regression." + "reason": "Torchvision-missing processor loading and a mistral_common ReasoningEffort import error are different failures." }, { - "left": "issue:43097", - "right": "issue:45588", + "left": "issue:44261", + "right": "issue:44485", "accept": false, - "reason": "Removed tie_embeddings API and flash_attention s_aux=None crash are unrelated bugs." + "reason": "A layernorm epsilon precision discussion is unrelated to the max_shard_size question." }, { - "left": "issue:44261", - "right": "issue:44485", + "left": "issue:43097", + "right": "issue:45588", "accept": false, - "reason": "RMS norm epsilon precision discussion and GLM-5 RoPE implementation are different model math issues with no clear shared defect." + "reason": "Removal of tie_embeddings_and_encoder_decoder and a flash_attention.py sinkless-model crash are different code paths." }, { - "left": "issue:38617", - "right": "issue:42757", + "left": "issue:43097", + "right": "issue:46050", "accept": false, - "reason": "Both are import errors, but they involve different missing symbols from different modules and do not look like the same bug." + "reason": "A config/API removal issue is unrelated to a quantized save/load regression." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Optimizer-state corruption under FSDP2/PEFT is unrelated to Liger Kernel not being applied with model_init." + "reason": "Wrong optimizer states on FSDP2/PEFT ranks is unrelated to Liger Kernel application with model_init." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols and modules; no shared root cause is evident." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "AddedToken argument duplication and tokenizer mapping None are distinct tokenizer-layer bugs." + "reason": "A duplicate special kwarg in AddedToken is unrelated to TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:44410", "right": "issue:45375", "accept": false, - "reason": "Missing attn projections in qwen3next and dropped deepstack_visual_indexes in Qwen3_5MoeVisionConfig are different config/model issues." + "reason": "Missing qwen3next projections and a missing deepstack_visual_indexes config field are different model/config bugs." }, { - "left": "issue:43502", - "right": "issue:43992", + "left": "issue:44038", + "right": "issue:44410", "accept": false, - "reason": "local_files_only network leakage and missing embed_tokens.weight are unrelated loading bugs." + "reason": "Both are Qwen3-family issues, but they affect different model variants and different loading paths." }, { - "left": "issue:44038", - "right": "issue:44410", + "left": "issue:43502", + "right": "issue:43992", "accept": false, - "reason": "Both touch Qwen3 variants, but one is a Qwen3-VL-Moe from_pretrained issue and the other is missing projections in qwen3next; not the same code path." + "reason": "local_files_only leaking API calls is unrelated to a missing embed_tokens.weight in UMT5Encoder." }, { "left": "issue:42890", "right": "issue:46298", "accept": false, - "reason": "Test seeding failures and noisy logging are unrelated." + "reason": "A flaky integration test lacking set_seed and a noisy logging message are unrelated." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "Different missing imports from different APIs; same symptom type, not the same underlying bug." + "reason": "Different missing imports from different modules; not the same bug." }, { - "left": "issue:43064", + "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "FSDP2/PEFT optimizer-state corruption and quantization offloading are different training-system issues." + "reason": "Tie-embeddings removal and quantized offloading behavior are different regression classes." } ] }, @@ -14863,10 +15026,11 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36683", "issue:38617", + "issue:42175", "issue:42490", "issue:42757", "issue:42907", @@ -14886,11 +15050,10 @@ "issue:43976", "issue:43986", "issue:43992", - "issue:44038", + "issue:44077", "issue:44117", "issue:44162", "issue:44222", - "issue:44230", "issue:44368", "issue:44380", "issue:44483", @@ -14904,7 +15067,7 @@ "issue:45103", "issue:45478", "issue:45479", - "issue:45600", + "issue:45584", "issue:45712", "issue:45753", "issue:46132", @@ -14915,6 +15078,7 @@ "issue:46421" ], "soft_pairs": [ + "issue:44936|issue:46365", "issue:44704|issue:45103", "issue:44380|issue:46421", "issue:43986|issue:44117", @@ -14933,241 +15097,241 @@ "issue:43352|issue:44380", "issue:43352|issue:43976", "issue:44162|issue:46172", + "issue:45584|issue:46365", + "issue:44077|issue:46365", "issue:44380|issue:45479", "issue:44380|issue:46132", "issue:43064|issue:44380", - "issue:44483|issue:44514", "issue:44222|issue:45478", + "issue:44483|issue:44514", "issue:36683|issue:43582", "issue:43208|issue:43526", - "issue:44936|issue:46365", + "issue:42175|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" + "issue:38617|issue:43502" ], - "prepared_review_unit_hash": "f478899ba0bb9a519c3823b9164d8a5cf1857ac2711d799affcdf3e77ebd17b3", + "prepared_review_unit_hash": "7807488d83a7d61986eab1433db7d6db7d5cf2c1a914f9dd58e866cea4e05d77", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37089, - "estimated_input_tokens": 9273, - "estimated_eval_tokens": 18802 + "soft_pair_count": 33, + "serialized_chars": 37445, + "estimated_input_tokens": 9362, + "estimated_eval_tokens": 18980 }, "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, so I would not merge any of these issues as duplicates. The cluster is mostly a mix of unrelated import errors, model-specific regressions, and a few attention/compile/config bugs.", - "confidence": 0.84, - "canonical_issue_reason": "Issue 45103 is the cleanest standalone bug report: it has a precise crash condition, a clear subsystem, and a concrete reproduction path, making it the best canonical issue among this set.", + "summary": "The set is heterogeneous: most soft pairs are superficially similar regression/import/runtime reports but not the same underlying bug. No PRs are present. I would not merge any soft edge pair.", + "confidence": 0.16, + "canonical_issue_reason": "issue:43824 is the clearest, most self-contained bug report in the set, with a specific import failure and strong impact signal; it is the best representative of the overall import/regression theme, even though the cluster is not truly duplicate-heavy.", "canonical_pr_reason": null, - "best_issue_reason": "45103 is the most suitable representative issue because it is narrowly scoped, actionable, and currently open, while the rest are largely unrelated single-bug reports.", + "best_issue_reason": "issue:43824 is the most actionable and representative issue here: concrete failure mode, clear reproduction context, and broader relevance than the other candidates.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "Different failures: trainer lifecycle/eval breakage vs an import dependency conflict affecting Trainer import." + }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "Different subsystems and failures: AutoProcessor kwargs propagation vs auto_docstring crash with future annotations." + "reason": "Different code paths: cached_file kwargs propagation vs auto-docstring annotation handling." }, { "left": "issue:44380", "right": "issue:46421", "accept": false, - "reason": "One is GPT2 attention scaling ignored in SDPA/FlashAttention; the other is lazy cache init causing recompilation with chunked prefill." + "reason": "Attention backend config bug vs cache lazy initialization/recompilation behavior; not the same defect." }, { "left": "issue:43986", "right": "issue:44117", "accept": false, - "reason": "Video AutoProcessor/torchvision import crash is unrelated to TOKENIZER_MAPPING_NAMES returning None." + "reason": "Video-model AutoProcessor crash without torchvision is unrelated to TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:43824", "right": "issue:43986", "accept": false, - "reason": "Qwen2_5_VL class import failure is distinct from AutoProcessor failing without torchvision." + "reason": "Both are import-related, but one is a missing class export and the other is an optional-dependency crash." }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 import/load failure is unrelated to a backend mapping error for tensorflow_text." + "reason": "Python-version loading failure and invalid backend mapping are unrelated." }, { "left": "issue:43066", "right": "issue:43906", "accept": false, - "reason": "Wrong tokenizer decoder type is not the same as an isolated reproduction of another issue." + "reason": "The second item is only a reproduction issue reference, not evidence of the same tokenizer bug." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM training blockers and Bamba SDPA compile failure are different models and code paths." + "reason": "xLSTM training blockers and Bamba torch.compile/SDPA failure are distinct model-specific issues." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing embed_tokens weight on UMT5Encoder is unrelated to max_shard_size behavior in save_pretrained()." + "reason": "Missing encoder weights on load vs an unrelated save_pretrained shard-size question." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "A PyTorch precision warning during compile is not the same bug as Bamba SDPA compile failure." + "reason": "Precision warning during bfloat16 compilation is not the same as a Bamba SDPA compile failure." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "Float32 warning on compile and tie_word_embeddings warning are unrelated configuration/runtime behaviors." + "reason": "Different warnings and different triggers: bfloat16 precision warning vs tie_word_embeddings warning." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "Translation-task pipeline error is unrelated to saving Pipeline objects with save_pretrained()." + "reason": "Pipeline translation messaging bug vs inability to save Pipeline objects." }, { "left": "issue:45712", "right": "issue:46298", "accept": false, - "reason": "Leftover dummy classes failing repo checks is not the same as a noisy logging message." + "reason": "Dummy class cleanup/check_repo issue is unrelated to a noisy logging message." }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Dequantized model save failure is different from generic pipeline serialization failure." - }, - { - "left": "issue:42907", - "right": "issue:45753", - "accept": false, - "reason": "Model save failure and Qwen3_5 image infinite loop are unrelated." + "reason": "Saving dequantized models and saving Pipeline objects are different persistence paths." }, { "left": "issue:43097", "right": "issue:46298", "accept": false, - "reason": "Removed tie_embeddings_and_encoder_decoder is a behavior/regression issue, not a logging-noise issue." + "reason": "A removed config field is not the same as a noisy logging message." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "Unsupported Flash Attention 2.0 in NemotronH is unrelated to GPT2 attention scaling being ignored." + "reason": "Unsupported Flash Attention on Nemotron is a model capability issue, not GPT2 attention scaling being ignored." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "Flash Attention support in a model is unrelated to a Python version compatibility problem." + "reason": "Unsupported Flash Attention and a Python-version compatibility issue are unrelated." }, { "left": "issue:44162", "right": "issue:46172", "accept": false, - "reason": "ESM2 breakage and a DeepSeek tensor transpose issue do not share the same failure mode." + "reason": "ESM2 breakage and a deepseek tensor transpose bug do not appear to share a code path." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper empty-transcription generation failure is unrelated to a Trainer import conflict." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "PatchTSMixer post_init validation and Trainer import failure are different problems." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "Attention scaling config loss is unrelated to a degenerate zero-loss classification bug." + "reason": "Attention scaling being ignored is not the same as degenerate zero loss in sequence classification." }, { "left": "issue:44380", "right": "issue:46132", "accept": false, - "reason": "GPT2 backend scaling bug is distinct from AttentionInterface.register changing function behavior." + "reason": "AttentionInterface.register behavior change is a different API regression than GPT2 backend scaling." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling being ignored." + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling." }, { - "left": "issue:44483", - "right": "issue:44514", + "left": "issue:44222", + "right": "issue:45478", "accept": false, - "reason": "Chat completions API rejection and Qwen2_5_VLProcessor batch-padding crash are different endpoints/code paths." + "reason": "FP8 save_pretrained behavior and Qwen3.5 MOE from_pretrained errors are distinct failures." }, { - "left": "issue:44222", - "right": "issue:45478", + "left": "issue:44483", + "right": "issue:44514", "accept": false, - "reason": "FP8 save_pretrained behavior is different from a Qwen3.5 MoE from_pretrained error." + "reason": "Chat completions API rejection and batched processor crash are different surfaces." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size and Apple Silicon caching_allocator_warmup TypeError are unrelated." + "reason": "Gemma3Config vocab_size AttributeError and Apple Silicon allocator TypeError are unrelated." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training bugs do not match BeitImageProcessorFast reduce_labels behavior." + "reason": "xLSTM training bugs and BeitImageProcessorFast reduce_labels behavior are unrelated." }, { - "left": "issue:44936", + "left": "issue:42175", "right": "issue:46365", "accept": false, - "reason": "trainer.evaluate() failing after train is not the same as failing to import Trainer until datasets is uninstalled." + "reason": "Tensorflow backend packaging and Trainer import conflict are not the same issue." }, { "left": "issue:42490", "right": "issue:43012", "accept": false, - "reason": "Trailing-slash local model import warning is unrelated to a bfloat16 compile warning." + "reason": "Local-model trailing-slash warning and bfloat16 precision warning are different behaviors." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "General compile warning and compiled flex_attention failure on newer torch versions are different bugs." + "reason": "Precision warning during compilation and compiled flex_attention failure are different bugs." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Missing is_offline_mode import from huggingface_hub is unrelated to missing Qwen2_5_VLForConditionalGeneration export." + "reason": "Missing is_offline_mode import and missing Qwen2_5_VLForConditionalGeneration export are distinct import errors." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "layer_type_validation import breakage is unrelated to local_files_only still allowing API requests." - }, - { - "left": "issue:45600", - "right": "issue:46298", - "accept": false, - "reason": "Broken auto_mappings references and a noisy logging message are not the same issue." - }, - { - "left": "issue:44038", - "right": "issue:44230", - "accept": false, - "reason": "Generic Qwen3-VL-Moe breakage and fp8 inference support are related in model family only, not the same concrete bug." + "reason": "ImportError for layer_type_validation and local_files_only still making API requests are unrelated." } ] }, @@ -15189,11 +15353,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:36010", "issue:36683", - "issue:42175", "issue:42757", "issue:42994", "issue:43010", @@ -15212,8 +15375,10 @@ "issue:43937", "issue:43976", "issue:43992", + "issue:44038", "issue:44075", "issue:44077", + "issue:44230", "issue:44242", "issue:44355", "issue:44380", @@ -15229,22 +15394,24 @@ "issue:45278", "issue:45290", "issue:45310", - "issue:45375", - "issue:45584", "issue:45593", + "issue:45600", "issue:45698", "issue:46077", "issue:46082", "issue:46172", - "issue:46202", "issue:46209", + "issue:46298", "issue:46365" ], "soft_pairs": [ + "issue:45600|issue:46298", + "issue:44038|issue:44230", "issue:45698|issue:46172", "issue:43618|issue:44380", "issue:43618|issue:43761", "issue:44483|issue:45290", + "issue:44625|issue:46365", "issue:43299|issue:43867", "issue:36683|issue:43475", "issue:42994|issue:44655", @@ -15254,253 +15421,236 @@ "issue:43901|issue:44655", "issue:36010|issue:42757", "issue:43611|issue:44655", - "issue:43352|issue:43937", "issue:36683|issue:44077", + "issue:43352|issue:43937", "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:43352|issue:45593", "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:44075|issue:46365", "issue:43010|issue:45278", - "issue:44355|issue:44855", - "issue:43867|issue:46202", - "issue:43867|issue:45375" + "issue:44355|issue:44855" ], - "prepared_review_unit_hash": "79ce4c8f80d2e8929e55232626f12177563a029b38a6b0a1861fe26298e59010", + "prepared_review_unit_hash": "4606d0334b26b9e327a3138ba8b6780f60505dd217df21bba20db6ad0fff43ce", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37164, - "estimated_input_tokens": 9291, - "estimated_eval_tokens": 18838 + "soft_pair_count": 31, + "serialized_chars": 36465, + "estimated_input_tokens": 9117, + "estimated_eval_tokens": 18490 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is a loose mix of Transformers v5 regressions and unrelated import/load/save bugs. None of the soft pairs are the same concrete defect, so they should stay separate rather than be collapsed into one duplicate set.", - "confidence": 0.92, - "canonical_issue_reason": "No true canonical issue: the items are heterogeneous. If forced to pick one umbrella, issue 45278 is the broadest import-regression bucket and is the most reusable triage anchor among the set.", + "summary": "This cluster is only loosely related to Transformers v5 regressions; most soft pairs point to different model-, API-, or serialization bugs rather than true duplicates. No soft edge looks safe to merge.", + "confidence": 0.23, + "canonical_issue_reason": "issue:45278 is the broadest umbrella in the set: it reports multiple import failures after upgrading Transformers, which best matches the cluster\u2019s overall regression-heavy theme.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the best representative only as a broad umbrella for import failures after upgrading; the rest of the cluster is too diverse to treat as one bug.", + "best_issue_reason": "issue:45278 is the best representative issue because it is broad enough to summarize the cluster, while the others are narrower model-specific or feature-specific bugs.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45600", + "right": "issue:46298", + "accept": false, + "reason": "Unrelated: removed Sam3LiteText auto-mapping CI break vs a noisy logging message." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Both mention Qwen3-VL/MoE, but one is a general loading bug and the other is fp8 inference support; different underlying problems." + }, { "left": "issue:45698", "right": "issue:46172", "accept": false, - "reason": "Save/load custom-module corruption and a deepseek-v4 tensor transpose bug are different code paths and symptoms." + "reason": "Different areas: save/load of custom modules vs a DeepSeek-V4 tensor transpose bug." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "Missing CLIP attentions and ignored GPT2 attention scaling under SDPA/FlashAttention are unrelated regressions." + "reason": "Both are regressions, but CLIP attentions missing is not the same bug as GPT-2 attention scaling being ignored in SDPA/FlashAttention." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Both involve CLIP/v5 output behavior, but one drops attentions while the other drops hidden_states; different concrete defects." + "reason": "Same model family, different symptoms and code paths: missing attentions vs missing hidden_states." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "An API request rejection and a chat-template crash on tool-call messages are different layers and bugs." + "reason": "Chat API request validation is a different problem from apply_chat_template crashing on tool-call assistant messages." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Qwen3.5 config propagation bug vs an import-order issue involving datasets/Trainer." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3VL MoE loading breakage and a state_dict sorting load error are not the same failure mode." + "reason": "Both are model-loading failures, but they involve different models and failure modes." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Missing Gemma3Config vocab_size and missing Sam3VisionEncoderOutput fpn_position_embeddings are unrelated attribute errors." + "reason": "AttributeError on Gemma3Config vs missing field on Sam3VisionEncoderOutput; unrelated models and attributes." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Quantized model saving failure and Pipeline save_pretrained failure are both save-related but not the same bug." + "reason": "Quantized model saving failure is not the same as saving Pipeline objects with save_pretrained." }, { "left": "issue:43066", "right": "issue:46209", "accept": false, - "reason": "Wrong tokenizer decoder type and FP8 inference failure on B200 are unrelated issues." + "reason": "Tokenizer decoder type bug vs FP8 inference failure on B200; unrelated." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Different RoPE-related model bugs in different models; not the same concrete implementation defect." + "reason": "Different RoPE-related model issues; same general area but not the same concrete bug." }, { "left": "issue:45276", "right": "issue:46077", "accept": false, - "reason": "Gemma4 embedding resize not propagating and Gemma4 attention-mask construction are distinct code paths." + "reason": "Both Gemma4-related, but resize_token_embeddings behavior differs from bidirectional attention mask construction." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Docs wording for return_all_scores and Pipeline serialization are unrelated." + "reason": "Docs mismatch for TextClassificationPipeline vs serialization of Pipeline objects." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they concern different symbols and different dependency breakages." + "reason": "Both are import errors, but for different symbols and different dependency packages." }, { "left": "issue:43611", "right": "issue:44655", "accept": false, - "reason": "Base_model_prefix loading regression and Pipeline save_pretrained failure are not the same bug." + "reason": "Model loading with base_model_prefix is not the same as saving Pipeline objects." }, { - "left": "issue:43352", - "right": "issue:43937", + "left": "issue:36683", + "right": "issue:44077", "accept": false, - "reason": "FlashAttention support limitation and invalid GenerationConfig are unrelated model/config failures." + "reason": "Gemma3Config missing vocab_size and patchtsmixer post_init validation are unrelated." }, { - "left": "issue:36683", - "right": "issue:44077", + "left": "issue:43352", + "right": "issue:43937", "accept": false, - "reason": "A missing config attribute and a deprecation/allowlist issue in patchtsmixer are different bugs." + "reason": "Flash Attention support error for NemotronH vs invalid GenerationConfig for GLM-5 are different bugs." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "State_dict ordering causing load failure is not the same as an incorrect checkpoint produced by save_pretrained." + "reason": "Both involve loading/saving checkpoints, but one is a state_dict ordering error and the other is a Qwen3.5 save_pretrained regression." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Gemma4 bidirectional-attention masking are unrelated." + "reason": "Qwen3.5 num_labels propagation is unrelated to Gemma4 bidirectional attention masks." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "These are both version-compatibility problems, but for different Python versions and different failures." - }, - { - "left": "issue:45584", - "right": "issue:46365", - "accept": false, - "reason": "Whisper empty-transcription generation and a Trainer import conflict with datasets are unrelated." + "reason": "Different Python-version compatibility failures with no clear shared code path." }, { - "left": "issue:42175", - "right": "issue:46365", + "left": "issue:46077", + "right": "issue:46082", "accept": false, - "reason": "Missing TensorFlow backend in an install extra is unrelated to the Trainer import conflict." + "reason": "Same version window, but Gemma4 attention masks and LlamaConfig head_dim validation are unrelated." }, { "left": "issue:43352", "right": "issue:45593", "accept": false, - "reason": "No FlashAttention 2 support and missing auxiliary losses are unrelated model bugs." - }, - { - "left": "issue:46077", - "right": "issue:46082", - "accept": false, - "reason": "Gemma4 causal-mask behavior and LlamaConfig head_dim validation are different configuration/architecture issues." + "reason": "FlashAttention support limitation vs D-FINE auxiliary loss logic; unrelated." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Two missing config attributes on different model families are similar in shape but not the same underlying bug." + "reason": "Different AttributeErrors on different model configs." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "Missing CLIP attentions and incorrect Qwen3.5 saved checkpoints are unrelated regressions." - }, - { - "left": "issue:44077", - "right": "issue:46365", - "accept": false, - "reason": "patchtsmixer post_init policy and the datasets-induced Trainer import error are unrelated." + "reason": "CLIP output regression and Qwen3.5 checkpoint saving regression are unrelated." }, { "left": "issue:36683", "right": "issue:44242", "accept": false, - "reason": "A missing Gemma3 config attribute and a missing load-balancing loss are different defects." - }, - { - "left": "issue:44075", - "right": "issue:46365", - "accept": false, - "reason": "SGD arguments being ignored and a Trainer import conflict are unrelated." + "reason": "Gemma3Config vocab_size missing is unrelated to MoE load-balancing loss omission." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "UMT5Encoder missing embed_tokens.weight and Qwen3.5 MoE from_pretrained failure are different loading bugs." + "reason": "Both concern Qwen3.5, but one misses embed_tokens.weight and the other has a from_pretrained error; not the same concrete bug." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "A distributed optimizer-state bug and a typo in cache strategies are not the same issue." + "reason": "Optimizer state corruption under FSDP2/PEFT is unrelated to a cache strategy typo." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "SGD args ignored is an optimizer bug; importing Trainer being blocked by datasets is a different import/dependency issue." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Cache update no_grad behavior and broad import errors after upgrade do not share the same code path." + "reason": "No_grad decoration for update methods is unrelated to import errors after upgrading Transformers." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Compiled-file execution errors and a Python 3.13 JIT parsing/indentation problem are distinct." - }, - { - "left": "issue:43867", - "right": "issue:46202", - "accept": false, - "reason": "State_dict-sorting load failure and Gemma pipeline roundtrip failure on \u2581 are different serialization bugs." - }, - { - "left": "issue:43867", - "right": "issue:45375", - "accept": false, - "reason": "Load failure from sorted state_dict and a silently dropped config field are unrelated." + "reason": "Compiled Python file handling is unrelated to the Python 3.13 DebertaV2Model parsing error." } ] }, @@ -15522,7 +15672,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:28282", "issue:30333", @@ -15555,7 +15705,6 @@ "issue:44387", "issue:44485", "issue:44556", - "issue:44625", "issue:44749", "issue:44857", "issue:44938", @@ -15570,16 +15719,17 @@ "issue:45600", "issue:45850", "issue:46132", - "issue:46285", - "issue:46365" + "issue:46202", + "issue:46285" ], "soft_pairs": [ + "issue:43867|issue:46202", + "issue:43867|issue:45375", "issue:41669|issue:44246", - "issue:36683|issue:43064", "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", @@ -15607,236 +15757,239 @@ "issue:43010|issue:43232", "issue:43502|issue:44222" ], - "prepared_review_unit_hash": "dc5256ed377d879792e3997c42dece091188d2f610b88d6750e03b21b525b25f", + "prepared_review_unit_hash": "fba921e4e0f07433ba7038ca95ce222eba0e81f99df9c1a2904b2d903102dea8", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36840, - "estimated_input_tokens": 9210, - "estimated_eval_tokens": 18676 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36540, + "estimated_input_tokens": 9135, + "estimated_eval_tokens": 18526 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous mix of unrelated Transformers issues. The only plausible duplicate thread is import-time slowdown/import-surface regression (41669 and 44246); the rest do not look like the same underlying bug.", - "confidence": 0.68, - "canonical_issue_reason": "44246 is the best available anchor because it captures the broad user-facing symptom in the only plausible shared theme (slow imports), though the cluster is not truly coherent.", + "summary": "These issues are largely a grab-bag of unrelated regressions and bugs. None of the proposed soft pairs looks like the same underlying problem closely enough to merge as duplicates, so I reject all soft edges. If a single issue must stand in for the set, the broad v5 reload regression report is the best representative, but only weakly so.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 44556 is the broadest, most user-facing regression in the set: a version-upgrade load/reload failure spanning v4.57 to v5.2/v5.3. It is still not a true duplicate of the others, but it is the least niche representative.", "canonical_pr_reason": null, - "best_issue_reason": "44246 is the most representative issue overall: it is broad, user-facing, and closest to the shared import-performance theme; the other issues are mostly unrelated point bugs.", + "best_issue_reason": "Issue 44556 best captures the general class of v5 loading regressions and is the most plausible umbrella representative among these unrelated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41669", - "right": "issue:44246", - "accept": true, - "reason": "Same import-performance regression theme: 41669 gives a likely root cause (`import *` in models), while 44246 reports the user-visible slowdown." + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Both mention loading-related failures, but one is a generic state_dict loading error while the other is a Gemma pipeline roundtrip/tokenization issue. Different code paths and root causes." }, { - "left": "issue:36683", - "right": "issue:43064", + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "State_dict load sorting error and a missing Qwen3.5 vision config field are different bugs in different components." + }, + { + "left": "issue:41669", + "right": "issue:44246", "accept": false, - "reason": "Different failures: missing `vocab_size` in Gemma3 config vs distributed optimizer-state corruption in FSDP2/PEFT." + "reason": "Import-time slowdown from import * cleanup and intermittent import slowness are both performance-related, but not the same change or failure mode." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Both are version-regression reports, but one is Python 3.14 import failure and the other is `PretrainedConfig` type-checking breakage; different bugs." + "reason": "Python 3.14 import/load failure is an environment compatibility issue; the other is PretrainedConfig type checking in v5.4.0. Not the same bug." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config missing vocab_size and FSDP2/PEFT optimizer state corruption are unrelated model/config vs training-distribution bugs." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Both are flaky test issues, but one is missing `set_seed()` in SamHQ tests and the other is an xdist race on `captured_info.txt`." + "reason": "A flaky SamHQ integration test due to missing set_seed differs from a pytest-xdist race in captured_info.txt." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "`@torch.no_grad` on cache/layer updates is a specific API bug; 45850 is a broad complaint about many models breaking in v5." - }, - { - "left": "issue:44625", - "right": "issue:46365", - "accept": false, - "reason": "Unrelated: Qwen3.5 `num_labels` propagation vs inability to import `Trainer` until uninstalling `datasets`." + "reason": "A cache/layer @torch.no_grad issue is a specific implementation bug; the other is a broad complaint that v5 breaks too many models." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Both involve `from_pretrained`, but the concrete problems differ: missing `embed_tokens.weight` vs Qwen3.5 MoE load error." + "reason": "UMT5Encoder missing embed_tokens.weight and Qwen3.5 Moe from_pretrained errors are different model-loading failures." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "Different save/load paths and model families: sorted state_dict loading error vs incorrect visual-encoder keys on save." + "reason": "Generic load error with sorted state_dict and incorrect visual encoder save_pretrained keys are different save/load bugs." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "No relation beyond being v5-related regressions; one is cache/layer update semantics, the other is a Python-version compatibility failure." + "reason": "Cache update decoration and Python version compatibility are unrelated." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bug vs OOM while loading a specific Qwen2 model with TP; related area, but not the same concrete bug." + "reason": "Tensor parallel training bug and OOM when loading Qwen2-57B-A14B-Instruct with tp are both TP-related, but the reported failures and likely fixes differ." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "Unrelated model/API issues: CLIP output attentions assignment vs patchtsmixer `post_init` allowance." + "reason": "CLIPOutput attentions missing assignment and patchtsmixer optional post_init restriction are separate model API issues." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "Different subsystems and symptoms: missing PyTorch import for AutoModel vs tokenizer-loading regression." + "reason": "PyTorch-not-found ImportError and tokenizer loading regression are both import/load themed, but not the same underlying defect." }, { "left": "issue:43618", "right": "issue:46285", "accept": false, - "reason": "Both touch CLIP, but one is missing attentions in output and the other is an inconsistent internal structure change; not the same bug." + "reason": "Missing CLIPOutput attentions and inconsistent CLIP text model structure are distinct CLIP API regressions." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "CLIP attentions regression and LwDetr float16 AMP crash are unrelated." + "reason": "Missing attentions assignment and LwDetrImageLoss AMP CUDA crash are unrelated model-output vs training/numerics bugs." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "No shared concrete code-path: DSA indexer activation issue vs DeepseekV3MoE divergence from remote implementation." + "reason": "A DSA indexer ReLU discussion and DeepseekV3MoE divergence are different implementation concerns." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "Different bug classes: local_files_only still making network requests vs AddedToken constructor argument collision." + "reason": "local_files_only still making API requests is a download/offline bug; AddedToken special kwarg duplication is a tokenizer construction bug." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Test flakiness in SamHQ vs failure loading a Phi multimodal model; not the same underlying issue." + "reason": "Seed-related test flakiness and Phi-4 multimodal model loading failure are unrelated." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Both are CI/test-related, but one is missing seeding and the other is a stale device override in GraniteSpeech." + "reason": "A missing set_seed in a test and a stale GraniteSpeech device override test failure are different CI/test problems." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "Unrelated test flake vs auto_mappings.py referencing removed Sam3LiteText configs." + "reason": "Seed flakiness and auto_mappings referencing removed Sam3LiteText configs are unrelated test vs mapping issues." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both are loading regressions, but one is tokenizer loading in v5 and the other is checkpoint reload incompatibility after upgrading from v4.57." + "reason": "Both are loading regressions after v5, but one is tokenizer loading and the other is checkpoint reload across versions; different concrete failures." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "OOM under int4 quantization vs reload failure after version upgrade; different mechanisms and fixes." + "reason": "Int4 quantization OOM from higher reserved CUDA memory is a memory-management regression, not a checkpoint reload failure." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "Different Qwen3.5 MoE issues: FP8 save_pretrained bug vs missing tensor-parallel plan." + "reason": "FP8 save_pretrained for moe and missing _tp_plan for tensor parallelism are different model export vs parallelism metadata bugs." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "No relation: cache/layer no_grad decoration vs dummy Conv2d in Sam3PixelDecoder." + "reason": "Torch.no_grad decoration for cache/layer updates and a dummy nn.Conv2d in Sam3PixelDecoder are unrelated implementation changes." }, { "left": "issue:43576", "right": "issue:46132", "accept": false, - "reason": "`transformers env` command breakage is unrelated to AttentionInterface.register behavior changes." + "reason": "Broken transformers env command and AttentionInterface.register behavior change are separate CLI vs attention API issues." }, { "left": "issue:43502", "right": "issue:44485", "accept": false, - "reason": "Local-files-only network bug vs GLM-5 RoPE implementation discussion; different problems." + "reason": "Offline download leakage and GLM-5 RoPE implementation concern different subsystems and failure modes." }, { "left": "issue:42548", "right": "issue:44351", "accept": false, - "reason": "Both are import-name errors from `transformers`, but they concern different missing symbols and do not look like the same concrete defect." + "reason": "cannot import PreTrainedModel and cannot import HybridCache are both import errors, but they affect different symbols and likely different export paths." }, { "left": "issue:45125", "right": "issue:45478", "accept": false, - "reason": "Both concern Qwen3.5 MoE, but one is missing `_tp_plan` and the other is a `from_pretrained` error in newer transformers." + "reason": "Missing tensor-parallel plan for Qwen3_5MoeForConditionalGeneration and Qwen3.5 Moe from_pretrained error are related only by model family, not the same bug." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled-Python-file errors vs performance slowdown after upgrading from 4.57.3 to 5.3.0 are unrelated." + "reason": "Compiled Python file loading errors and slow filtering after upgrade are unrelated." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLFlow reporting on failed training jobs has nothing to do with tokenizer loading regression." + "reason": "MLFlow not reporting failed training jobs and tokenizer loading regression are entirely different problems." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "Both are Qwen3.5-related, but one is tensor parallelism setup and the other is a missing vision-config field silently dropped by strict validation." + "reason": "Missing _tp_plan and missing deepstack_visual_indexes are different Qwen3.5 configuration/model export issues." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "No shared bug: cache/layer update decoration vs generation state sync after `sync_gpus`." + "reason": "Cache update no_grad decoration and generation kwargs sync handling are different generation internals." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "Different issues: unwanted API requests during offline loading vs FP8 MoE save_pretrained behavior." + "reason": "local_files_only offline leakage and FP8 save_pretrained moE are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. It correctly treats 44246 as the best anchor for the only clearly related thread (import-time slowdown, with 41669), and the pair verdicts avoid overstating weaker similarities. No mergeability issue since there are no PR pairs." - }, - "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 }, @@ -15851,7 +16004,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:27957", "issue:30333", @@ -15919,12 +16072,12 @@ "issue:43576|issue:43824", "issue:31795|issue:38617", "issue:44961|issue:45468", - "issue:44222|issue:44230", "issue:43576|issue:44938", + "issue:44222|issue:44230", "issue:41628|issue:42548", + "issue:46143|issue:46144", "issue:44908|issue:44961", "issue:44485|issue:44961", - "issue:46143|issue:46144", "issue:36296|issue:43901", "issue:44315|issue:46209", "issue:36296|issue:43526", @@ -15936,233 +16089,236 @@ "issue:44230|issue:46095", "issue:44393|issue:45071" ], - "prepared_review_unit_hash": "ebd052d6e9a63d460f3df0529544a4047502c3fcc16261c22f9d2c3e942e584e", + "prepared_review_unit_hash": "3af8d5b294438fee42a60e4f8195947bf5d8b2795d3a6e7ea2583c24e1c1320b", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36415, - "estimated_input_tokens": 9104, - "estimated_eval_tokens": 18464 + "serialized_chars": 36411, + "estimated_input_tokens": 9103, + "estimated_eval_tokens": 18462 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is heterogeneous: a few pairs share broad themes like v5 regressions, loading failures, or import errors, but they describe different concrete bugs and should not be deduplicated. No pull requests are present.", - "confidence": 0.86, - "canonical_issue_reason": "No single issue is a strong canonical duplicate target here; the items span unrelated problems across loading, docs, FP8, import/export, CLI, and tensor-parallel behavior.", + "summary": "The cluster is mostly a grab bag of Transformers v5-era regressions, import failures, model-loading issues, and unrelated docs/feature bugs. Only the RoFormer pair looks like the same underlying code-path problem; the rest are too different to merge as duplicates.", + "confidence": 0.41, + "canonical_issue_reason": "issue:45850 is the broadest umbrella in the set, capturing the shared theme of widespread v5 regressions across models and loading paths, even though it is very generic.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45850 is the broadest umbrella-style report, but it is too vague to be a reliable dedup target.", + "best_issue_reason": "issue:45850 is the best single representative because it summarizes the overarching cluster theme better than the more specific bugs, but it is still too vague to serve as a precise canonical fix.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "One is a Sentence Transformers load hang; the other is a docs wording complaint for model.forward." + "reason": "Loading-stuck bug vs model.forward documentation issue; different problem and code path." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both involve version-related loading regressions, but for different models and likely different failure modes." + "reason": "Both are reload regressions, but one is Qwen3VL-Moe-specific and the other is a generic v4.57\u2192v5.2/5.3 checkpoint reload failure." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "local_files_only API leakage is unrelated to the unrelated 'racoon' report." + "reason": "Local-files-only network leak vs unrelated placeholder issue; no shared bug." }, { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "FP8 accuracy regression and a missing image-processor method are different bugs." + "reason": "FP8 accuracy regression vs missing image-processor method causing multimodal token counting failure; different issues." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLFlow reporting on failed jobs is unrelated to tensor-parallel training." + "reason": "MLflow reporting failure after job crash vs tensor-parallel training bug; unrelated." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled-file loading errors are not the same as PretrainedConfig type-checking breakage." + "reason": "Compiled-file execution error vs PretrainedConfig type-checking breakage; distinct regressions." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Documentation references to removed pipeline() APIs are unrelated to 'racoon'." + "reason": "Docs still referencing pipeline() vs unrelated placeholder issue." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Compiled Python file errors and save_pretrained pipeline issues affect different code paths." + "reason": "Compiled Python file errors and save_pretrained pipeline serialization are separate bugs." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel not applied with model_init is unrelated to the DSA/ReLU discussion." + "reason": "Liger Kernel application bug vs DSA indexer/ReLU discussion; unrelated." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Tokenizer loading regression and Llama 3.2 vision model loading failure are different concrete load bugs." + "reason": "Tokenizer-loading regression vs Llama 3.2 90B Vision model-loading regression; similar symptom, different failure." }, { "left": "issue:27957", "right": "issue:33453", "accept": false, - "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to tokenizer loading regression." + "reason": "Flash Attention 2 support for XLMRoberta vs tokenizer-loading regression; different code paths." }, { "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "Missing huggingface_hub import and the v5 env CLI breakage are not the same issue." + "reason": "Missing huggingface_hub import vs broken transformers env command; both are import/CLI symptoms but not the same bug." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Broken env command and PretrainedConfig type checking are separate regressions." + "reason": "Broken env command and PretrainedConfig type-checking are separate regressions." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "CLI env failure is unrelated to the missing Qwen2_5_VL export." + "reason": "env CLI breakage vs missing Qwen2_5_VL export; different export surfaces." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "A docs confusion issue does not match an ImportError for layer_type_validation." + "reason": "Documentation confusion vs ImportError for layer_type_validation; unrelated." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "The placeholder 'racoon' report does not match the Gemma-4 audio positional encoding bug." + "reason": "Placeholder issue vs Gemma-4 audio positional encoding bug; no overlap." }, { - "left": "issue:44222", - "right": "issue:44230", + "left": "issue:43576", + "right": "issue:44938", "accept": false, - "reason": "FP8 save_pretrained and FP8 inference support are different failures, not one code-path bug." + "reason": "CLI env issue vs Python 3.14 load failure; too broad to be the same bug." }, { - "left": "issue:43576", - "right": "issue:44938", + "left": "issue:44222", + "right": "issue:44230", "accept": false, - "reason": "Broken transformers env CLI and Python 3.14 import/load failure are distinct problems." + "reason": "FP8 save_pretrained bug vs FP8 inference support request; related theme, but not the same concrete defect." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Different missing exports: AutoImageProcessor versus PreTrainedModel." + "reason": "Different missing top-level imports; likely separate export regressions." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": true, + "reason": "Both point to the same RoFormer forwarding/interface problem: kwargs are not propagated because the attention implementation bypasses the expected interface." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs handling is unrelated to the 'racoon' issue." + "reason": "Scheduler kwargs bug vs unrelated placeholder issue." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to the 'racoon' report." - }, - { - "left": "issue:46143", - "right": "issue:46144", - "accept": false, - "reason": "Both concern RoFormer, but one is about kwargs propagation and the other about attention-interface usage; not the same concrete bug." + "reason": "GLM-5 RoPE implementation discussion vs unrelated placeholder issue." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor-parallel training and pipeline docs behavior are unrelated." + "reason": "Tensor-parallel training bug vs pipeline docs mismatch; unrelated." }, { "left": "issue:44315", "right": "issue:46209", "accept": false, - "reason": "Liger Kernel application in model_init is unrelated to FP8 inference on B200." + "reason": "Model-init Liger Kernel application bug vs FP8 inference failure on B200; different failures." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor-parallel training bug and BeitImageProcessorFast label reduction are different code paths." + "reason": "Tensor-parallel training bug vs BeitImageProcessorFast reduce_labels bug; unrelated." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Generation cache missing during chunked_prefill is not a typo in cache strategy docs." + "reason": "Missing GenerationMixin cache in chunked_prefill vs documentation typo; not the same issue." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "A specific Qwen3.5-35B chat crash is not the same bug as a broad 'many models break in v5' complaint." + "reason": "Specific Qwen3.5-35B chat crash vs broad complaint about many v5 breakages; not a duplicate." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Missing files in processor.save_pretrained is unrelated to the placeholder bug report." + "reason": "Processor save_pretrained missing files vs generic bug report placeholder; no evidence of sameness." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading failures and the TensorFlow/tensorboard backend error are different issues." + "reason": "Lazy loading failure vs undefined tf backend error when only tensorboard is installed; unrelated." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "Qwen3-VL MoE breakage and missing _tp_plan for tensor parallelism are separate model-support gaps." + "reason": "Qwen3-VL-Moe v5 bug vs missing _tp_plan for Qwen3.5Moe tensor parallelism; same family, but not the same concrete code-path problem." }, { "left": "issue:44230", "right": "issue:46095", "accept": false, - "reason": "FP8 inference support and a weight-conversion-script question are not the same underlying bug." + "reason": "FP8 inference support request vs DeepSeekV4 weight-conversion question; unrelated." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Qwen3-VL 2D bounding-box output problems are unrelated to PretrainedConfig type-checking regressions." + "reason": "Qwen3-VL bounding-box output problem vs PretrainedConfig type-checking breakage; different bugs." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet: issue:45850 is a reasonable broad canonical issue, and the cluster is accurately described as mostly heterogeneous v5 regressions rather than a tight duplicate set. The soft-edge calls are generally conservative, and the only accepted pair (46143/46144) is plausibly the same RoFormer interface/kwargs defect from the titles alone." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -16177,7 +16333,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:31795", "issue:33453", @@ -16244,8 +16400,8 @@ "issue:43408|issue:44315", "issue:44297|issue:45230", "issue:44246|issue:44273", - "issue:43352|issue:44393", "issue:44295|issue:45230", + "issue:43352|issue:44393", "issue:43576|issue:44908", "issue:44704|issue:44908", "issue:43976|issue:44393", @@ -16272,20 +16428,20 @@ "issue:41669|issue:43704", "issue:31795|issue:34689" ], - "prepared_review_unit_hash": "42adaef6177f8e02448c17a48b8f60049262999e74cb4d1b61b20971c3fa16ca", + "prepared_review_unit_hash": "11951b5bfa07b38257532f90d8dd76794deb3d82c8978347926045f27e831bc6", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 42, - "serialized_chars": 39548, + "serialized_chars": 39545, "estimated_input_tokens": 9887, "estimated_eval_tokens": 20030 }, @@ -16294,264 +16450,264 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The soft pairs are mostly false positives: they share broad themes like v5 regressions, import/loading problems, or model-family-specific bugs, but not the same concrete underlying defect. The closest umbrella issue is the general v5 import/loading breakage report, but there is no strong duplicate cluster here.", - "confidence": 0.89, - "canonical_issue_reason": "issue:45278 is the broadest concrete regression report in the set (many import errors after upgrading 4.57.0 -> 5.5.0), so it is the closest umbrella candidate for the import/load-related subset, though it is not a true duplicate of most others.", + "summary": "Cluster is mostly a grab bag of Transformers v5 regressions, import/loading issues, and model-specific bugs; there are a few thematic similarities, but no clear duplicate pair.", + "confidence": 0.66, + "canonical_issue_reason": "issue:45850 is the broadest umbrella for the cluster\u2019s recurring \u201cv5 breaks models\u201d theme, even though it is very generic.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most representative and actionable issue for this cluster because it clearly describes a widespread upgrade regression affecting imports/loading, while the other issues are mostly narrower model-specific bugs or unrelated regressions.", + "best_issue_reason": "issue:45850 is the best representative of the cluster because it captures the common version-regression theme shared by several nearby reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Both involve import/lazy-loading behavior, but one is a generic lazy-loading complaint and the other is a specific sys.modules access concern; not clearly the same bug." + "reason": "Related import/lazy-loading area, but different bugs: intermittent import slowness vs unsafe `sys.modules` access." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Both are v5 breakage reports, but one is about the `transformers env` command and the other about checkpoint reload compatibility; different code paths." + "reason": "Different failures: broken `transformers env` command vs checkpoint reload compatibility after upgrade." }, { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Same model family, but one is a 2D bounding-box output issue and the other is an infinite-loop on a specific image; different failures." + "reason": "Same model family, but different symptoms and code paths: bbox output bug vs image-specific infinite loop." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "One is a specific `Llama4Config` attribute error; the other is a vague general complaint about many models breaking in 5.0." + "reason": "Specific `Llama4Config` attribute error vs a broad complaint about many models breaking." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "One concerns expensive `import *` usage and import-time overhead; the other is a GenerationMixin cache problem during chunked prefill." + "reason": "Import-star cleanup/perf issue is unrelated to missing cache in chunked prefill." }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both mention v5 reload/incompatibility after upgrades, but they concern different model families and likely different underlying breakages." + "reason": "Both are v5 loading regressions, but for different models and different root causes." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "Specific missing `pad_token_idx` in StableLmConfig is not the same as a broad claim that 5.0 breaks many models." + "reason": "Model-specific missing `pad_token_idx` issue vs a vague global breakage report." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A cache-strategy typo is unrelated to the generic v5 model-breakage complaint." + "reason": "Cache-strategy typo is unrelated to the broad v5 regression complaint." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "Both are broad v5 regression reports, but one is specifically about import errors after upgrade and the other is an unspecific blanket complaint; not the same defect." + "reason": "Both describe v5 breakages, but not the same concrete failure or fix." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API guidance and a cache-strategy typo are unrelated." + "reason": "TF32 API migration issue has nothing to do with a cache-strategy typo." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "One is about import-time cost from `import *`; the other is a missing `Qwen3_5ForCausalLM` import, which is a separate symbol/export issue." + "reason": "Both touch imports, but one is import-time bloat and the other is a missing symbol." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "A flaky test due to missing `set_seed()` is unrelated to scheduler kwargs handling." + "reason": "Missing `set_seed()` in tests is unrelated to scheduler kwargs handling." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Model-type warning for sam3_tracker/sam3_video and Liger Kernel not applied with `model_init` are unrelated." + "reason": "Model-type mismatch warning vs Liger Kernel not being applied under `model_init`." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer save metadata mismatch and a generic bug report do not describe the same bug." + "reason": "Tokenizer save metadata mismatch vs a too-generic bug report with no concrete match." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Slow imports and broken lazy loading may be related at a high level, but the reports do not identify the same concrete defect." + "reason": "Same import/lazy-loading area, but one is intermittent slowness and the other is a broader lazy-loading failure." }, { - "left": "issue:43352", - "right": "issue:44393", + "left": "issue:44295", + "right": "issue:45230", "accept": false, - "reason": "Flash Attention 2 support for Nemotron and Qwen3-VL bounding-box errors are unrelated." + "reason": "Buffer access error vs generic bug report; no specific overlap." }, { - "left": "issue:44295", - "right": "issue:45230", + "left": "issue:43352", + "right": "issue:44393", "accept": false, - "reason": "A buffer/`position_ids` access error is not the same as a generic bug report." + "reason": "Different model bugs and symptoms: Flash Attention 2 support error vs Qwen3-VL bbox issue." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "The `transformers env` command and inverse_sqrt scheduler kwargs are different areas of the codebase." + "reason": "Broken env CLI and scheduler kwargs are unrelated." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor kwarg forwarding and scheduler kwargs are unrelated parameter-passing bugs." + "reason": "AutoProcessor kwargs forwarding bug vs scheduler kwargs bug." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python-version compatibility and Qwen3-VL bounding-box output are unrelated issues." + "reason": "Python-version compatibility regression vs Qwen3-VL output bug." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwargs handling and `PretrainedConfig` type-checking are different bugs." + "reason": "Scheduler kwargs bug is unrelated to `PretrainedConfig` type-checking regression." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import performance from `import *` is unrelated to failing to import `HybridCache`." + "reason": "Import performance cleanup vs missing `HybridCache` symbol; different underlying problems." }, { "left": "issue:44492", "right": "issue:46328", "accept": false, - "reason": "A typo in cache strategies is unrelated to `SequenceFeatureExtractor.pad` converting arrays inefficiently." + "reason": "Cache-strategy typo vs feature-extractor padding inefficiency." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-star cleanup/performance and a small testing_utils bug are unrelated." + "reason": "Import cleanup/perf issue vs a separate testing_utils bug." }, { "left": "issue:43976", "right": "issue:45850", "accept": false, - "reason": "Python version support and generic v5 model breakage are not the same underlying bug." + "reason": "Python support regression vs a broad claim that v5 breaks many models." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Both concern Qwen-related models, but one is a VL bounding-box issue and the other is a Qwen3.5 MoE `from_pretrained` failure." + "reason": "Different Qwen3.5-related problems: vision bbox hallucination vs from_pretrained error." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "A `patchtsmixer` initialization policy issue is unrelated to `_get_tied_weight_keys` crashing on a list." + "reason": "PatchTSMixer post-init policy issue is unrelated to `_tied_weights_keys` list handling." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python-version compatibility and a missing `Qwen3_5ForCausalLM` import are unrelated." + "reason": "Python compatibility issue vs missing Qwen3_5 import; unrelated." }, { "left": "issue:43827", "right": "issue:44908", "accept": false, - "reason": "Docs referencing removed pipelines and scheduler kwargs are unrelated." + "reason": "Docs still mention pipelines, but the other issue is about scheduler kwargs." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion in `model.forward` is unrelated to tokenizer loading regression." + "reason": "Documentation confusion in `forward` vs tokenizer loading regression." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel not applied with `model_init` and degenerate training under flash_attention_3 are different training/model-init problems." + "reason": "Liger Kernel application issue vs flash_attention_3 training degeneration." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "`transformers env` being broken and lazy loading not working are both broad runtime complaints, but not the same specific bug." + "reason": "Env command failure vs lazy-loading issue; same broad area, but not the same bug." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Saving Pipeline objects and `_get_tied_weight_keys` crashing are unrelated failures." + "reason": "Pipeline save_pretrained issue vs tied-weights crash; no concrete match." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "A generic bug report and a specific Qwen3.5 MoE `from_pretrained` error are not duplicates." + "reason": "`45230` is too generic to establish the same bug as the Qwen3.5 MoE from_pretrained error." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "A DSA indexer activation issue is unrelated to Gemma4AudioRelPositionalEncoding." + "reason": "DSA/ReLU discussion vs Gemma-4 audio positional-encoding bug; unrelated." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading problems and inverse_sqrt scheduler kwargs are unrelated." + "reason": "Lazy-loading/import issue vs scheduler kwargs handling; different code paths." }, { "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "Broken env command and missing Qwen3_5 import are different issues." + "reason": "Broken env CLI vs missing model import; unrelated." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are different model-specific bugs." + "reason": "Qwen3VL timestamp calculation bug vs GLM-5 RoPE implementation issue." }, { "left": "issue:46018", "right": "issue:46172", "accept": false, - "reason": "DeepSeek-V4 shared expert gating and a DeepSeek-V4 transpose/dtype bug are different concrete defects." + "reason": "DeepSeek-V4 shared-expert gating bug vs a separate tensor transpose bug." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Timestamp calculation in a processor and scheduler kwargs handling are unrelated." + "reason": "Timestamp calculation bug vs scheduler kwargs handling; unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-star cleanup/performance and a VRAM leak in multi-threaded dataloaders are unrelated." + "reason": "Import-star cleanup/perf issue vs VRAM leak in multi-threaded dataloading." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "A documentation issue and a model-loading regression for Llama 3.2 Vision are not the same bug." + "reason": "Confusing docs issue vs Llama 3.2 Vision model-loading regression." } ] }, @@ -16573,7 +16729,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-10000-450", + "cluster_id": "cluster-10000-451", "nodes": [ "issue:30333", "issue:31515", @@ -16664,286 +16820,286 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "41043ad1267420e234cb6cfd9c1b4bf3355c2f12583e8f29ffad08d8e038b73b", + "prepared_review_unit_hash": "569e822d64c6453389c7d4c69dceb57fb57c246e504fe0c434082cd440e02159", "component_budget": { - "node_count": 450, - "item_count": 450, - "soft_pair_count": 1664, - "serialized_chars": 861331, - "estimated_input_tokens": 215333, - "estimated_eval_tokens": 430922 + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 }, "budget": { "node_count": 44, "item_count": 44, "soft_pair_count": 42, - "serialized_chars": 37124, - "estimated_input_tokens": 9281, - "estimated_eval_tokens": 18818 + "serialized_chars": 37114, + "estimated_input_tokens": 9279, + "estimated_eval_tokens": 18814 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: most items are model-specific regressions, tokenizer/config issues, or unrelated API/docs bugs. The soft-similarity pairs mostly share only a broad subsystem, not the same concrete defect, so I reject all soft edges.", - "confidence": 0.69, - "canonical_issue_reason": "Issue 45850 is the broadest umbrella in the set and best matches the apparent recurring theme of Transformers v5 regressions affecting multiple models, but the cluster is not truly duplicate-heavy.", + "summary": "The cluster is highly heterogeneous: most items are unrelated regression reports, doc issues, or model-specific bugs, so none of the soft edges look like true duplicates. No PRs are present; the closest thing to a thematic umbrella is a vague v5-regression complaint, but it is not a safe duplicate hub.", + "confidence": 0.96, + "canonical_issue_reason": "No strong canonical issue exists because the items span unrelated subsystems and failure modes. If one must be chosen, issue 45850 is the broadest umbrella for the v5-regression theme, but it is too vague to be a true duplicate target.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45850 is the best overall representative because it captures the widest common theme across the set (version-5 breakages), even though it is too broad to be a precise duplicate target.", + "best_issue_reason": "Issue 45850 is the only broad, umbrella-style report in the set, so it is the closest fit for a cluster representative. That said, it is still too generic to serve as a high-quality canonical duplicate.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Different problems: checkpoint loading speed versus documentation wording for forward inputs." + "reason": "Checkpoint loading slowdown and documentation confusion are different problem types." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "Different model/code paths: VRAM leak in dataloader threads versus GLM-5 RoPE implementation." + "reason": "VRAM leak in Qwen3 causal LM is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Both are model-specific bugs, but they affect different models and distinct logic." + "reason": "GLM-5 RoPE math and Gemma-4 audio positional encoding are different model-specific bugs." }, { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Continuous batching audio support and DeepSeek-V4 weight conversion are unrelated." + "reason": "Continuous batching for audio models is unrelated to DeepSeek-V4 weight conversion support." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "One is a broad version-5 complaint; the other is an import error for a specific class." + "reason": "A broad v5 regression complaint is not the same as a specific import error for Qwen3.5." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "Local-files-only network requests and Qwen3VL timestamp math are different bugs." + "reason": "Local-files-only network calls and Qwen3VL timestamp calculation are different code paths." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 settings guidance and a VRAM leak are unrelated." + "reason": "TF32 configuration guidance is unrelated to a VRAM leak bug." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Timestamp calculation and missing save_pretrained files are different code paths." + "reason": "Timestamp calculation and missing processor save files are separate processor issues." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "No concrete overlap beyond both being runtime bugs." + "reason": "Qwen3VL timestamp logic and Qwen3 causal LM VRAM leakage are not the same bug." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch and Gemma-4 positional encoding are unrelated." + "reason": "CPU offload device mismatch and Gemma-4 audio positional encoding do not share a concrete failure mode." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return handling in GLM-MoE is not the same as Gemma-4 audio positional encoding." + "reason": "torch.split return handling in GlmMoeDsaIndexer is unrelated to Gemma-4 audio positional encoding." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "Different model internals and different failure modes." + "reason": "Missing rms_norm_eps in MLA q_a_layernorm is a different bug from Gemma-4 audio encoding." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "Both are import errors, but for different symbols and likely different fixes." + "reason": "HybridCache import failure and Qwen3.5 module import failure are separate missing-symbol issues." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash Attention support gating and a module import failure are unrelated." + "reason": "Flash Attention support in NemotronH is unrelated to Qwen3.5 importability." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both concern tokenizer v5 behavior, but one is a broad behavior change and the other is a decoder-type mapping bug." + "reason": "Both mention tokenizer v5 behavior, but one is a broad behavior change and the other is a specific decoder-type bug." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs bug versus an unrelated placeholder issue." + "reason": "Scheduler kwargs handling is unrelated to the unrelated 'Unemployment' issue." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "A typo in cache strategies is unrelated to the other issue." + "reason": "A cache-strategy typo and an unrelated issue titled 'Unemployment' do not match." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "One is not a technical bug; the other is a broad Transformers regression report." + "reason": "The broad v5 regression complaint does not describe the same bug as the unrelated 'Unemployment' issue." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression and an unrelated placeholder issue." + "reason": "CLIPOutput attentions regression is unrelated to the 'Unemployment' issue." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Tied-weights crash and chat-template tool-call crash are different code paths." + "reason": "Tied-weights key handling and chat template tool-call crashes are different failures." }, { "left": "issue:44998", "right": "issue:45341", "accept": false, - "reason": "Unrelated issues; no shared defect." + "reason": "The 'Unemployment' issue and a testing_utils bug are unrelated." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing config attribute in Llama4 versus tied-weights list handling are unrelated." + "reason": "Missing pad_token_id on Llama4Config is unrelated to tied-weights key crashes." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "StableLm config padding field regression is not the same bug as tied-weights handling." + "reason": "Missing pad_token_idx in StableLmConfig is a config issue, not tied-weight key handling." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow checkpoint loading and an ImportError for layer_type_validation are unrelated." + "reason": "Slow checkpoint loading and a missing import symbol are unrelated." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 support request and processor.save_pretrained missing files are different issues." + "reason": "FP8 inference support and processor save_pretrained file output are different topics." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLFlow reporting on failed training jobs is unrelated to checkpoint loading speed." + "reason": "MLflow job-status reporting and checkpoint loading speed are unrelated." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Different symptoms and different subsystems." + "reason": "Checkpoint loading performance and tensor-parallel training bugs are different problems." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 feature support request versus from_pretrained error in Qwen3.5 MoE are not the same defect." + "reason": "FP8 support request and Qwen3.5 from_pretrained failure are not the same underlying issue." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "Checkpoint loading performance and tokenizer loading regression are different problems." + "reason": "Checkpoint loading slowdown and tokenizer-loading regression are separate issues." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Audio continuous batching support and FP8 inference support are unrelated." + "reason": "Continuous batching for audio models is unrelated to FP8 inference support." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba init_weights and RoFormer attention interface are separate model implementations." + "reason": "Mamba init_weights behavior and RoFormer attention-interface usage are different model implementation bugs." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Different model-specific bugs with no shared concrete code path." + "reason": "Mask2Former ignore_value handling and DeepSeek-V4 shared-expert gating are unrelated." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "Incorrect type hint and tokenizer regex warning are unrelated." + "reason": "ModernBERT norm_eps typing and tokenizer regex warnings are unrelated." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "One is a missing use of ignore_value; the other is kwargs not passed into loss calculation." + "reason": "Mask2Former ignore_value and loss-function kwargs passthrough are different code paths." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss kwargs handling and RoFormer attention interface are unrelated." + "reason": "Loss kwargs passthrough and RoFormer attention-interface behavior are unrelated." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Different bugs: ignore_value usage versus regex warning." + "reason": "A model-specific ignore_value bug does not match a global tokenizer regex warning." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "Different model internals and failure modes." + "reason": "Mask2Former loss handling and RoFormer attention implementation are different bugs." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "A model-loss bug and a typo in quantization operation naming are unrelated." + "reason": "Ignore_value usage and a typo in quantization operation naming are unrelated." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs propagation and a config type-hint issue are unrelated." + "reason": "Loss kwargs passthrough and config type hints are unrelated." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "Type-hint correction and typo in cache conversion naming are not the same bug." + "reason": "ModernBERT norm_eps typing and a quantization-operation typo are different issues." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "RMSNorm question for Llama4 and a ModernBERT type-hint bug are unrelated." + "reason": "A Llama4 norm question and a ModernBERT type-hint bug are unrelated." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Different model-specific issues with no concrete overlap." + "reason": "Mask2Former ignore_value handling and ModernBERT norm_eps typing are different bugs." } ] }, @@ -16976,52 +17132,52 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "e3740ca505d98751732cb5dfb22c3cef79ead20a31a24fb521d5b38b2628895a", + "prepared_review_unit_hash": "b2abcc9379be3cc77abe5fd2ea0dfaebe4795400f1064d4a1f527edc69169032", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2878, - "estimated_input_tokens": 720, - "estimated_eval_tokens": 1696 + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 2878, - "estimated_input_tokens": 720, - "estimated_eval_tokens": 1696 + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three issues are not duplicates of one another; they request unrelated features in different parts of the library.", + "summary": "All three issues are unrelated feature requests in different areas: tokenizer truncation behavior, learning-rate scheduling functionality, and Wav2Vec2 word boundary offsets. No duplicate cluster emerges from the titles or descriptions.", "confidence": 0.98, - "canonical_issue_reason": "No single canonical issue: the items cover unrelated requests (token truncation behavior, ReduceLROnPlateau-like scheduling, and Wav2Vec2 word offsets).", + "canonical_issue_reason": "No clear canonical duplicate: the issues describe distinct feature requests with different code paths and no shared targets.", "canonical_pr_reason": null, - "best_issue_reason": "issue:10519 is the most concrete and narrowly scoped request, but it should not be treated as a canonical duplicate for the others.", + "best_issue_reason": "issue:10519 is the most representative only by slightly broader wording and discussion, but it is still unrelated to the other two; if one issue must anchor the cluster, this is the least specific.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Different topics: truncation from the beginning vs. Wav2Vec2 word start/end offsets." + "reason": "Tokenizer truncation options and Wav2Vec2 word offset extraction are unrelated problems." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "Different topics: truncation behavior vs. ReduceLROnPlateau-like learning-rate scheduling." + "reason": "Truncation-from-beginning behavior is unrelated to ReduceLROnPlateau-like scheduler functionality." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "Different topics: scheduler functionality vs. extracting Wav2Vec2 word timing offsets." + "reason": "Learning-rate scheduling features and Wav2Vec2 word offset timing are different subsystems and not the same bug or change." } ] }, @@ -17053,7 +17209,7 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "ff92f07787566a7c34c16ba09c17d7b146ba179e536a63c41cc087e29de0c599", + "prepared_review_unit_hash": "6a61ee3b18fc0c0a7876602dc51e8fab3e76b2aceacce1923f3b8eba0b928801", "component_budget": { "node_count": 3, "item_count": 3, @@ -17075,24 +17231,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These issues do not look like true duplicates: one is about Tapas tokenization mismatch, one about an incorrect auth-related error when downloading private models, and one about a generic failure to load bert-base-uncased. The only mild overlap is around model-loading UX, not the same underlying bug.", - "confidence": 0.31, - "canonical_issue_reason": "No strong canonical issue exists because the topics differ. If one must be chosen, #16618 is the broadest model-loading report and the closest to the other two by title, but it is still not the same bug as the auth error or Tapas tokenization issue.", + "summary": "These three issues do not appear to be true duplicates: one is about Tapas tokenization deviating from TensorFlow, one is about a misleading auth/download error for private models, and one is about a generic inability to load bert-base-uncased. The soft links are too broad to merge confidently.", + "confidence": 0.89, + "canonical_issue_reason": "No exact duplicate stands out; if a single anchor is needed, issue:16618 is the broadest and most central model-loading report, with the most discussion activity.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the best representative only in the sense that it is the most general and closest to the shared 'model loading' wording; however, it should not be treated as a true duplicate target for the other issues.", + "best_issue_reason": "issue:16618 is the best cluster anchor because it is the most general user-facing failure report and could subsume nearby model-loading problems better than the more specialized tokenization or auth-message issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Tapas tokenization mismatch and bert-base-uncased loading failure are unrelated code paths; one concerns tokenizer parity, the other a model load error." + "reason": "Different underlying problems: Tapas tokenization parity with TensorFlow vs a generic bert-base-uncased loading failure. Same ML area, but not the same bug." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "#15980 is specifically about the wrong error when a private model download is attempted without login, while #16618 is a generic load failure for a public model; same broad area, different underlying bug." + "reason": "This is about an error message when downloading a private model without login, not the same concrete failure as loading bert-base-uncased." } ] }, @@ -17122,12 +17278,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "8c45da95a32f4cf0d9ad84a91386f4dfd723ffda7cd05e0da5c978614dd51fce", + "prepared_review_unit_hash": "78e8a0417f12fdb9243cb6b8276913359cac768ddbf3622671c47b42564fa50f", "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 }, @@ -17135,7 +17291,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -17144,18 +17300,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate community feature requests: one for Flash Attention 2 support across more architectures, and one for GGUF support across more architectures. They share a broad theme but not the same underlying change or code path.", - "confidence": 0.98, - "canonical_issue_reason": "Issue 26350 is the stronger anchor only by activity/engagement, but it is not a duplicate of 33260 because the requested features are different.", + "summary": "These are two separate feature-request issues: one for Flash Attention 2 architecture support and one for GGUF architecture support. They share a loose \u201ccommunity contribution\u201d framing but not the same underlying change or code path.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "26350 has much higher discussion activity and is the more established tracker, so it is the best representative issue here even though the requests are unrelated.", + "best_issue_reason": "Issue 26350 is the better representative issue only because it is older and has substantially more discussion activity; however, it is not a duplicate of 33260.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different feature requests (Flash Attention 2 vs GGUF support); shared wording about 'more architectures' is too broad to indicate the same underlying bug or change." + "reason": "Different requested capabilities (Flash Attention 2 vs GGUF) and different implementation areas; the shared wording is only a broad support request, not the same bug/change." } ] }, @@ -17185,12 +17341,12 @@ "soft_pairs": [ "issue:26413|issue:31501" ], - "prepared_review_unit_hash": "37670e0b969fa8bf6c1a2ee26fa2a530c83bd1951e908505673cc72dc9a44a50", + "prepared_review_unit_hash": "9f38b887dc888240898ea2a40c8b8a639f18aa85bac6ca53c242f896f4cff4b0", "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 }, @@ -17198,7 +17354,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1593, + "serialized_chars": 1594, "estimated_input_tokens": 399, "estimated_eval_tokens": 1054 }, @@ -17207,18 +17363,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues both concern training startup failures, but they do not clearly describe the same underlying bug. The checkpoint-resume issue has a concrete error path; the FineWeb training issue is broader and could have many causes.", - "confidence": 0.76, - "canonical_issue_reason": "issue #26413 is the more specific, technically grounded report with a clear failure signature (`resume_from_checkpoint` / empty `epoch_iterator`), making it the better anchor for this cluster.", + "summary": "The two issues are only loosely related at the level of training failing to start, but they do not clearly describe the same underlying bug. One is a specific `resume_from_checkpoint`/epoch iterator failure, the other is a broad FineWeb SLM training startup problem.", + "confidence": 0.87, + "canonical_issue_reason": "Issue 26413 is the better canonical issue because it names a concrete failure mode (`resume_from_checkpoint` causing an epoch_iterator error) and is more specific and actionable than the broader training-start symptom in 31501.", "canonical_pr_reason": null, - "best_issue_reason": "issue #26413 is the best representative because it states an exact error condition and code path, while #31501 is a high-level symptom report without enough evidence of the same root cause.", + "best_issue_reason": "Issue 26413 is the best representative because it targets a distinct code-path and failure message, making it easier to map to a single fix or regression.", "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 generic \u201ctraining doesn\u2019t start\u201d report. There is not enough evidence they describe the same underlying bug." + "reason": "Both concern training problems, but the evidence is too weak to treat them as the same bug; 26413 is specifically about checkpoint resumption and an empty epoch iterator, while 31501 is a broad startup failure with no clear match to that code path." } ] }, @@ -17277,7 +17433,7 @@ "issue:44370|issue:46226", "issue:46222|issue:46226" ], - "prepared_review_unit_hash": "17c198444253ef129a9b2afc134de84cbdaac3f35c23cd662abb3b2fe589547e", + "prepared_review_unit_hash": "a577970bda1f6ae0f151100d2b91000a313d229e7b86a584d90338db695a8e34", "component_budget": { "node_count": 8, "item_count": 8, @@ -17299,166 +17455,166 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of similarly worded i18n doc-translation requests, but they are not the same underlying issue across languages. The only clear duplicate-like pair is the exact Thai duplicate 46222/46226; the rest look like separate translation requests despite high textual similarity from the shared template.", - "confidence": 0.67, - "canonical_issue_reason": "issue:46222 is the strongest duplicate anchor because it is text-identical to issue:46226, making it the clearest representative of the only confirmed duplicate subcluster.", + "summary": "This cluster looks like one soft-similarity group around the generic i18n template issues plus a very similar Thai-text pair, but the packet does not support treating any pair as confirmed duplicates because every edge is marked deterministic_accept:false. The Turkish, Spanish, and Romanian issues are best treated as separate language-specific requests rather than one merged underlying issue.", + "confidence": 0.66, + "canonical_issue_reason": "issue:43834 is the best representative of the generic-template i18n subgroup because it is one of the earliest near-identical placeholder issues, but this should be treated as a soft representative only, not a confirmed duplicate anchor.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46222 is the best overall issue representative because it cleanly matches an exact duplicate (46226) and has the strongest evidence of being the canonical copy among the issues here.", + "best_issue_reason": "issue:43834 is the strongest global representative of the cluster\u2019s generic-template pattern, since it sits at the center of the near-identical placeholder issues and is the cleanest summary node for triage.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different language translation requests; same template wording but not the same underlying work." + "reason": "Same broad i18n topic, but Turkish vs generic placeholder request; not enough evidence for a true duplicate." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different language translation requests; similarity is from the shared i18n template, not a duplicate." + "reason": "Turkish language request does not match the generic placeholder issue as the same underlying report." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Turkish vs a separate placeholder translation issue; not enough evidence of the same concrete request." + "reason": "Turkish request and generic template issue are only loosely similar, not the same issue." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Different languages (Turkish vs Romanian); these are separate translation tasks." + "reason": "Turkish and Romanian translation requests are distinct language-specific issues." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different languages (Turkish vs Spanish); not a duplicate." + "reason": "Turkish and Spanish translation requests are different language targets, so not duplicates." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish translation request versus a separate placeholder translation issue; high similarity is just template overlap." + "reason": "Spanish request is only soft-similar to the generic template issue, not the same underlying report." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Different translation requests; not the same underlying issue." + "reason": "Spanish request and the generic placeholder issue are similar in structure but not clearly the same issue." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Spanish vs Romanian translation requests are distinct." + "reason": "Spanish and Romanian requests are separate language-specific issues." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Shared template only; no evidence they refer to the same language/change." + "reason": "Spanish request is not the same concrete issue as the generic placeholder report." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian translation request versus a separate placeholder issue; not clearly the same work." + "reason": "Romanian request and generic template issue are only soft-similar, not confirmed duplicates." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Same broad i18n area, but different translation requests." + "reason": "Romanian request does not appear to be the same underlying report as the generic template issue." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian vs another placeholder translation issue; similarity is not sufficient to call a duplicate." + "reason": "Romanian request is only structurally similar to the generic placeholder issue." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Different language requests; not the same underlying change." + "reason": "Romanian request is unrelated to the Thai-text issue." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Different language requests; not a duplicate." + "reason": "Romanian request is unrelated to the Thai-text issue." }, { "left": "issue:43834", "right": "issue:44370", "accept": false, - "reason": "Near-identical template text, but no concrete evidence they are the same translation request." + "reason": "These are very similar generic-template issues, but the packet still only gives soft similarity, not a confirmed duplicate signal." }, { "left": "issue:43834", "right": "issue:44365", "accept": false, - "reason": "Looks like separate placeholder issues rather than one underlying task." + "reason": "Near-identical placeholder structure, but not enough evidence to merge as confirmed duplicates." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "Different language/issue context; not enough to merge." + "reason": "Generic template issue vs Thai-text issue; different content and not the same request." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Different language/issue context; not a confirmed duplicate." + "reason": "Generic template issue vs Thai-text issue; different content and not the same request." }, { "left": "issue:44365", "right": "issue:44370", "accept": false, - "reason": "Very similar template wording, but no explicit evidence of the same underlying issue." + "reason": "Strong textual similarity, but still only soft-similarity evidence in the packet." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Different translation requests; do not merge on template similarity alone." + "reason": "Generic template issue and Thai-text issue are different underlying reports." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Different translation requests; not the same bug/change." + "reason": "Generic template issue and Thai-text issue are different underlying reports." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Different translation requests; shared template text is not enough." + "reason": "Generic template issue vs Thai-text issue; not the same concrete bug/request." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Different translation requests; not a duplicate." + "reason": "Generic template issue vs Thai-text issue; not the same concrete bug/request." }, { "left": "issue:46222", "right": "issue:46226", - "accept": true, - "reason": "These are text-identical duplicate Thai translation issues with no distinguishing evidence." + "accept": false, + "reason": "These Thai-text issues are highly similar, but the packet does not justify upgrading that similarity to a confirmed duplicate." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative: the only fully supported duplicate pair is issue:46222/issue:46226, which have identical titles and maximal similarity, while the other high-similarity pairs are reasonably treated as separate translation requests because the titles indicate different languages or placeholder variants. No overreach on mergeability or PR pairing." + "accept": false, + "feedback": "The clustering summary is too strong relative to the packet. The packet only provides title/body similarity and shows `deterministic_accept: false` for every listed pair, so statements like \"true duplicate subgroup,\" \"same underlying request,\" or exact duplicate acceptance between the generic template issues and the Thai pair are not grounded. In particular, marking `issue:46222|issue:46226` as accepted and treating the generic template issues as confirmed duplicates overstates the evidence. The conservative part about the Turkish/Spanish/Romanian issues being distinct language-specific requests is reasonable, but the overall verdict should stay at the soft-similarity level." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -17530,12 +17686,12 @@ "issue:43837|issue:45177", "issue:45322|issue:46133" ], - "prepared_review_unit_hash": "3868bfb8e2cf6ceb881c130fe300e5589df9acbf7e016de416c1d232464df89c", + "prepared_review_unit_hash": "dd024e2b05e7aa45e9a7fc264b29f34b5feb98ca483af56e584430f86394bae5", "component_budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 36, - "serialized_chars": 21660, + "serialized_chars": 21658, "estimated_input_tokens": 5415, "estimated_eval_tokens": 11086 }, @@ -17543,7 +17699,7 @@ "node_count": 16, "item_count": 16, "soft_pair_count": 36, - "serialized_chars": 21660, + "serialized_chars": 21658, "estimated_input_tokens": 5415, "estimated_eval_tokens": 11086 }, @@ -17552,228 +17708,228 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a set of independent model-addition/support requests across different model families and tasks (BERT variants, detection, audio/TTS/ASR, OCR, vision-language). They share the broad pattern of \"add model support,\" but none appear to be the same underlying change.", - "confidence": 0.92, - "canonical_issue_reason": "issue:43671 is the most central and broadly connected request in the set, with the strongest soft links to many of the other model-support issues. It best represents the cluster theme, but it is still a specific Qwen3-TTS request rather than a duplicate of the others.", + "summary": "This is a loose thematic cluster of model-support requests, not a duplicate set. Most items ask to add different models or modalities (text encoders, detectors, ASR/TTS, OCR, vision-language), so they should stay separate.", + "confidence": 0.56, + "canonical_issue_reason": "issue:43671 is the best cluster anchor: it is a concrete model-support request and sits closest to the other speech/audio-related items, but it is still not a duplicate of them.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43671 is the best representative issue for this cluster because it sits at the center of the similarity graph and reflects the common model-support request pattern most clearly. It should not be treated as a duplicate target for the other items.", + "best_issue_reason": "issue:43671 is the most representative anchor for this loose group of model-add requests, with the strongest overlap in theme and terminology among the speech/audio proposals.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "Different model families: EuroBert config vs DEIMv2 support." + "reason": "Different model requests: EuroBert config support vs DEIMv2. Same broad 'add model' pattern, but not the same change." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "BERT config request vs Qwen3-TTS audio support; different task and code path." + "reason": "EuroBert is a text model/config request; Qwen3-TTS is a speech synthesis request. Different underlying code paths." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert vs RF-DETR are unrelated models." + "reason": "EuroBert and RF-DETR are unrelated model families (text encoder vs object detector)." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR detection support is unrelated to DEIMv2." + "reason": "Both are detection-related, but RF-DETR and DEIMv2 are distinct detector models with separate integrations." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "RF-DETR detection vs LFM2.5 audio support are different changes." + "reason": "RF-DETR is an object detector; LFM2.5 Audio 1.5B is an audio model. Different functionality." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "Different model families and implementation paths." + "reason": "RF-DETR and the weight-sparse transformer request are different architectures and different implementation targets." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR vs Qwen3-ASR are different modalities and models." + "reason": "Object detection support vs ASR support are different model capabilities and code paths." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR and Qwen3-TTS are unrelated requests." + "reason": "RF-DETR and Qwen3-TTS are unrelated tasks: detection vs text-to-speech." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT support is unrelated to Qwen3-ASR." + "reason": "NeoBERT and Qwen3-ASR are different model families and different modalities." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT and weight-sparse transformer are different model additions." + "reason": "NeoBERT support vs a sparse-transformer model request are distinct changes." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT vs audio model support are different changes." + "reason": "NeoBERT is a text model; LFM2.5 Audio 1.5B is an audio model." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are distinct model requests." + "reason": "RF-DETR and NeoBERT are unrelated model additions." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are different model families." + "reason": "NeoBERT and ModernVBERT are different model requests, not the same underlying fix." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 support is separate from Qwen3-ASR support." + "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different model integrations." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 and weight-sparse transformer are not the same code change." + "reason": "MiniCPM3 and the OpenAI sparse-transformer request are unrelated models." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 vs ModernVBERT are different model requests." + "reason": "Both are model-add requests, but MiniCPM3 and ModernVBERT are distinct families and likely separate implementation work." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 and Voxtral TTS are unrelated models." + "reason": "MiniCPM3 support and Voxtral TTS support are different modalities and code paths." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are different models and tasks." + "reason": "DEIMv2 detection support and Qwen3-TTS support are different model types." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models and Qwen3-ASR are unrelated." + "reason": "ModernVBERT and Qwen3-ASR are unrelated model families and modalities." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT and Qwen3-TTS are different support requests." + "reason": "ModernVBERT is a BERT-style text model; Qwen3-TTS is speech synthesis." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "Related BERT wording, but not the same underlying change: model addition vs RoPE variant support." + "reason": "ModernVBERT and 'BERT-like models with RoPE' are only superficially related; one is a specific model request, the other is a general architectural pattern." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like RoPE support is unrelated to Voxtral TTS." + "reason": "General BERT-like RoPE discussion does not match Voxtral TTS support; different model classes." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS and Qwen3-ASR are different tasks and would not be one merged PR." + "reason": "Same model family, but TTS and ASR are different capabilities and would not be merged into one PR." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Both are TTS-related in title only; they target different model families and code paths." + "reason": "Both are speech models, but Qwen3-TTS and Voxtral-TTS are different model families with separate integrations." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "TTS support and OCR2 support are unrelated." + "reason": "TTS support and DeepSeek-OCR2 support are different modalities and changes." }, { "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "Qwen3-TTS support is unrelated to the sparse transformer request." + "reason": "Speech model support vs sparse-transformer support are unrelated changes." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio are distinct model additions." + "reason": "Both are audio-related, but Qwen3-TTS and LFM2.5 Audio 1.5B are distinct models/capabilities." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "ASR support vs audio model support are different requests." + "reason": "ASR support and LFM2.5 Audio support are not the same underlying model change." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR and EUPE are different model families and tasks." + "reason": "ASR support and EUPE perception-encoder support are different model requests." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "ASR support is unrelated to the weight-sparse transformer request." + "reason": "ASR support and a weight-sparse transformer request are unrelated." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "Audio model support vs SAM3-LiteText are unrelated." + "reason": "Audio model support vs SAM3-LiteText support are different modalities and integrations." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer and EUPE are different underlying changes." + "reason": "Sparse-transformer support and EUPE support are different underlying changes." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS are different model support changes." + "reason": "ASR and TTS are distinct speech tasks; these would not be one merged fix." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral TTS are unrelated model requests." + "reason": "NeoBERT and Voxtral-TTS are unrelated model families." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "ASR support and DeepSeek-OCR2 support are different tasks." + "reason": "ASR support and OCR2 support are different code paths and model types." }, { "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE and TIPSv2 are distinct vision-language model requests." + "reason": "EUPE and TIPSv2 are separate vision-language model requests; related broad area, but not the same concrete change." } ] }, @@ -17803,12 +17959,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "afcf19f5cfa2617fa95ec40963375d01fc2748d28373b1e14f13e5d9b068fa8f", + "prepared_review_unit_hash": "b46e848e79ebfdefd15f150a66bd8098cc95365a74a19ae402de11e8a870f6cf", "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 }, @@ -17816,7 +17972,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, + "serialized_chars": 1479, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -17825,18 +17981,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one asks for Set Block Decoding, the other requests DeepSeek v3.2 support. They describe different user-facing problems and have no shared targets or code-path evidence.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "These two issues do not look like duplicates: one asks for generic block decoding changes, while the other requests support for a specific model version (DeepSeek v3.2).", + "confidence": 0.95, + "canonical_issue_reason": "No canonical issue: the titles/body indicate different topics rather than the same underlying bug or feature request.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #43037 is the more concrete and actionable request, but it is still unrelated to #41084 rather than a duplicate of it.", + "best_issue_reason": "Issue 43037 is the more concrete and user-facing request, but it is still not a duplicate of 41084.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different topics entirely: block decoding vs model support for DeepSeek v3.2. No shared filenames, targets, or code path evidence." + "reason": "The requests are unrelated at the level needed for duplicate triage: block decoding vs. DeepSeek v3.2 support." } ] }, @@ -17866,12 +18022,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "ee7f3caea1d265221a4c99a923b80b6e92f37127fcfe2ab261200f352e17fbe9", + "prepared_review_unit_hash": "bb8ba802e3d3406b3870d9fea03e377faad2b70268cabc37cffe7697f6bc6ed2", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1583, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -17879,7 +18035,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1583, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -17888,18 +18044,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at a very high level (Trainer customization), but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. I would not treat them as duplicates.", - "confidence": 0.94, - "canonical_issue_reason": "Issue 43598 is the more general Trainer customizability request and is the better representative if one issue must be kept.", + "summary": "These are related thematically around Trainer customization, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPs estimation for custom trainers. I would not merge them as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "Issue #43598 is the broader, more representative Trainer customizability request; it centers on a core Trainer behavior rather than a narrower metrics/estimation edge case.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43598 is the better global anchor for this cluster because it centers on Trainer behavior customization rather than a narrower utility concern.", + "best_issue_reason": "#43598 is the best issue to keep as the cluster representative because it addresses a foundational Trainer customization point and is phrased as a general design concern.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "They concern different code paths and outcomes: `remove_unused_column` customizability versus FLOPS estimation for custom trainers. Shared Trainer context is too broad to imply the same underlying bug." + "reason": "They share the general theme of custom trainers in `Trainer`, but the underlying problems are different: one is about column removal/customizability, the other about FLOPs counting. They would not be fixed by the same change." } ] }, @@ -17929,12 +18085,12 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "888e9bf61323e8c0a8cdd3ba357fe6d5dc49ce63bf5a15a6f5b244489aa8219b", + "prepared_review_unit_hash": "a3192c02b5be04c25afe036aa839dbd523d11d62078a4ad42ce8e5760883af27", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1611, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -17942,7 +18098,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1611, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -17951,18 +18107,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at a very broad level (Qwen model support), but they describe different model families and different auto-class registration gaps, so they should not be deduplicated.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 43684 is the broader, more central request: it asks for model registration in AutoModel and AutoModelForConditionalGeneration, which is the more general missing-support problem.", + "summary": "The two issues are about different Qwen model families and different API surfaces: one requests Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration, while the other requests AutoModelForSequenceClassification support for Qwen3.5. They are related only at a broad \"Qwen model support\" level, not the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 43684 is the best representative because it covers the core registration gap in the auto-model machinery, whereas 44405 is a narrower sequence-classification-specific request for a different config family.", + "best_issue_reason": "Issue 43684 is the broader model-registration request, so it is the better cluster anchor, but it is still not a duplicate of 44405.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different underlying changes: one requests Qwen3-Omni registration for generic/conditional-generation auto classes, the other requests AutoModelForSequenceClassification support for Qwen3.5. Same broad area, but not the same bug or mergeable fix." + "reason": "Different model variants and different heads/APIs; this is not the same bug or change and would not plausibly be merged as one issue." } ] }, @@ -17999,12 +18155,12 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "0092da70d7afde1e86b61af67a0bfb8ef34dbc43c9f41841072d29b6b41776b0", + "prepared_review_unit_hash": "5623e52b10313574a62cb62f7b95cd6d9a207e33f2b4a76c59da6690b07dc1cd", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4267, + "serialized_chars": 4266, "estimated_input_tokens": 1067, "estimated_eval_tokens": 2390 }, @@ -18012,7 +18168,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4267, + "serialized_chars": 4266, "estimated_input_tokens": 1067, "estimated_eval_tokens": 2390 }, @@ -18021,55 +18177,58 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All four items are low-information closed issues. Three share the exact title \"huggingface \" but there is no body/context to confirm they describe the same underlying bug; the other issue titled \"Fett\" looks unrelated. I would keep one of the repeated \"huggingface \" reports as the representative issue, but I would not merge any pair based on the available evidence.", - "confidence": 0.89, - "canonical_issue_reason": "issue:43845 is the best representative of the repeated \"huggingface \" reports because it is the earliest of the near-identical titles and has no less evidence than the later duplicates.", + "summary": "Three issues are near-identical low-signal \"huggingface\" reports; one \"Fett\" issue appears unrelated. I treat 43845 and 43846 as the clearest duplicates, while 43859 is too under-specified to confidently merge with them from metadata alone.", + "confidence": 0.78, + "canonical_issue_reason": "issue:43845 is the best representative of the duplicate \"huggingface\" pair: it is the earliest of the two identical-title, identical body-length reports and has the clearest match to issue:43846.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43845 is the strongest global issue candidate: it matches the repeated title cluster, is earliest among the duplicates, and there is no evidence that another issue is a better source report.", + "best_issue_reason": "issue:43845 is the strongest global issue representative because it anchors the only convincingly duplicate pair (43845/43846) and is more complete than 43859.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no body details. \"Fett\" has no visible relation to the repeated \"huggingface \" reports." + "reason": "Different titles and no shared concrete evidence beyond generic similarity." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no substantive description to indicate the same underlying bug." + "reason": "Different titles and no shared concrete evidence beyond generic similarity." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles and no evidence tying this issue to the \"huggingface \" reports." + "reason": "Different titles and no shared concrete evidence beyond generic similarity." }, { "left": "issue:43845", "right": "issue:43846", - "accept": false, - "reason": "Titles match, but the issues are too empty to confirm they report the same concrete problem; identical phrasing alone is not enough." + "accept": true, + "reason": "Same exact title and same body length, which strongly suggests the same underlying issue." }, { "left": "issue:43845", "right": "issue:43859", "accept": false, - "reason": "Same generic title, but there is no body or technical context to prove these are duplicate reports of one bug." + "reason": "Same broad title, but the shorter body and later activity make this too weak to treat as the same issue from metadata alone." }, { "left": "issue:43846", "right": "issue:43859", "accept": false, - "reason": "Same generic title, but the available evidence is insufficient to establish they describe the same underlying issue." + "reason": "Same broad title, but the shorter body and later activity make this too weak to treat as the same issue from metadata alone." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet: 43845/43846 share the same title and body length, while 43835 is clearly different and 43859 is weaker evidence from metadata alone. The soft-edge verdicts are conservative overall, and the single accepted duplicate pair is supported by the strongest available signals." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -18103,88 +18262,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "7ade19427739b93dd009c2ff6f807100171492bf62a5d0bd034c9ce677180429", + "prepared_review_unit_hash": "99c954f4682a076783c6f8338c4953f5c90d346194c4ca447711b14e722b6731", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5788, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5788, - "estimated_input_tokens": 1447, - "estimated_eval_tokens": 3150 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster looks like unrelated low-information/spam issues rather than true duplicates. There is no evidence of a shared bug or request, and the only somewhat descriptive title is still too vague to anchor a canonical duplicate.", - "confidence": 0.93, - "canonical_issue_reason": "No strong canonical issue: the items have different titles, no bodies or linked references, and no shared technical details. They do not appear to describe the same underlying problem.", + "summary": "This cluster is dominated by low-information spam/generic placeholder issues with no substantive bug or feature signal. The only apparent overlaps are superficial title matches, not evidence of the same underlying issue.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "If forced to pick a representative, issue 44418 is the least opaque because it has the only non-spam-like title, but it is still not a convincing canonical duplicate target.", + "best_issue_reason": "No issue in the set meaningfully represents a real problem statement; all are either spam placeholders or an unrelated generic documentation suggestion.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are titled \"\" with no body or links, so there is no concrete evidence they report the same issue." + "reason": "Both are titled \"\", but that is only a superficial match; there is no concrete bug, target, or code-path evidence to show they are the same underlying issue." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Same generic spam title, but no technical content or references to show a shared underlying bug." + "reason": "Identical spam titles do not establish a shared defect or change request; no supporting body or references indicate the same underlying issue." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is spam-like and the other is a generic documentation suggestion; they do not look like the same report." + "reason": "One is spam and the other is a generic documentation suggestion; these are not the same underlying issue." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "The titles are unrelated and there is no content suggesting a common root cause." + "reason": "A spam placeholder and a titled issue about \"Google LLC\" do not indicate the same bug or change." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "A vague documentation suggestion does not match a spam placeholder issue." + "reason": "A documentation enhancement request and a spam placeholder are unrelated." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "No shared subject matter beyond being low-information issues; not the same underlying bug or change." + "reason": "A documentation enhancement request and a spam placeholder are unrelated." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Identical spam titles are not enough to merge them as duplicates without any substantive body or evidence." + "reason": "Both are spam placeholders with no concrete problem description; identical titles alone are insufficient to treat them as the same underlying issue." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "The titles are unrelated and the available metadata gives no indication of the same problem." + "reason": "A spam placeholder and an unrelated titled issue do not match on any concrete defect or change." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "No shared technical details, targets, or references; these are not the same issue." + "reason": "A spam placeholder and an unrelated titled issue do not match on any concrete defect or change." } ] }, @@ -18214,12 +18373,12 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "ec34912142b4604103177982616d4fa4e7c58c20934f235dbb2601b3e38fc1a5", + "prepared_review_unit_hash": "f86c200bb5cc212e2035e96a5ed8f510eb9c0cb7648c5a3bce1f52b95ad34f1d", "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 }, @@ -18227,7 +18386,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1645, + "serialized_chars": 1646, "estimated_input_tokens": 412, "estimated_eval_tokens": 1080 }, @@ -18236,24 +18395,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues appear likely to be duplicates: they share the same title and very similar body sizes, and #45145 has the only inbound reference, so it is the better canonical representative.", - "confidence": 0.87, - "canonical_issue_reason": "#45145 appears to be the better canonical issue because it has an inbound reference, making it slightly better connected in the discussion trail.", + "summary": "These two closed issues are very likely duplicates of the same report: they have identical titles, very similar body lengths, and a soft-similarity signal. #45145 is the better canonical issue because it was created first and has the stronger reference activity.", + "confidence": 0.84, + "canonical_issue_reason": "#45145 is earlier and has the stronger reference activity, so it is the best canonical issue if these are treated as duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the strongest issue representative because it appears to be the more connected copy of the same likely duplicate report.", + "best_issue_reason": "#45145 is the strongest representative because it predates #45151 and is the better anchor for a probable duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": true, - "reason": "Identical titles and very similar body lengths make these issues appear to describe the same likely report, though the evidence is limited." + "reason": "Identical titles plus similar body size and a soft-similarity signal make this a probable duplicate, though the body text is not visible here so the match is not fully provable." } ] }, "evaluator_result": { "accept": false, - "feedback": "The pairing is plausible, but the writeup is a bit too strong for the evidence available. The packet only shows identical titles, similar body lengths, and a soft_similarity signal; it does not establish that these are definitely duplicates. Please soften the language to \u2018likely\u2019/\u2018appear to be\u2019 throughout, especially in the soft-edge reason and confidence." + "feedback": "The conclusion is plausible, but the writeup overstates the evidence. The packet only shows identical titles, similar body lengths, and a soft-similarity signal; it does not expose the body text needed to claim the reports are 'highly similar' or that they are definitively the same underlying issue. The canonical-choice reasoning is fine, but the soft-edge verdict and summary should be phrased more conservatively (e.g. 'likely duplicate' or 'probable duplicate') and the confidence should be lowered." }, "evaluator_used": true, "retried": true, @@ -18334,7 +18493,7 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "d22306be406110f45c523b58ae1363b61120985338f421e14dc925759f2a4732", + "prepared_review_unit_hash": "6eac56f30fd4015113e8620f5457ea5c4a07a42522a5f79094f263f96e1459c5", "component_budget": { "node_count": 2, "item_count": 2, @@ -18356,18 +18515,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related 4-bit inference issues, but they describe different bugs: one is a Gemma4/audio-module uint8 dtype failure in `torch.finfo()`, the other is a `BitsAndBytesConfig.llm_int8_skip_modules` regression that drops the default `lm_head` exclusion and triggers an assertion. They should not be deduplicated.", - "confidence": 0.91, - "canonical_issue_reason": "Issue #45674 is the more general and actionable report: it names the configuration regression, the affected setting, and the concrete 4-bit inference failure path.", + "summary": "These two issues are related only at a high level (4-bit inference / bitsandbytes), but they describe different failure modes and code paths: one is a Gemma4 audio-module uint8/`torch.finfo()` TypeError, the other is a `BitsAndBytesConfig.llm_int8_skip_modules` regression that removes the default `lm_head` exclusion and triggers an AssertionError. They should not be merged as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "Issue #45674 is the better canonical issue for this cluster because it has the clearer, configuration-centric regression description and appears broader/more reusable as a report of a BitsAndBytesConfig bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #45674 is the best representative of the cluster because it describes a central library-side bug rather than a model-specific symptom.", + "best_issue_reason": "Issue #45674 is the strongest representative issue: it states the concrete failing behavior, the triggering option, and the 4-bit inference context more explicitly than #45672.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Different underlying bugs and code paths. #45672 is Gemma4/audio-module-specific `torch.finfo()` TypeError on uint8 weights; #45674 is a `BitsAndBytesConfig` regression where `llm_int8_skip_modules` clears the default `lm_head` exclusion and causes an `AssertionError` in 4-bit inference." + "reason": "Both involve 4-bit inference, but the bugs are different: #45672 is a Gemma4/audio-module `torch.finfo()` TypeError on uint8 weights, while #45674 is an `llm_int8_skip_modules`/`lm_head` exclusion regression causing AssertionError. Different underlying code paths and fixes." } ] }, @@ -18397,7 +18556,7 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "bfbb7a22b10662deda6f06fa65fa922505f2f0dcd0cbe7881d8d2718c7b2406d", + "prepared_review_unit_hash": "4b6726ebcc38c080f374d0f065765c5eb44730f84d4173ecf246af7d1704d230", "component_budget": { "node_count": 2, "item_count": 2, @@ -18419,18 +18578,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues are not duplicates; they are unrelated placeholder/personal-name titles with no shared bug report context.", - "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the issues do not describe the same problem or request, so neither can serve as a representative duplicate target.", + "summary": "The two closed issues are superficially similar only in that they are both empty, name-like entries; they do not describe the same bug or change and should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "Issue 45762 is the only reasonable placeholder canonical because it was created first, but it is not a substantive report and has no content to represent a real bug cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a good global representative; both are standalone, unrelated entries with no actionable problem statement.", + "best_issue_reason": "Issue 45762 is marginally preferable as the cluster representative due to earlier creation, but neither issue is suitable as a true canonical bug report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Titles only share a surname-like token and both are bare name/date strings; there is no evidence they report the same underlying issue." + "reason": "Reject: the titles are different personal-name/date strings, with no body, targets, or discussion tying them to the same underlying issue." } ] }, @@ -18460,30 +18619,30 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "5c956266222604b78382b0c18326765328a9b9a0b956c2cd79323214457cc791", + "prepared_review_unit_hash": "c261127724d3050348ceb6d3ccbf1dc51c2f857c99c96845f5766dca2b326f1a", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, - "estimated_input_tokens": 370, - "estimated_eval_tokens": 996 + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1480, - "estimated_input_tokens": 370, - "estimated_eval_tokens": 996 + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are not convincing duplicates: they are just two different personal-name/date titles with no body, no links, and no shared technical context. The soft similarity signal is too weak to treat them as the same underlying issue.", - "confidence": 0.96, + "summary": "The two closed issues appear unrelated: both are personal-name/date strings with no body or linked context, so there is no evidence they describe the same bug or request.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -18493,7 +18652,7 @@ "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "Different names and dates, no description or evidence of a shared bug/report; not the same underlying issue." + "reason": "Different named/date-only issue titles with no shared context, body, or links; not enough evidence they are duplicates." } ] }, @@ -18523,40 +18682,40 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "da00cc043a3158f7fd01531d3ec003a10d2f5e0b4ac107f1930e188cf5e09d38", + "prepared_review_unit_hash": "dc9c24502f0d75bef955755724e8489d63dcbb8ca3bd69b464e736e6e6a2d0c8", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1540, - "estimated_input_tokens": 385, - "estimated_eval_tokens": 1026 + "serialized_chars": 1541, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1540, - "estimated_input_tokens": 385, - "estimated_eval_tokens": 1026 + "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": "These two issues are not duplicates: one requests a new Qwen token-classification model class, while the other asks for FP8 training support.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "These are unrelated open feature requests: one asks for Qwen3_5 token-classification support for value-model use, the other asks for FP8 training support.", + "confidence": 0.96, + "canonical_issue_reason": "issue:46295 is the broader, more general request, so it is the better cluster anchor if one must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:46295 is the more generally applicable issue, but there is no evidence these two are duplicates or the same underlying problem.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Different feature requests in different areas: model class addition vs FP8 training support. They do not describe the same underlying bug or change." + "reason": "These requests concern different areas of the stack and different changes: model-head support for token classification versus FP8 training infrastructure. They do not describe the same underlying bug or change." } ] }, @@ -18596,88 +18755,88 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "3fd48f9b1e16ab7fbc31db8c262ebaef503f8fe96e9b9af2d69cff3719906124", + "prepared_review_unit_hash": "84c7f83c4b30ecd2a58ae76043daf86cc1eb963765f5fd019786411290962b9b", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5727, - "estimated_input_tokens": 1432, - "estimated_eval_tokens": 3120 + "serialized_chars": 5723, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5727, - "estimated_input_tokens": 1432, - "estimated_eval_tokens": 3120 + "serialized_chars": 5723, + "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 most likely contains two strong duplicate pairs\u201445870/45871 and 45872/45873\u2014plus one unrelated dataset post (45869). The evidence supports likely duplicates, but not confirmed exact duplication from the packet alone.", - "confidence": 0.9, - "canonical_issue_reason": "No single canonical issue fits all five items because the cluster splits into two separate likely-duplicate pairs with different concrete topics, plus one unrelated issue.", + "summary": "Best-effort split: the packet looks like two likely duplicate issue pairs (45870/45871 and 45872/45873) plus one separate dataset request (45869). The duplicate evidence is strong but still soft, so I would treat the pairs as likely/near-identical rather than confirmed exact duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "No single canonical issue for the whole cluster. Best-effort grouping suggests two near-identical issue pairs (45870/45871 and 45872/45873) and one distinct issue (45869), but the evidence is soft rather than definitive.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the earliest clean representative of the 'Example: Financial time series prediction with DataForge' pair, but it only represents that pair and not the separate scraping-service pair.", + "best_issue_reason": "45870 is the best representative if one issue must be chosen, because it is the earliest issue in the near-identical 45870/45871 pair. It should be treated as a representative for that pair only, not for the full cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Dataset bundle promo vs. financial prediction example: related finance wording, but different underlying request." + "reason": "Different request type: dataset bundle vs. financial time-series example. Related theme only, not the same underlying issue." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Dataset bundle promo vs. financial prediction example; not the same issue." + "reason": "Different request type: dataset bundle vs. financial time-series example. Related theme only, not the same underlying issue." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Same title and matching body length with very strong similarity; this is a likely duplicate pair, though not proven exact from metadata alone." + "reason": "Near-identical issues: same title, matching body length, and strong soft-similarity signals. Likely duplicates." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "Prediction example vs. scraping-service integration are different concrete asks." + "reason": "Same broad DataForge/financial theme, but the concrete ask is different: prediction example versus scraping service." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Prediction example vs. scraping-service integration are different concrete asks." + "reason": "Same broad DataForge/financial theme, but the concrete ask is different: prediction example versus scraping service." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "Different underlying topics despite broad DataForge/finance overlap." + "reason": "Same broad DataForge/financial theme, but the concrete ask is different: prediction example versus scraping service." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Different underlying topics despite broad DataForge/finance overlap." + "reason": "Same broad DataForge/financial theme, but the concrete ask is different: prediction example versus scraping service." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Same title and matching body length with very strong similarity; this is a likely duplicate pair, though not proven exact from metadata alone." + "reason": "Near-identical issues: same title, matching body length, and strong soft-similarity signals. Likely duplicates." } ] }, "evaluator_result": { "accept": false, - "feedback": "The analysis is too strong for the evidence in the packet. We only have titles, body lengths, and soft-similarity signals; there is no body text proving exact duplication. In particular, calling 45870/45871 and 45872/45873 \"exact duplicate\" pairs overstates what is grounded here. A more conservative summary would say they are likely duplicates or strong matches, not confirmed exact duplicates. " + "feedback": "The cluster split is plausible, but the wording is slightly too strong for the packet. The evidence only shows soft similarity plus matching titles/body lengths, not confirmed exact duplicates. Please soften the verdicts to \"near-identical\" or \"likely duplicates\" rather than asserting exact duplication, and keep the canonical-issue statement framed as a best-effort split rather than a definitive duplicate grouping." }, "evaluator_used": true, "retried": true, @@ -18717,8 +18876,8 @@ ], "soft_pairs": [ "pull_request:39895|pull_request:43424", - "pull_request:41855|pull_request:43823", "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", @@ -18739,172 +18898,172 @@ "pull_request:44827|pull_request:46048", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "66c7f538f489f1a16daabb1d18d018c2f853d2d72de246a5d90ffc4651784c72", + "prepared_review_unit_hash": "5fe63c347924732467957d5b352624b66714f24ecde2a50a7fc11460dd85690c", "component_budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 22, - "serialized_chars": 43086, - "estimated_input_tokens": 10772, - "estimated_eval_tokens": 21800 + "serialized_chars": 43093, + "estimated_input_tokens": 10774, + "estimated_eval_tokens": 21804 }, "budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 22, - "serialized_chars": 43086, - "estimated_input_tokens": 10772, - "estimated_eval_tokens": 21800 + "serialized_chars": 43093, + "estimated_input_tokens": 10774, + "estimated_eval_tokens": 21804 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Only one true duplicate pair stands out: two PRs that both extend `fixed_cross_entropy` to forward `weight` and `label_smoothing` into `nn.functional.cross_entropy`. The rest are unrelated changes that only overlap at a broad subsystem level or via generic soft similarity.", - "confidence": 0.94, + "summary": "This cluster is mostly heterogeneous: the PRs cover unrelated model additions, refactors, docs, and bug fixes. The only clear duplicate-style overlap is the fixed_cross_entropy kwargs pair (43251/43254), which targets the same loss-path change and the same tracking issue.", + "confidence": 0.84, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the cleaner canonical representative of the shared loss-function fix: it has the explicit issue target (43240) and directly patches `fixed_cross_entropy` to pass the missing kwargs through.", + "canonical_pr_reason": "PR 43251 is the best representative because it implements the concrete cross_entropy kwargs fix, explicitly references the tracking issue, and is effectively the same change as 43254.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the best overall representative of the cluster because it is the only clearly grounded bug-fix PR here, with a concrete issue link and a narrowly scoped implementation in the core loss helper.", + "best_pr_reason": "PR 43251 is the strongest global representative: it fixes a specific runtime bug in fixed_cross_entropy and is the only PR in the set with a true duplicate partner.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Different changes: VideoPrism model/docs scaffolding vs an ExecuTorch dynamic-shape test." + "reason": "Unrelated changes: new VideoPrism model addition vs. an executorch dynamic-shape test." }, { "left": "pull_request:41855", - "right": "pull_request:43823", + "right": "pull_request:43424", "accept": false, - "reason": "Mistral tokenizer missing methods vs adding MobileLLM; unrelated model areas and code paths." + "reason": "Tokenizer API additions are unrelated to executorch export testing." }, { "left": "pull_request:41855", - "right": "pull_request:43424", + "right": "pull_request:43823", "accept": false, - "reason": "Tokenizer method additions vs ExecuTorch export test; no shared bug or change." + "reason": "Mistral tokenizer helper methods and adding MobileLLM are different model areas and different changes." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Processor auto-loading robustness changes vs MobileLLM model addition; unrelated." + "reason": "Processor robustness changes are not the same underlying bug as adding a new model." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss-function kwargs fix vs MobileLLM model addition; different concrete changes." + "reason": "Loss-function kwarg forwarding is unrelated to adding MobileLLM." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Loss-function kwargs fix vs Mistral4 test fixes; unrelated code paths." + "reason": "Different code paths: cross-entropy kwargs fix vs. Mistral4 test repairs." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "ExecuTorch dynamic-shape test vs Mistral4 test fixes; not the same underlying change." + "reason": "Executorch export coverage and Mistral4 test fixes are not the same change." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "ExecuTorch export test vs MobileLLM model addition; unrelated." + "reason": "No concrete overlap beyond both touching model-related code." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "MobileLLM model addition vs config-resolution ordering fix; different targets." + "reason": "A new model addition and a config-resolution fix are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "MobileLLM model addition vs Falcon output-interface refactor; unrelated." + "reason": "New model implementation vs. Falcon output-interface refactor are unrelated." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor vs model-creation skill draft; not the same bug or feature." + "reason": "Falcon refactoring and the model-creation skill are not the same code-path change." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output-interface refactor vs Mistral4 test fix; unrelated." + "reason": "Falcon refactor and Mistral4 test fixes are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:46344", "accept": false, - "reason": "Mistral4 test fixes vs VitPose flip_back graph-break fix; different models and issues." + "reason": "Mistral4 test fixes and VitPose graph-break fix target different bugs in different models." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test fixes vs layernorm renaming/mapping fixes; not the same change." + "reason": "A typo rename in CLIP-like model conversion is unrelated to Mistral4 tests." }, { "left": "pull_request:44827", "right": "pull_request:46267", "accept": false, - "reason": "Mistral4 test fixes vs chat-response parsing updates; unrelated." + "reason": "Mistral4 tests and chat parsing/serve updates do not share the same underlying change." }, { "left": "pull_request:44827", "right": "pull_request:46367", "accept": false, - "reason": "Mistral4 test fixes vs finegrained FP8 kernel rename/update; unrelated." + "reason": "Different fixes: Mistral4 tests vs. FP8 kernel integration changes." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 test fixes vs model-creation skill draft; unrelated." + "reason": "The model creation skill is a meta/documentation PR, not the same bug or code change as Mistral4 tests." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM model addition vs model-creation skill draft; unrelated artifacts." + "reason": "Adding MobileLLM is unrelated to the model creation skill PR." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 test fixes vs config-resolution ordering fix; unrelated." + "reason": "Mistral4 test fixes and config resolution ordering are different issues." }, { "left": "pull_request:46048", "right": "pull_request:46367", "accept": false, - "reason": "TDT loss kernel integration vs finegrained FP8 kernel API rename; different kernels and purposes." + "reason": "TDT loss kernel support and FP8 kernel integration changes affect different integrations and code paths." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 test fixes vs TDT loss kernel integration; unrelated." + "reason": "Mistral4 test fixes and TDT loss kernel support are unrelated." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs fix the same `fixed_cross_entropy` bug by forwarding supported kwargs (`weight`, `label_smoothing`) into `nn.functional.cross_entropy`, and both point at issue 43240." + "reason": "Both PRs implement the same concrete fixed_cross_entropy kwargs support on the same function and same issue target, so they are mergeable duplicates." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the lone accepted pair is a conservative call: PRs 43251 and 43254 make the same narrow `fixed_cross_entropy` kwargs-forwarding change in the same file and share issue target 43240, so they look mergeable into one bugfix PR. The rejected soft pairs are also conservatively framed as unrelated or merely broadly similar." + "feedback": "The assessment is well grounded in the packet: the cluster is mostly unrelated, and the only accepted pair (43251/43254) clearly shares the same file, the same issue target, and the same fixed_cross_entropy kwargs fix. The rejection calls are conservative and do not overstate similarity." }, "evaluator_used": true, "retried": false, @@ -18933,12 +19092,12 @@ "pull_request:40546|pull_request:42781", "pull_request:42781|pull_request:46331" ], - "prepared_review_unit_hash": "506819152e21cdc077232213ab320c8eba4f72998d0c73feb12eca92731b379a", + "prepared_review_unit_hash": "34e86a7080c8f763ef7ba233fcabdc2f7096a6e924f96c67b5198b9408f08d92", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10379, + "serialized_chars": 10380, "estimated_input_tokens": 2595, "estimated_eval_tokens": 5446 }, @@ -18946,7 +19105,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10379, + "serialized_chars": 10380, "estimated_input_tokens": 2595, "estimated_eval_tokens": 5446 }, @@ -18955,24 +19114,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three distinct feature PRs: an initial VibeVoice integration, a later VibeVoice Realtime extension, and an unrelated Parakeet-RNNT integration. They share some auto-registry and speech-model scaffolding, but they do not look like the same concrete change.", - "confidence": 0.95, + "summary": "This cluster mixes one base VibeVoice implementation PR, one separate VibeVoice Realtime follow-up, and one unrelated Parakeet-RNNT PR. The shared auto-registry/docs files are not enough to make any pair a true duplicate or mergeable single PR.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 40546 is the clearest canonical PR for the VibeVoice cluster because it is the original, broad VibeVoice implementation and docs addition; the others are separate follow-on or unrelated model integrations rather than the same change.", + "canonical_pr_reason": "pull_request:40546 is the original, broad VibeVoice implementation and is the clearest representative of the VibeVoice subset in this cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 40546 best represents the underlying VibeVoice feature introduction. PR 42781 is a later realtime/acoustic-tokenizer extension, and PR 46331 is for a different model family (Parakeet-RNNT).", + "best_pr_reason": "pull_request:40546 is the most representative single PR here because it introduces the core VibeVoice model support without mixing in the separate realtime extension or an unrelated model family.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Both are about VibeVoice, but 42781 is a distinct follow-up adding realtime and acoustic-tokenizer support plus additional generation/runtime changes. It is not the same concrete fix/feature as the original VibeVoice implementation and would not plausibly be merged as a single duplicate PR." + "reason": "Both are VibeVoice-related, but 42781 is a separate realtime/acoustic-tokenizer extension rather than the same concrete implementation change; they are not the same underlying bug or mergeable into one PR." }, { "left": "pull_request:42781", "right": "pull_request:46331", "accept": false, - "reason": "These are different model integrations: VibeVoice Realtime versus Parakeet-RNNT. Shared auto-mapping edits are generic registry updates, not evidence of the same underlying bug or change." + "reason": "These target different model families (VibeVoice vs Parakeet-RNNT); overlap is only in generic auto-registration files, so they are unrelated." } ] }, @@ -19002,12 +19161,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "313538dd2696a58deaa39e47334fe6003ca3e7adab0fa25912113659a43c82b9", + "prepared_review_unit_hash": "68c5322b25f3c6cbc8be67595f188004fdb30fcf4aaa76d103fca27ef8de8ee1", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4675, + "serialized_chars": 4674, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -19015,7 +19174,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4675, + "serialized_chars": 4674, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -19024,18 +19183,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are unrelated: one adds covariate forecasting support for TimesFM, while the other adds DINOv3 image-classification support. They do not describe the same bug, change, or code path.", - "confidence": 0.99, + "summary": "These are two unrelated feature PRs in different model families: TimesFM covariate forecasting versus DINOv3 image classification support. They do not target the same bug or change set.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: the PRs target different model families and implement unrelated features.", + "canonical_pr_reason": "pull_request:41224 is the cleaner canonical representative because it is a focused model capability addition with clear code, auto-model, and test updates in one subsystem.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a good global representative for the cluster because there is no duplicate relationship to collapse.", + "best_pr_reason": "pull_request:41224 is the best standalone representative of the cluster because it is a concrete, self-contained feature PR with explicit model support and tests, whereas 40755 is a separate TimesFM enhancement.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different subsystems and features: TimesFM covariate forecasting vs. DINOv3 image classification. They do not fix or implement the same underlying change." + "reason": "Reject: these PRs touch different models and different code paths. TimesFM covariate forecasting and DINOv3 image classification support are unrelated feature additions, not a single mergeable change." } ] }, @@ -19130,12 +19289,12 @@ "pull_request:41797|pull_request:45186", "pull_request:45186|pull_request:45630" ], - "prepared_review_unit_hash": "1758fc83c7883da502f043db78a9be497b6b6f78c94810db2edcb33ca13c32c0", + "prepared_review_unit_hash": "2c5fe783972a3997226169de37cf28f7d2f074771c717c1f01f13aad98f93e40", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10433, + "serialized_chars": 10436, "estimated_input_tokens": 2609, "estimated_eval_tokens": 5474 }, @@ -19143,7 +19302,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10433, + "serialized_chars": 10436, "estimated_input_tokens": 2609, "estimated_eval_tokens": 5474 }, @@ -19152,24 +19311,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three unrelated model-addition PRs were grouped by generic Hugging Face model-registration/doc changes, but they target different models and codepaths, so they are not duplicates.", - "confidence": 0.97, + "summary": "These are three separate new-model addition PRs (DeepSeek OCR, Isaac, Kimi2-6) that overlap mainly in Hugging Face boilerplate and auto-registration files, not in the same underlying change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR: each PR introduces a different model package and docs (DeepSeek OCR, Isaac, Kimi2-6), and the shared auto-registration/doc edits are boilerplate rather than the same underlying change.", + "canonical_pr_reason": "No clear canonical duplicate PR: the PRs target different models and unrelated code paths, so any single PR would be only a loose representative.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45630 is the strongest standalone candidate because it has the most discussion/review activity and the most complete model integration surface, but it is still not merge-equivalent to the others.", + "best_pr_reason": "PR #45630 is the most mature and heavily reviewed of the set, so it is the strongest pragmatic representative if one must be chosen.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different model additions (DeepSeek OCR vs Isaac) with separate model directories and docs; shared registration updates are generic boilerplate, not the same fix/change." + "reason": "Different model additions (DeepSeek OCR vs Isaac). Shared docs/auto-mapping scaffolding is boilerplate, not the same concrete bug or change." }, { "left": "pull_request:45186", "right": "pull_request:45630", "accept": false, - "reason": "Different model additions (Isaac vs Kimi2-6); despite overlapping auto-mapping/doc files, they introduce distinct model implementations and should not be merged as one PR." + "reason": "Different model additions (Isaac vs Kimi2-6) with only generic registration and docs overlap. They do not fix or implement the same underlying code path." } ] }, @@ -19199,7 +19358,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "ed8da67cb21ef2260ba97a52f108240dea1b860e1414c41db26da9f797c6830f", + "prepared_review_unit_hash": "e6b85d0c619464d82f549bff0f7f8b06dfbf7fea6705a1c0250cf2d1db649339", "component_budget": { "node_count": 2, "item_count": 2, @@ -19221,18 +19380,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two separate GitHub Actions maintenance PRs: one bumps core workflow actions for Node 24 compatibility, the other updates Docker build actions and a few related workflow actions to newer versions. They overlap in files but address different action families and different upgrade goals.", - "confidence": 0.91, + "summary": "These two PRs both update GitHub Actions in workflow files, but they target different action families and different upgrade goals: one is a Node 24 compatibility sweep for checkout/cache/upload-artifact, while the other bumps docker/build-push-action. They overlap in scope, not in the underlying change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the more clearly scoped representative of the cluster because its title and diff directly target a specific compatibility issue across many workflows; PR 42976 is a separate generic action-version refresh focused on Docker-related actions.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the best representative PR for this cluster because it has the clearest, most concrete problem statement and broad workflow impact, while still being distinct from the Docker action upgrades in PR 42976.", + "best_pr_reason": "No single PR is a true canonical duplicate; if forced to pick a representative, 42975 is the broader workflow-action compatibility upgrade, but 42976 is a separate docker action bump.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Do not merge as duplicates: PR 42975 upgrades checkout/cache/upload-artifact for Node 24 compatibility, while PR 42976 upgrades docker/build-push-action and other workflow actions to latest versions. They overlap only in general CI/workflow maintenance, not the same underlying code-path change." + "reason": "Different dependency tracks and fix intents: Node 24-compatible GitHub Actions vs docker/build-push-action v6 upgrades. Shared workflow files alone is not enough to treat them as the same change." } ] }, @@ -19274,94 +19433,94 @@ "pull_request:43451|pull_request:46411", "pull_request:45490|pull_request:46411" ], - "prepared_review_unit_hash": "e16c1d768cfbdd61b5dc731a854a114d5deec2e2fed90db57ab261baacba004e", + "prepared_review_unit_hash": "dcccff41718c60ea2ca51e82b758c569e219cda15725dffe6ef50d54dfd233a3", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 10, - "serialized_chars": 17815, - "estimated_input_tokens": 4454, - "estimated_eval_tokens": 9164 + "serialized_chars": 17822, + "estimated_input_tokens": 4456, + "estimated_eval_tokens": 9168 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 10, - "serialized_chars": 17815, - "estimated_input_tokens": 4454, - "estimated_eval_tokens": 9164 + "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 items are separate model-addition PRs for different architectures (ViT NEPA, Molmo, Molmo2, CTSM, MiniMax M2.7). The overlap is only at the template level of adding model docs/registry/tests, not the same underlying bug or change.", - "confidence": 0.98, + "summary": "These are distinct model-addition PRs across different model families; the shared docs/auto-registry edits are broad integration boilerplate, not evidence of the same underlying change. No soft edges should be merged.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No true duplicate PR exists in this cluster; if one representative is needed, 43451 is the strongest candidate because it is the most extensive and has the highest review activity, but it is still a separate model addition.", "best_issue_reason": null, - "best_pr_reason": "PR 43451 is the most substantial and actively reviewed model-addition PR, but it still represents a distinct Molmo2 integration rather than a duplicate of any other PR here.", + "best_pr_reason": "43451 is the best representative PR only in the sense of completeness and activity; it covers the broadest end-to-end model integration, but it is not equivalent to the other PRs.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different models and code paths: ViT NEPA vs Molmo. Shared auto/docs plumbing is generic model-integration work, not the same change." + "reason": "Different models and codepaths: ViT NEPA vs Molmo. Shared docs/auto-registration files are generic model wiring, not the same change." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "Different model families: ViT NEPA vs Molmo2. Similar file patterns, but no shared concrete bug or feature." + "reason": "Different model families: ViT NEPA vs Molmo2. The overlap is only in shared registry/docs updates." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "ViT NEPA and CTSM are unrelated architectures (vision-language vs time-series); only the repo integration pattern overlaps." + "reason": "Unrelated additions: ViT NEPA vision model vs CTSM time-series model. No shared bug or concrete implementation target." }, { "left": "pull_request:42978", "right": "pull_request:46411", "accept": false, - "reason": "ViT NEPA model addition is unrelated to the MiniMax demo/code drop; they do not fix or add the same thing." + "reason": "Completely different scope: ViT NEPA model integration vs a MiniMax demo/new model files. Not the same underlying change." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are related by name, but they are distinct model ports with different files and capabilities; not the same underlying change." + "reason": "Molmo and Molmo2 are related model families, but they are distinct versions with different implementations and assets; they are not the same PR-worthy code-path fix/change." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are completely different domains and implementations; shared top-level registry/doc edits are incidental." + "reason": "Molmo vs CTSM are different models in different domains; shared auto/docs edits are not enough to treat them as duplicates." }, { "left": "pull_request:43448", "right": "pull_request:46411", "accept": false, - "reason": "Molmo model integration and MiniMax demo/code are unrelated; no common code-path problem or mergeable duplicate change." + "reason": "Molmo model integration is unrelated to the MiniMax demo/original-code PR." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM address different model stacks; overlap is only in generic auto-registration/docs." + "reason": "Molmo2 and CTSM are separate model integrations with different codepaths; overlap in registry/docs files is incidental." }, { "left": "pull_request:43451", "right": "pull_request:46411", "accept": false, - "reason": "Molmo2 integration and MiniMax demo code are unrelated changes, not the same feature or bug fix." + "reason": "Molmo2 model integration is not the same as adding a MiniMax demo and model stub files." }, { "left": "pull_request:45490", "right": "pull_request:46411", "accept": false, - "reason": "CTSM model integration and a MiniMax demo file have no shared concrete code-path issue; they are separate additions." + "reason": "CTSM time-series model work is unrelated to the MiniMax demo/new model files; no shared underlying bug or mergeable change." } ] }, @@ -19398,70 +19557,70 @@ "pull_request:43665|pull_request:46380", "pull_request:44815|pull_request:46380" ], - "prepared_review_unit_hash": "5430e326dbfd546fd09438bef8ab2986e6a2900789ee99abff1b7f9229008ebf", + "prepared_review_unit_hash": "9e505281769031e3b41ae9103dfbe2b5479d9bd0ceea87c932717035255e3cc4", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 11142, - "estimated_input_tokens": 2786, - "estimated_eval_tokens": 5828 + "serialized_chars": 11145, + "estimated_input_tokens": 2787, + "estimated_eval_tokens": 5830 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 11142, - "estimated_input_tokens": 2786, - "estimated_eval_tokens": 5828 + "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 are four mostly unrelated pull requests that happen to have similar \u201cfix\u201d titles and some superficial overlap in touched areas/tests. They do not look like duplicate implementations of the same bug or change.", - "confidence": 0.89, + "summary": "These four pull requests are not the same underlying change: they target different subsystems and different failure modes, with only superficial overlap in being small fixes/tests. No duplicate cluster emerges.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46380 is the strongest canonical representative: it fixes a concrete runtime bug in generation config validation, has a clear code-path change, and includes targeted follow-up tests.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #46380 is the best overall PR in this set because it addresses a specific correctness issue in core generation behavior rather than test-only edits or an unrelated subsystem change.", + "best_pr_reason": "PR 46380 is the most substantial and concrete fix in the set, with a specific generation-config validation bug and corresponding code path update; however, it is not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Both are test-oriented and share a generic \u201cfix\u201d theme, but they target different models/code paths: Whisper tokenizer decoding vs. CLIP/ViT test edits. No same underlying bug." + "reason": "Whisper tokenizer decoding/tests versus CLIP/ViT test edits; no shared bug or code path." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decoding tests/code and FP8 dequantization/model loading are unrelated problems in different subsystems." + "reason": "Whisper tokenization change is unrelated to FP8 dequantization/core loading changes." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Whisper tokenizer decoding changes and generation watermarking validation are distinct code paths and bugs; only a broad \u201cfix\u201d label overlaps." + "reason": "Only commonality is Whisper appearing in filenames; one is tokenizer decode logic, the other is generation/watermarking plus broad test updates." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "One PR is test assertions in CLIP/ViT, the other is FP8 dequant/model-loading logic; they are not the same change or bug." + "reason": "Test-only CLIP/ViT changes do not match the FP8 dequant/model-loading bug fix." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "CLIP/ViT test-only edits do not match the generation-config runtime fix in PR #46380." + "reason": "Both are generic-looking fixes, but they address different areas: image-model tests versus generation config and many model tests." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "FP8 dequantization/padding and generation watermarking validation are different concrete bugs in different areas; they could not plausibly be merged as one fix." + "reason": "Both touch model-loading-related code, but they fix different concrete problems and would not plausibly merge into one PR." } ] }, @@ -19491,7 +19650,7 @@ "soft_pairs": [ "pull_request:43488|pull_request:43532" ], - "prepared_review_unit_hash": "68d63d27bc514066a7aa832f6bba3ec69f9236e2e33d4abf4b07a6dd2e598090", + "prepared_review_unit_hash": "a9dc98cc914a95013da69fb2e61f84edc3e75b012aae4531fc04f04778339814", "component_budget": { "node_count": 2, "item_count": 2, @@ -19513,18 +19672,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: 43488 is about setup/versioning and a test dependency change, while 43532 is about workflow/CI behavior and related utility changes. They only overlap on a test file and do not appear to fix the same underlying code-path problem.", - "confidence": 0.96, + "summary": "These two PRs are not duplicates: one is a small repo-bot/setup/versioning test change, while the other is a broader CI/workflow and testing-utilities change. The only overlap is a shared Vit test file, which appears incidental.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #43532 is the better representative PR because it is the larger, more substantive change with broader discussion and multiple affected workflow/utility files; PR #43488 is a narrow bot-format/setup tweak.", "best_issue_reason": null, - "best_pr_reason": "PR 43532 is the more substantial and self-contained change set, but it still addresses a different area than 43488 and should not be treated as the same fix.", + "best_pr_reason": "PR #43532 is the strongest standalone candidate in this cluster: it has the more meaningful code-path impact and broader change surface, whereas PR #43488 is a tiny setup/versioning test PR.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Shared vit test files are not enough here: 43488 changes packaging/version metadata and adds pytest-random-order, while 43532 rewrites CI/workflow behavior and notification utilities. Different code paths and different goals." + "reason": "Reject: the diffs target unrelated concerns. #43488 changes setup/versioning and adds a pytest dependency/test tweak; #43532 rewires CI workflows and touches testing/notification utilities. Sharing one Vit test file is not enough to indicate the same underlying fix." } ] }, @@ -19554,40 +19713,40 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "0775009ddf58f9203a2a77d244d3e358ef8f4e51bc1d8f89cccbbc19d0c57c5c", + "prepared_review_unit_hash": "edbfe99d2e579e1e25f5e5af3028aef50a7441d83e7a0c0f64389d9e0cb08299", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6080, - "estimated_input_tokens": 1520, - "estimated_eval_tokens": 3296 + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6080, - "estimated_input_tokens": 1520, - "estimated_eval_tokens": 3296 + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs add similarly named but distinct segmentation pipelines: one for promptable concept segmentation and one for promptable visual segmentation. They share some scaffolding, but they target different features/models and should not be treated as duplicates.", + "summary": "Two open PRs add different pipelines: promptable concept segmentation and promptable visual segmentation. They share scaffolding and docs patterns, but they target different concrete features and model/test sets, so they should not be merged as duplicates.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43613 is the broader and more complete addition, covering the visual segmentation pipeline with more model integrations and test coverage.", + "canonical_pr_reason": "PR 43613 is the better canonical representative because it has the broader, more integrated visual-segmentation implementation and test coverage; 43612 is a separate concept-segmentation feature, not a duplicate.", "best_issue_reason": null, - "best_pr_reason": "PR 43613 is the best representative of the cluster because it is the more substantial, end-to-end pipeline implementation.", + "best_pr_reason": "PR 43613 is the strongest overall representative of the cluster due to its wider scope and more substantial code/test surface, while still being clearly distinct from 43612.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Reject: these are different features, not the same underlying change. One adds Promptable Concept Segmentation, the other Promptable Visual Segmentation, with different docs, pipeline modules, and model/test coverage. Shared boilerplate does not make them mergeable into one PR." + "reason": "Different features and code paths: concept segmentation vs visual segmentation. Shared documentation/pipeline scaffolding is not enough to treat them as the same underlying change." } ] }, @@ -19617,7 +19776,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "de73c04e8345a45e7c0285b0fb51338c270535040863c5047fc1792ce0ed0dd3", + "prepared_review_unit_hash": "173fb11220d386ac58bff9d36b13ebb4f28debe13caef047983a77aa2d8c8c04", "component_budget": { "node_count": 2, "item_count": 2, @@ -19639,18 +19798,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs both touch `trainer.py` and add extensibility hooks, but they address different code paths: gradient-accumulation loss scaling vs. signature-based column filtering. They are related in theme, not duplicates.", + "summary": "These are two unrelated Trainer extension PRs in the same file: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers/PEFT models. They are not duplicates.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: 43651 and 43663 change different Trainer behaviors and would not be merged as one duplicate fix.", + "canonical_pr_reason": "No canonical PR: the PRs address different code paths in `trainer.py` and should not be collapsed into one duplicate group.", "best_issue_reason": null, - "best_pr_reason": "PR 43651 is the better representative of the cluster because it makes a concrete training-step behavior change with a clear API hook; 43663 is a separate hook for input-column filtering.", + "best_pr_reason": "Neither PR is a clear duplicate of the other; if a representative is needed, `#43651` is slightly more self-contained and specific, but it does not subsume the other change.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Reject: both are Trainer extensibility changes, but they fix different problems in different code paths. 43651 changes gradient-accumulation loss normalization; 43663 changes how signature columns are derived for filtering." + "reason": "Different trainer behaviors: `#43651` adds `_loss_is_scaled_for_ga` for gradient-accumulation loss scaling, while `#43663` adds `_get_signature_columns` to control column filtering. Shared file alone is not enough to treat them as the same change." } ] }, @@ -19680,7 +19839,7 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "b9723602bddeacfebe346049128f67c1783a4fe8a1c5c276403557d66b5e418e", + "prepared_review_unit_hash": "0969a0749195deea83495164a4dc502b34fc8673fb2b9f4ff8d1c906cbbf7033", "component_budget": { "node_count": 2, "item_count": 2, @@ -19702,18 +19861,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs touch the same file but address different TypeAdapter/runtime typing problems in `src/transformers/cli/serve.py`. They should stay separate; PR 43842 is the more focused standalone fix for the missing-pydantic case.", - "confidence": 0.91, + "summary": "The two PRs both touch `src/transformers/cli/serve.py`, but they fix different annotation/runtime-import problems: one is a broader postponed-annotation cleanup, the other specifically handles `TypeAdapter` when `pydantic` is unavailable. They are related but not the same change and should not be merged as duplicates.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "`pull_request:43842` is the better canonical PR because it has the clearest, most targeted fix and an explicit issue link; it addresses the concrete `TypeAdapter` import/annotation failure when `pydantic` is not installed.", "best_issue_reason": null, - "best_pr_reason": "43842 is the cleaner, more targeted fix: it explicitly guards the `TypeAdapter` import behind `TYPE_CHECKING` and quotes the annotation, matching the stated missing-pydantic NameError.", + "best_pr_reason": "`pull_request:43842` is the strongest overall PR in this cluster: it is narrowly scoped, directly states the failure mode, and cleanly fixes the runtime path that would break CLI import without `pydantic`.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": false, - "reason": "Reject: both touch `serve.py`, but they fix different concrete issues. 43656 is a broader annotation/future-import cleanup, while 43842 specifically fixes `TypeAdapter` when pydantic is absent." + "reason": "They overlap in filename and general annotation theme, but they fix different bugs: `43656` is a broader future-annotations/type-hint cleanup, while `43842` specifically makes `TypeAdapter` safe when `pydantic` is absent. They are not the same underlying change and would not be merged as one PR." } ] }, @@ -19905,334 +20064,304 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "351319817012d0d54a2c857aebe183700fdb17fa867f17c6cf597052a9dbc88b", + "prepared_review_unit_hash": "668598dbef95359f740f3c0aca511ea58cff14a546bd16a76688265692e5233b", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84896, - "estimated_input_tokens": 21224, - "estimated_eval_tokens": 42704 + "serialized_chars": 84898, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 }, "budget": { "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75092, - "estimated_input_tokens": 18773, - "estimated_eval_tokens": 37802 + "serialized_chars": 75094, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is mostly a theme cluster of standardized output-tracing refactors across many different models. Only the GPT-J/CodeGen and GPT-Neo pairs look like true duplicates; the rest are separate model-specific changes that merely share the same refactor pattern.", - "confidence": 0.85, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44722 is the best anchor for the only real duplicate subcluster: it covers the same GPT-J/CodeGen output-tracing refactor as the overlapping GPT-J PRs, with the clearest and broadest concrete code-path overlap.", - "best_issue_reason": null, - "best_pr_reason": "PR 44722 is the strongest single representative in this cluster because it reflects the shared GPT-J/CodeGen decorator migration most directly and appears to subsume the closely related GPT-J variants.", + "summary": "A shared output-tracing migration cluster centered on issue #43979. Most PRs are model-specific refactors and should not be deduped across different models; only same-model retries look like true duplicates.", + "confidence": 0.86, + "canonical_issue_reason": "Issue #43979 is the shared tracker for the standardized output-tracing migration and is explicitly linked by many of these PRs.", + "canonical_pr_reason": "PR #44066 is a strong representative of the cluster: it is explicitly tied to #43979 and makes the concrete GPT-J/CodeGen output-tracing refactor on the relevant code path.", + "best_issue_reason": "Issue #43979 is the common upstream coordination point for this migration work.", + "best_pr_reason": "PR #44066 is the most representative, well-scoped implementation in the set and sits at the center of the duplicate GPT-J/CodeGen attempts.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families; same refactor theme only." + "reason": "Different models and code paths (CVT/FNet vs GPT-J); same theme only." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "CVT/FNet output tracing vs DeBERTa output tracing are unrelated code paths." + "reason": "Different models/code paths; both reference #43979 but are not duplicates." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both target the GPT-J output-tracing refactor in the same file family; they look like overlapping versions of the same change." + "reason": "Same GPT-J output-tracing refactor on the same file; 44066 already covers the concrete GPT-J path." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "RegNet/ResNet refactors and EfficientNet refactors are separate model changes." + "reason": "Different model families; only the refactor theme matches." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet and GPT-J are unrelated model code paths." + "reason": "EfficientNet vs GPT-J/CodeGen are unrelated code paths." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen output tracing is unrelated to EfficientNet output tracing." + "reason": "Different models and implementation paths; not the same bug/change." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are unrelated models; only the refactor pattern matches." + "reason": "MobileNetV2 vs DeBERTa are unrelated model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen and MGP-STR are different code paths." + "reason": "Different models; shared migration style does not make them duplicates." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen and MPT are separate model implementations." + "reason": "GPT-J/CodeGen vs MPT are separate code paths." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Both are GPT-Neo output-tracing refactors against the same model/file path, so they appear to be the same underlying change." + "reason": "Both are GPT-Neo output-tracing refactors on the same file and same underlying code path." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo and GPT-J are different models; same umbrella refactor is not enough." + "reason": "GPT-Neo vs GPT-J/CodeGen are different model-specific refactors." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen vs GPT-Neo are distinct model code paths." + "reason": "Different models; same migration family only." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet output tracing and GPT-J/CodeGen output tracing are unrelated changes." + "reason": "ResNet vs GPT-J/CodeGen are unrelated code paths." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J touch different model implementations." + "reason": "MGP-STR vs GPT-J/CodeGen are different model changes." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are different concrete model paths." + "reason": "ResNet vs GPT-J are different concrete code paths." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are different models; not a duplicate change." + "reason": "GPT-Neo and ResNet are different models; not mergeable as one PR." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are unrelated implementations." + "reason": "Different models/code paths; only the output-tracing pattern is shared." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert are separate model-specific refactors." + "reason": "GPT-Neo vs Flaubert are separate model implementations." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert do not share the same code path." + "reason": "Different models and refactors; no shared concrete bug." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo output tracing is unrelated to GPT-J/CodeGen output tracing." + "reason": "GPT-Neo vs GPT-J/CodeGen are different model-specific changes." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are different model implementations." + "reason": "MPT vs GPT-J/CodeGen are unrelated code paths." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo and GPT-J are different models." + "reason": "GPT-Neo vs GPT-J are different model refactors." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated code paths." + "reason": "GPT-Neo vs MGP-STR are unrelated model changes." }, { - "left": "pull_request:44098", - "right": "pull_request:44101", + "left": "pull_request:44068", + "right": "pull_request:44086", "accept": false, - "reason": "ViLT and Flaubert are separate model-specific changes." + "reason": "Different models; shared decorator migration theme is too broad." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT-2 and MGP-STR are unrelated concrete implementations." + "reason": "GPT2 vs MGP-STR are unrelated code paths." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J and MGP-STR are different models." + "reason": "GPT-J vs MGP-STR are different model-specific changes." }, { "left": "pull_request:44098", - "right": "pull_request:44161", - "accept": false, - "reason": "ViLT and LongT5 are unrelated model paths." - }, - { - "left": "pull_request:44073", "right": "pull_request:44101", "accept": false, - "reason": "VisualBert and Flaubert are different models." + "reason": "ViLT vs Flaubert are different model implementations." }, { "left": "pull_request:44059", - "right": "pull_request:44071", + "right": "pull_request:44722", "accept": false, - "reason": "GPT-2 and MPT are unrelated code paths." + "reason": "GPT2 vs GPT-J/CodeGen are unrelated code paths." }, { - "left": "pull_request:44071", - "right": "pull_request:44076", + "left": "pull_request:44017", + "right": "pull_request:44722", "accept": false, - "reason": "MPT and ImageGPT are separate model refactors." + "reason": "Segformer vs GPT-J/CodeGen are different model refactors." }, { - "left": "pull_request:44074", + "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "TextNet and LongT5 are different model implementations." + "reason": "Segformer vs LongT5 are unrelated concrete changes." }, { - "left": "pull_request:44071", - "right": "pull_request:44129", - "accept": false, - "reason": "MPT and SpeechT5 are unrelated concrete changes." - }, - { - "left": "pull_request:44010", + "left": "pull_request:44017", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and VisualBert are different models." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44018", - "accept": false, - "reason": "SqueezeBert and GPT-Neo do not share the same code path." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44056", - "accept": false, - "reason": "RegNet and MPNet are separate model-specific refactors." + "reason": "Segformer vs VisualBert are different model paths." }, { - "left": "pull_request:44010", + "left": "pull_request:44068", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBert and MPT are unrelated implementations." + "reason": "GPT-Neo vs MPT are separate code paths." }, { - "left": "pull_request:43997", - "right": "pull_request:44074", + "left": "pull_request:44074", + "right": "pull_request:44101", "accept": false, - "reason": "RegNet and TextNet are different models." + "reason": "TextNet vs Flaubert are unrelated model implementations." }, { "left": "pull_request:43997", "right": "pull_request:44161", "accept": false, - "reason": "RegNet and LongT5 are unrelated changes." - }, - { - "left": "pull_request:44056", - "right": "pull_request:44161", - "accept": false, - "reason": "MPNet and LongT5 are different model paths." + "reason": "RegNet vs LongT5 are different code paths." }, { "left": "pull_request:43996", "right": "pull_request:44017", "accept": false, - "reason": "CVT/FNet refactors and Segformer refactors are separate model-specific changes." + "reason": "CVT/FNet vs Segformer are different model refactors." }, { "left": "pull_request:44010", - "right": "pull_request:44044", + "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBert and DeBERTa are unrelated models." + "reason": "SqueezeBERT vs MPT are unrelated models." }, { "left": "pull_request:44044", "right": "pull_request:44074", "accept": false, - "reason": "DeBERTa and TextNet are different code paths." + "reason": "DeBERTa vs TextNet are different concrete changes." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Same GPT-J/CodeGen output-tracing change; 44722 looks like the consolidated version of the overlapping refactor." + "reason": "Same GPT-J/CodeGen standardized output-tracing change on overlapping files; these are duplicate implementations." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both touch the same GPT-J output-tracing refactor in the same file area, so they are effectively duplicates." + "reason": "Same GPT-J output-tracing refactor; 44722 is a direct duplicate-style follow-up on the same code path." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated model changes." + "reason": "GPT-J vs SpeechT5 are different model-specific refactors." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "CVT/FNet and RegNet/ResNet are different model-specific refactors." + "reason": "Different model families; only the output-tracing migration theme overlaps." }, { "left": "pull_request:43996", "right": "pull_request:44129", "accept": false, - "reason": "CVT/FNet and SpeechT5 are unrelated code paths." + "reason": "CVT/FNet vs SpeechT5 are unrelated code paths." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and SpeechT5 are different models." + "reason": "DeBERTa vs SpeechT5 are different model changes." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and VITS are separate model implementations." + "reason": "DeBERTa vs ViTS are different concrete refactors." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded and conservative overall. The accepted soft edges stay within the same model/file families with clear diff overlap (GPT-J/CodeGen and GPT-Neo), while cross-family pairs are appropriately rejected. The summary does not overstate broader duplication beyond what the packet supports." + "feedback": "The summary is grounded in the packet: issue #43979 is the shared tracker for this migration set, and the chosen canonical PR (#44066) is well supported by its explicit link and overlapping GPT-J/CodeGen scope. The soft-edge calls are conservative overall, especially by only accepting pairs that share the same concrete model/code path. " }, "evaluator_used": true, "retried": false, @@ -20270,14 +20399,14 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "52d22dd3050bb234fe5a8981fa793614fa7b9e6244a6f815e5f50b8f63728e30", + "prepared_review_unit_hash": "fc34f9b0c613b961f0848339e18ca2f704e9789f7c607c660cbdf75d5611cd5b", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84896, - "estimated_input_tokens": 21224, - "estimated_eval_tokens": 42704 + "serialized_chars": 84898, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 }, "budget": { "node_count": 8, @@ -20292,48 +20421,48 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All items are model-specific PRs that apply the same output-tracing refactor pattern to different code paths, not duplicates of one another. The cluster is best represented by the SpeechT5 PR because it is the most concrete and has an explicit issue target.", - "confidence": 0.92, + "summary": "These PRs are all parallel model-specific refactors to standardized output tracing/capture_outputs, but they touch different architectures and code paths. They look thematically similar, not like duplicates or one mergeable change.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44129 is the strongest representative: it has an explicit issue target, a clear standardized output-capture refactor, and enough context to describe the cluster\u2019s common theme.", + "canonical_pr_reason": "PR 44073 is a clean representative of the cluster: it clearly expresses the shared standardized output-tracing refactor, and the diff directly shows the decorator/capture_outputs migration pattern.", "best_issue_reason": null, - "best_pr_reason": "44129 is the most suitable standalone representative because it is self-contained, clearly scoped to one model, and anchored to a linked issue, whereas the others are similar but equally model-specific refactors.", + "best_pr_reason": "PR 44073 is the best global representative because its title and diff most clearly capture the common change pattern across the cluster without being unusually narrow or issue-specific.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Both are output-tracing refactors, but they touch different model implementations (FocalNet vs ImageGPT) and different code paths; not the same underlying change." + "reason": "Both are standardized-output refactors, but they target different models (FocalNet vs ImageGPT) and different implementation details. Same theme, not the same underlying bug/change." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Same refactor pattern, but VisualBert and TextNet are separate model-specific implementations; these are parallel changes, not duplicate PRs." + "reason": "Both migrate output tracing, but they are separate model implementations (VisualBert vs TextNet) with different forward signatures and test coverage. Not one concrete change." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet and VisualBert are distinct model code paths; the similarity is architectural, not a shared concrete bug or change." + "reason": "These are model-specific output-tracing refactors for different architectures (MPNet vs VisualBert). They share a pattern, but not a mergeable single code-path fix." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "Both are standardized-output refactors, but they are separate model migrations in different files with no shared concrete fix to merge into one PR." + "reason": "MPNet and TextNet are unrelated model code paths; the similarity is only the generic migration to capture_outputs. Not duplicates." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated model implementations; the PRs share a refactor theme only, not the same underlying issue." + "reason": "RegNet and SqueezeBert are different models with separate forward/output plumbing. The shared decorator migration is too broad to count as the same change." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 modify different model-specific forward/output plumbing, so they are not duplicate fixes even though both standardize output capture." + "reason": "SqueezeBert and SpeechT5 are different subsystems and the diffs indicate different output-tracing adaptations, not one concrete bug or feature merge." } ] }, @@ -20412,274 +20541,274 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "12ae77e0d182a15f237b4e8ee98ea2c820851f0c9d062c0473b053e6b1f68a91", + "prepared_review_unit_hash": "8615dc41d7f316d63538a8d118b78531311e814ca41ecb0d0abc8af719b6a986", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 39, - "serialized_chars": 39788, - "estimated_input_tokens": 9947, - "estimated_eval_tokens": 20150 + "serialized_chars": 39790, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 39, - "serialized_chars": 39788, - "estimated_input_tokens": 9947, - "estimated_eval_tokens": 20150 + "serialized_chars": 39790, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is a set of similar-looking but model-specific output-tracing refactors across unrelated code paths. They share a theme and some naming, but not the same underlying bug/change, so none of the soft edges should be merged as duplicates.", - "confidence": 0.93, + "summary": "This cluster contains a series of independent, model-specific output-tracing refactors. They share a common pattern (`can_return_tuple` / tracing helpers), but they target different model files and code paths, so they should not be treated as duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44025 is a good cluster representative: it clearly shows the core refactor pattern (adding `can_return_tuple` and simplifying forward output handling) and spans two related files, making it a strong exemplar of the shared theme.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 44025 is the best overall representative for the cluster because its diff is straightforward, clearly documents the refactor intent, and captures the common output-tracing change pattern without being overly specialized to a single code path.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model families and different files; same refactor theme only, not the same concrete bug or change." + "reason": "Different model families (`timm_backbone` vs `upernet`); same refactor theme only." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Both are output-tracing refactors, but they target different architectures and code paths." + "reason": "`timm_backbone` and `vision_text_dual_encoder` are unrelated code paths despite similar tracing changes." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Unrelated model implementations; similar decorator/import cleanup is not enough to treat as one change." + "reason": "Separate models (`timm_backbone` vs `univnet`); not the same underlying bug or change." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "Timm backbone vs Mamba/Falcon Mamba: different model behavior and separate forward paths." - }, - { - "left": "pull_request:43998", - "right": "pull_request:44004", - "accept": false, - "reason": "CodeGen\u2019s forward/cache handling is a different change from timm_backbone output tracing." + "reason": "`timm_backbone` and `mamba`/`falcon_mamba` are distinct implementations; no shared concrete fix." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "These are distinct model stacks; similarity is limited to the generic output-tracing refactor pattern." + "reason": "Different model code paths (`timm_backbone` vs `rwkv`); only the refactor style overlaps." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "MobileNetV1 and UnivNet are unrelated models; not the same underlying fix." + "reason": "`mobilenet_v1` and `univnet` are unrelated models; same output-tracing motif only." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different model code paths and different output types; only broad refactor overlap." + "reason": "Different model implementations (`mobilenet_v1` vs `upernet`), so not mergeable as one PR." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "MobileNetV1 vs Depth Anything: same repo-wide refactor style, but not a shared bug/change." + "reason": "`mobilenet_v1` and `depth_anything` touch different forward paths and files." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Separate model wrappers with different forward signatures and paths." + "reason": "Separate models (`mobilenet_v1` vs `vision_encoder_decoder`); no common concrete bug." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobileNetV1 and Speech Encoder-Decoder are unrelated; no shared concrete defect." + "reason": "`mobilenet_v1` and `speech_encoder_decoder` are unrelated model wrappers." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Vision-Text Dual Encoder and UnivNet are different model families; similarity is only in the refactor pattern." + "reason": "`vision_text_dual_encoder` and `univnet` have different runtime behavior and code paths." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different output-tracing changes in distinct models, not one mergeable fix." + "reason": "Different models (`vision_text_dual_encoder` vs `upernet`); only the tracing refactor is common." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Vision-Text Dual Encoder and Mamba/Falcon Mamba touch different code paths and APIs." + "reason": "`vision_text_dual_encoder` and `mamba` are not the same underlying change." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "These are separate model implementations; the common theme is too broad to duplicate." + "reason": "`vision_text_dual_encoder` and `depth_anything` are separate model-specific refactors." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are unrelated; same cleanup style does not imply same underlying issue." + "reason": "`univnet` and `upernet` are unrelated models with independent forward implementations." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Different model stacks and semantics; not plausibly one combined PR." + "reason": "`univnet` and `mamba`/`falcon_mamba` do not fix the same concrete issue." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "Depth Anything vs UnivNet: no shared code path or concrete bug." + "reason": "Different models (`univnet` vs `depth_anything`); same refactor pattern only." }, { "left": "pull_request:44001", "right": "pull_request:44026", "accept": false, - "reason": "Unrelated model wrappers; similar decorator changes only." + "reason": "`univnet` and `vision_encoder_decoder` are distinct code paths." }, { "left": "pull_request:44002", - "right": "pull_request:44003", + "right": "pull_request:44029", "accept": false, - "reason": "UperNet and Mamba/Falcon Mamba are different subsystems with different output handling." + "reason": "`upernet` and `rwkv` are unrelated model implementations." }, { "left": "pull_request:44002", - "right": "pull_request:44004", + "right": "pull_request:44003", "accept": false, - "reason": "UperNet segmentation output refactor is not the same as CodeGen cache/output typing changes." + "reason": "`upernet` and `mamba`/`falcon_mamba` do not share the same bug or patch." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba/Falcon Mamba and CodeGen have different forward behavior and problem scope." + "reason": "`mamba`/`falcon_mamba` vs `codegen`: different model subsystems and code paths." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Timm backbone tracing cleanup and CodeGen output refactor are unrelated concrete changes." + "reason": "`timm_backbone` and `codegen` are unrelated models; similarity is only thematic." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet and CodeGen do not share the same code path or defect." + "reason": "`univnet` and `codegen` target different forward logic and cannot be merged as one fix." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "Different model families and different return semantics." + "reason": "`upernet` and `codegen` are separate model changes." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "`vision_text_dual_encoder` and `codegen` are different concrete code paths." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Depth Anything and Vision/Encoder-Decoder are distinct model wrappers; same refactor theme only." + "reason": "`depth_anything` and `vision_encoder_decoder` are different models; same refactor theme only." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Depth Anything and Speech Encoder-Decoder are not the same underlying bug/change." + "reason": "`depth_anything` and `speech_encoder_decoder` are independent wrappers with separate fixes." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Depth estimation vs SuperPoint keypoint model: unrelated implementation paths." + "reason": "`depth_anything` and `superpoint` are unrelated model implementations." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Depth Anything and Mamba/Falcon Mamba are different subsystems; no shared fix." + "reason": "`depth_anything` and `rwkv` touch different model stacks and code paths." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Depth Anything and DPR touch different model behaviors and output contracts." + "reason": "`depth_anything` and `dpr` are distinct model-specific refactors." }, { "left": "pull_request:44026", - "right": "pull_request:44027", + "right": "pull_request:44029", "accept": false, - "reason": "Vision Encoder-Decoder and Speech Encoder-Decoder are separate wrappers; similarity is broad architectural pattern only." + "reason": "`vision_encoder_decoder` and `rwkv` are unrelated models." }, { "left": "pull_request:44026", - "right": "pull_request:44028", + "right": "pull_request:44027", "accept": false, - "reason": "Different tasks and code paths; not the same concrete issue." + "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are different wrappers, but not the same change." }, { "left": "pull_request:44026", - "right": "pull_request:44029", + "right": "pull_request:44028", "accept": false, - "reason": "Distinct model families; no evidence they fix the same bug." + "reason": "`vision_encoder_decoder` and `superpoint` are separate code paths." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Vision/Speech encoder-decoder wrappers are unrelated to DPR's output handling." + "reason": "`vision_encoder_decoder` and `dpr` fix different model implementations." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Speech Encoder-Decoder and SuperPoint are unrelated models with different forward contracts." + "reason": "`speech_encoder_decoder` and `superpoint` are unrelated model files." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Different model families and output mechanisms." + "reason": "`speech_encoder_decoder` and `rwkv` do not share a concrete underlying bug." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Speech Encoder-Decoder and DPR are separate code paths; not a duplicate." + "reason": "`speech_encoder_decoder` and `dpr` are distinct model-specific changes." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and Mamba/Falcon Mamba are unrelated; same refactor motif is insufficient." + "reason": "`superpoint` and `rwkv` are different model paths; only the refactor pattern matches." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "SuperPoint keypoint model vs DPR retrieval model: different concrete changes." + "reason": "`superpoint` and `dpr` are unrelated changes." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "Mamba/Falcon Mamba and DPR do not share the same underlying bug or change." + "reason": "`rwkv` and `dpr` are separate model implementations, not one mergeable fix." } ] }, @@ -20709,7 +20838,7 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "15e041515752d0ca8f0dc56774432caae3973f9bc02c45f63f795910be52f14b", + "prepared_review_unit_hash": "1318a490434d4a6d3478457ae12151b1dce8b8d077582f1e36871f513196779a", "component_budget": { "node_count": 2, "item_count": 2, @@ -20731,18 +20860,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two unrelated model-addition PRs: one adds PenguinVL, the other adds Voxtral TTS support. They share only broad Transformers integration files and docs plumbing, not the same underlying feature or bug.", + "summary": "These two PRs add different model integrations: PenguinVL vs Voxtral TTS. They share some boilerplate auto-registration/docs paths, but they do not fix the same underlying code-path or belong in one merged change.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45401 is the more self-contained representative: it targets a specific tracked issue, has a focused scope around Voxtral TTS support, and includes the full implementation/testing stack for that model.", + "canonical_pr_reason": "No canonical PR as a duplicate representative: the PRs target different models and different feature sets, so neither subsumes the other.", "best_issue_reason": null, - "best_pr_reason": "PR #45401 is the best cluster representative because it is the more explicit, issue-backed feature addition and is easier to interpret as the cluster's exemplar model-support PR.", + "best_pr_reason": "PR #45401 is the more self-contained candidate because it has an explicit issue target and a narrower, clearly scoped model-support addition.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different models and code paths: PenguinVL vs Voxtral TTS. Shared files are only common Transformers registration/documentation surfaces, which is too broad to indicate the same underlying change." + "reason": "Reject: different model implementations (PenguinVL vs Voxtral TTS). Shared files are generic registration/docs plumbing, not evidence of the same bug or change." } ] }, @@ -20776,7 +20905,7 @@ "pull_request:45363|pull_request:46339", "pull_request:45666|pull_request:46339" ], - "prepared_review_unit_hash": "f290302ae2b505321b825f2508a04d3f5f0eb1744efeafc904120166a57c02b4", + "prepared_review_unit_hash": "6a9cdc14bb680094c56e09404250dc42fa71c9ca697a4aa20c821ce5320f6855", "component_budget": { "node_count": 4, "item_count": 4, @@ -20798,40 +20927,37 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a chain of related kernel-fusion PRs. The latest PR, 46339, looks like the best canonical PR for the concrete n-to-1 KernelConfig fusion feature, while 44979 is a separate lower-level module-fusion API and should not be merged with the later KernelConfig work.", - "confidence": 0.87, + "summary": "All four PRs are on the same kernel-fusion feature line, but they are sequential expansions of the API rather than the same concrete change. The latest PR is the broadest and most representative of the final direction.", + "confidence": 0.66, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46339 is the most complete and latest version of the same n-to-1 KernelConfig fusion feature, adding the broader integration/simplification work that builds on the earlier PRs.", + "canonical_pr_reason": "PR 46339 is the most complete and recent kernel-fusion implementation: it simplifies and extends the KernelConfig-based n-to-1 fusion path, touching the core loader, hub kernel integration, modeling utils, and tests.", "best_issue_reason": null, - "best_pr_reason": "46339 is the best standalone representative because it appears to subsume the earlier KernelConfig fusion iterations and is the most polished implementation of the shared code-path change.", + "best_pr_reason": "PR 46339 best represents the cluster because it subsumes the earlier KernelConfig fusion work and includes the widest implementation and test coverage.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "44979 introduces a foundational Module Fusion API, while 45363 adds a specific KernelConfig-based n-to-1 kernel fusion integration; related, but not the same concrete change." + "reason": "Related feature area, but not the same change: 44979 introduces a lower-level Module Fusion API, while 45363 builds a separate KernelConfig-based n-to-1 fusion integration on top of it." }, { "left": "pull_request:45363", "right": "pull_request:46339", - "accept": true, - "reason": "Both target the same concrete n-to-1 kernel fusion via KernelConfig path, touch overlapping integration files, and 46339 reads like a follow-up simplification/extension of 45363." + "accept": false, + "reason": "Same general feature family, but 46339 is a later extension/simplification of the KernelConfig fusion path with additional loader and integration changes, so it is not the same concrete PR." }, { "left": "pull_request:45666", "right": "pull_request:46339", - "accept": true, - "reason": "These are the same KernelConfig-based n-to-1 fusion feature area, with 46339 appearing to extend and simplify the exact mechanism introduced in 45666." + "accept": false, + "reason": "Both work on n-to-1 kernel fusion, but 45666 is an intermediate extension and 46339 further refactors/simplifies the implementation; they are successive iterations, not a single duplicate change." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is well grounded in the packet: 44979 is clearly the separate Module Fusion API PR, while 45363/45666/46339 all center on KernelConfig-based n-to-1 kernel fusion and share overlapping files/titles. The soft-edge verdicts are conservative and do not overstate the evidence." - }, - "evaluator_used": true, + "evaluator_result": null, + "evaluator_used": false, "retried": false, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -20854,7 +20980,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "61cbe46094d9d2184a0bec3bb0f07407b7bd7d745ab33d022c2c7c9d62954757", + "prepared_review_unit_hash": "2a14e027e89c2fceb643be078cdceb96160df2bb98b769722811c0eae3c0c6bf", "component_budget": { "node_count": 2, "item_count": 2, @@ -20876,18 +21002,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests target the same Mistral regex patch bug in `tokenization_utils_tokenizers.py`; the later PR adds tests and is the stronger canonical candidate.", + "summary": "Both pull requests target the same Mistral regex patch bug in the same file, with the same issue target and nearly identical code change; 45317 is the more complete version because it also adds tests.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No issue node is present. Between the PRs, #45317 is the better canonical PR because it fixes the same AttributeError but also adds test coverage, making it the more complete change.", + "canonical_pr_reason": "PR 45317 is the stronger canonical PR because it is the later, more complete fix for the same AttributeError path and includes test coverage in addition to the code change.", "best_issue_reason": null, - "best_pr_reason": "PR #45317 is the best overall representative: it addresses the same concrete `_patch_mistral_regex` AttributeError and includes validation tests, while #45086 is a narrower earlier version of the same fix.", + "best_pr_reason": "PR 45317 is the best representative of the cluster because it fixes the concrete bug directly and adds tests, making it the more self-contained and merge-ready change.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same bug, same code path, same file, and nearly identical fix; #45317 is essentially an expanded version of #45086 with tests, so they are duplicate/mergeable." + "reason": "Same underlying AttributeError in `_patch_mistral_regex`, same affected file and issue target, and the diffs change the same code path in a mergeable way." } ] }, @@ -20917,12 +21043,12 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "25ca93cd0c9d3a9f49c7d80a6c84dcda1e42b265617ee910f4bbbe9c8bfb6254", + "prepared_review_unit_hash": "c314a3b8b7feac8e08f4ea514b2d1f77d5631ecfb8ec438ccfeb10df4794831b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6066, + "serialized_chars": 6065, "estimated_input_tokens": 1517, "estimated_eval_tokens": 3290 }, @@ -20930,7 +21056,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6066, + "serialized_chars": 6065, "estimated_input_tokens": 1517, "estimated_eval_tokens": 3290 }, @@ -20939,18 +21065,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related NemotronH model work items, but they are not the same fix: one is a narrower refactor/conversion-mapping update, while the other splits NemotronH into dense/sparse variants with broader API, docs, and test changes.", + "summary": "Two NemotronH-related pull requests, but they address different scopes: one is an inheritance/refactor toward GraniteMoeHybrid, the other is a broader model split introducing NemotronHDense/NemotronHSparse and docs. They are related, not duplicates.", "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45569 is the better canonical PR because it is the broader, more complete model-structure change: it introduces the dense/sparse variants, updates auto/model registrations, adds docs, and includes tests. PR 45115 is narrower and mainly refactors inheritance/conversion mappings.", + "canonical_pr_reason": "PR #45569 is the better canonical PR because it appears to deliver the fuller, user-facing model packaging change (new dense/sparse variants, auto-mapping, docs, and tests) rather than a narrow refactor.", "best_issue_reason": null, - "best_pr_reason": "PR 45569 is the best representative of the cluster since it addresses the main architectural change and touches the full user-facing surface. PR 45115 looks like a partial supporting refactor rather than the primary end-state implementation.", + "best_pr_reason": "PR #45569 is the strongest representative of the cluster: it has the broadest and most complete implementation of the NemotronH workstream and is more likely to stand alone as the primary change. PR #45115 is a narrower preparatory refactor.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "They overlap on NemotronH internals, but 45115 is about inheritance/conversion mapping cleanup for NemotronH, while 45569 is a larger split into NemotronH dense/sparse variants with docs, auto-mapping, and tests. Not the same concrete change and not safely mergeable into one PR." + "reason": "They share NemotronH files, but the changes are not the same underlying fix. #45115 is a targeted refactor/inheritance mapping toward GraniteMoeHybrid, while #45569 splits NemotronH into dense/sparse models with docs and tests. Different goals, not mergeable as one PR." } ] }, @@ -20980,7 +21106,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "fbef38615b0bdc887bdabbda27a14d8ef4bbd53418cdad7f8b46a2de1920bc78", + "prepared_review_unit_hash": "4686420d3bcd26cfbc83bcdd77248f001ae3a8c2e34620aa0e20bc5bba3af36b", "component_budget": { "node_count": 2, "item_count": 2, @@ -21002,18 +21128,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two related but distinct PRs for heterogeneous model support: one adds per-layer config infrastructure, the other adds heterogeneous modeling/cache changes. They overlap on shared config plumbing but are not the same concrete change.", - "confidence": 0.84, + "summary": "The two PRs are related pieces of the same heterogeneous-model effort, but they are not duplicates: one adds per-layer config support, while the other builds out broader heterogeneous modeling/cache/masking support on top of that.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45332 is the broader umbrella change and includes the modeling-side integration; #45333 is a narrower config-only foundation. If picking one representative, #45332 is the better canonical PR, but they are not duplicates.", + "canonical_pr_reason": "PR 45332 is the broader umbrella change and is the better canonical representative of the heterogeneous-model feature work; PR 45333 looks like a narrower config-layer prerequisite rather than the same end-to-end change.", "best_issue_reason": null, - "best_pr_reason": "#45332 is the best representative of the cluster because it spans the larger end-to-end heterogeneous model support surface, while #45333 is a prerequisite/config subset.", + "best_pr_reason": "PR 45332 is the best standalone PR because it covers the full code-path change set (config, modeling, cache, masking, tests) rather than only the configuration layer.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "They share heterogeneous-config plumbing, but one is a broad modeling/cache/masking PR and the other is a config-only PR. Different concrete code-paths and not safe to treat as the same change." + "reason": "Related but not the same concrete change: 45333 is per-layer configuration support, while 45332 is a broader heterogeneous model implementation that depends on and extends that foundation. They would not be a single mergeable PR without collapsing distinct scopes." } ] }, @@ -21043,7 +21169,7 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "19e3bcce600d7e3f35500917b22fadb3a3354b17d5227fe61d35da67962e0bee", + "prepared_review_unit_hash": "3e5a1c3392d12dabdb6584f5aec738ab240e9a0dd743ef0545bc4b5e5f74f111", "component_budget": { "node_count": 2, "item_count": 2, @@ -21065,18 +21191,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are typing-related PRs, but they address different scopes: #45415 is a broad repo-wide type-checking sweep, while #45425 is a more targeted typing/refactor PR centered on modeling_utils and shared typing protocols. They are not duplicates.", - "confidence": 0.94, + "summary": "Both items are typing-related PRs, but they are not the same change: one is a broad type-checking sweep across many Transformer files, while the other adds a smaller set of typing protocols/helpers around modeling_utils/PEFT integration. They overlap in theme and a few files, but not in concrete code-path or patch scope.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "No clear duplicate PR in this cluster. If one must be selected, #45425 is the better representative because it is a focused, self-contained typing change, whereas #45415 is a wide-ranging type-checking cleanup across many files.", + "canonical_pr_reason": "PR 45415 is the broader umbrella change: it touches many Transformer modules plus pyproject typing settings, so it best represents the cluster\u2019s main typing-effort theme.", "best_issue_reason": null, - "best_pr_reason": "#45425 is the best representative of the cluster because it is narrowly scoped and technically coherent; #45415 is broader and mixes many unrelated typing edits, making it a poorer canonical stand-in.", + "best_pr_reason": "PR 45415 is the strongest representative of the cluster because it is the most comprehensive and central typing cleanup among the two.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Shared typing-related files are not enough here: #45415 is a broad typing sweep, while #45425 introduces specific typing protocols and annotations. They do not look like the same underlying fix and would not plausibly merge into one PR." + "reason": "Same general typing topic, but different concrete changes and scopes: 45415 is a wide type-checking sweep, while 45425 adds typing helpers/protocols for specific integrations. They do not look mergeable as one PR for the same bug/change." } ] }, @@ -21106,7 +21232,7 @@ "soft_pairs": [ "pull_request:45513|pull_request:46084" ], - "prepared_review_unit_hash": "8d6dd30b13f1bae327f83e28b182395920e3236527105fc96895cef4ab218543", + "prepared_review_unit_hash": "f9f3d5987e143fd3a815c2f7adce6a32b1e0d83e9310d497500e32b09c388a65", "component_budget": { "node_count": 2, "item_count": 2, @@ -21128,18 +21254,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address cache-enabled forward paths with seq_len > 1, but they target different model families and distinct code paths, so they do not look like the same underlying fix.", - "confidence": 0.95, + "summary": "The two pull requests are about different models and different cache bugs: PR #45513 fixes multi-token cached forward behavior in Qwen3.5/related linear-attention code, while PR #46084 fixes use_cache handling in Mamba2 when seq_len > 1. They are not the same underlying change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46084 is the cleaner representative: it is narrowly scoped to Mamba2 cache handling and explicitly tied to issue #46032.", + "canonical_pr_reason": "PR #45513 is the broader, more representative cache-fix PR in this cluster, spanning multiple Qwen-family model files and a shared multi-token cached-forward bug pattern.", "best_issue_reason": null, - "best_pr_reason": "PR 46084 is the best single PR to represent this cluster because it is tightly scoped, issue-linked, and more clearly describes one concrete bug fix.", + "best_pr_reason": "PR #45513 is the best representative artifact because it addresses a concrete cache-forward bug across several related model implementations and is more central to the cluster than the unrelated Mamba2 fix.", "soft_edge_verdicts": [ { "left": "pull_request:45513", "right": "pull_request:46084", "accept": false, - "reason": "Different architectures and different implementations: Qwen/OLMo linear-attention cached forward vs Mamba2 CUDA-kernel cache handling. Same broad theme, not the same concrete bug." + "reason": "Different model families and different code paths: Qwen3.5 linear-attention cached forward vs. Mamba2 use_cache with seq_len > 1. They do not look mergeable into one PR." } ] }, @@ -21169,12 +21295,12 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "6c3cb2761dce33a74873dcb9c6035c69d7d54407de6a5a7aa75e73798fdcb76f", + "prepared_review_unit_hash": "9955ab9939942577b3c46fb2364e3fbd50069f6a83de3792590986cf81e2e5f6", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4507, + "serialized_chars": 4508, "estimated_input_tokens": 1127, "estimated_eval_tokens": 2510 }, @@ -21182,7 +21308,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4507, + "serialized_chars": 4508, "estimated_input_tokens": 1127, "estimated_eval_tokens": 2510 }, @@ -21191,18 +21317,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related by expert-parallel loading infrastructure, but they fix different bugs: PR 45548 is about preserving EP/DeepSpeed ZeRO-3 loading metadata through accelerate launch, while PR 45662 targets expert parameters being overwritten during FSDP2 loading. Not a duplicate pair.", - "confidence": 0.92, + "summary": "The two PRs are related through expert-parallel loading, but they fix different bugs and touch different code paths. PR 45548 is about EP + DeepSpeed ZeRO-3 loading and config serialization; PR 45662 is about EP + FSDP2 experts being overwritten during rank-0 broadcast.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45662 is the better cluster representative because it addresses a concrete expert-parameter corruption bug and includes the broader set of loading-path changes around that failure mode.", + "canonical_pr_reason": "PR 45662 is the more concrete and representative fix: it targets a specific expert-overwrite loading bug in the core parameter-loading path, with coordinated changes across loading and MoE integration. PR 45548 addresses a separate EP/ZeRO-3 accelerate-launch issue.", "best_issue_reason": null, - "best_pr_reason": "PR 45662 is the strongest overall representative: its title and diff describe a specific, high-signal EP loading bug with direct code-path fixes, whereas 45548 is narrower and centered on config serialization/launch behavior.", + "best_pr_reason": "PR 45662 is the best representative of the cluster because it fixes a clearly defined code-path bug with broader supporting changes, while PR 45548 is narrower and concerns a different loading/configuration problem.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Shared expert-parallel context and one common file are not enough; 45548 fixes config serialization/launch handling for ZeRO-3, while 45662 fixes expert tensors being overwritten in FSDP2 loading. Different underlying bugs and not plausibly one merged PR." + "reason": "They share the EP theme, but the bugs are different: 45548 handles EP + DeepSpeed ZeRO-3 loading/config serialization, while 45662 fixes expert tensors being overwritten during FSDP2/rank-0 broadcast. Not the same underlying change." } ] }, @@ -21232,7 +21358,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "b401dbd0659916b5294fc7147db4101cd24cf9e6aa07295e15cf0f3377db2caf", + "prepared_review_unit_hash": "6d47cb1f0682f7a5a3a1267460846669b04a645cb109ef1ef7dd9a9089f431e3", "component_budget": { "node_count": 2, "item_count": 2, @@ -21254,18 +21380,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs address the same xdist collision problem around captured_info output in testing_utils; PR 45645 appears to be the broader, more complete version with CI log aggregation added.", + "summary": "Both PRs address the same xdist-related captured_info handling in `testing_utils.py`, but #45645 is a broader follow-up that also changes CI log collection and notification handling. They are related, yet not close enough to treat as the same PR change.", "confidence": 0.79, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the better canonical PR because it covers the core xdist-safe output fix plus the downstream CI/workflow handling needed to read multiple captured_info files.", + "canonical_pr_reason": "#45645 is the more complete representative: it includes the core xdist-safe `captured_info` fix plus the CI-side log aggregation changes around that artifact.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 is the strongest representative: it addresses the same underlying captured_info xdist collision and also preserves/debug-prints the resulting artifacts in CI.", + "best_pr_reason": "#45645 is the strongest overall fix in the cluster because it covers the main bug and the surrounding CI behavior needed to preserve and collect the split logs.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", - "accept": true, - "reason": "Same underlying bug: xdist workers colliding on captured_info output for patched testing logs. PR 45645 extends the same fix into CI artifact collection, so these are plausibly one combined change." + "accept": false, + "reason": "They overlap on the xdist-safe debug-log fix, but #45645 adds additional workflow and notification changes, so they are not the same underlying change and would not cleanly merge as one PR." } ] }, @@ -21297,46 +21423,46 @@ "pull_request:45670|pull_request:46320", "pull_request:46320|pull_request:46401" ], - "prepared_review_unit_hash": "7307da740cca9642070b10d873f943f0e2e69e50b2858d5ec6f3993400c75195", + "prepared_review_unit_hash": "f32ab67a40c856f8d2f1befbb646932e3acdda3502bedc6d1fe53d378ccd8173", "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": "The cluster is mostly about model-specific attn_implementation dispatch support, plus one unrelated AutoModel registration nit. The RecurrentGemma and BLIP-2 PRs are similar in theme but target different code paths and should not be treated as duplicates.", - "confidence": 0.95, + "summary": "These PRs are related only at a theme level (model support / attention dispatch), but they fix different models and different code paths, so they should not be merged as duplicates.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46320 is the best representative of the cluster\u2019s main theme: a focused model-specific attn_implementation dispatch implementation for RecurrentGemma, with supporting conversion and tests.", + "canonical_pr_reason": "No single canonical PR: the items are separate changes for different model families and different behaviors, not the same underlying fix.", "best_issue_reason": null, - "best_pr_reason": "PR 46320 is the strongest canonical PR because it is the clearest, most self-contained example of the attention-dispatch feature work in this cluster.", + "best_pr_reason": "PR 46401 is the strongest standalone change: it adds SDPA dispatch support across BLIP-2/InstructBLIP/InstructBLIPVideo and includes tests, making it the most concrete and substantial PR in the set.", "soft_edge_verdicts": [ { "left": "pull_request:45670", "right": "pull_request:46320", "accept": false, - "reason": "Different changes: AutoModel mapping for glmasr versus RecurrentGemma attention-dispatch support. They do not fix the same code-path problem." + "reason": "These are unrelated changes: 45670 only adjusts AutoModel registration for glmasr, while 46320 adds attn_implementation support for RecurrentGemma." }, { "left": "pull_request:46320", "right": "pull_request:46401", "accept": false, - "reason": "Both add attn_implementation dispatch, but for different model families and different implementations. They are similar feature work, not the same concrete bug/change, and would not naturally merge into one PR." + "reason": "Both mention attn_implementation, but they target different model families (RecurrentGemma vs BLIP-2/InstructBLIP) and different code paths, so they are not the same underlying bug/change." } ] }, @@ -21366,7 +21492,7 @@ "soft_pairs": [ "pull_request:45720|pull_request:46259" ], - "prepared_review_unit_hash": "b326c2c80e43af865f995d624df965cc2cce76de41b05bb9696d24425e0c1268", + "prepared_review_unit_hash": "2a1f4dad7f082fa7bd6dfc2256eff4933a4699b3b583acc1f4780e77a2523c62", "component_budget": { "node_count": 2, "item_count": 2, @@ -21388,18 +21514,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two unrelated open PRs were clustered only by soft similarity. PR #45720 implements TRF003 tuple/return_dict handling across model forwards, while PR #46259 adds a different modeling rule about do_* flags in image/video processors. They do not look like the same underlying change.", + "summary": "These two PRs are not duplicates: #45720 implements TRF003-style forward/return_dict tuple handling across several model forwards, while #46259 adds a different linter rule (TRF018/TRF016) for image/video processor boolean flags and updates unrelated model processors. The overlap in some model filenames is incidental to broad repo-wide cleanup work.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR as a duplicate target: the PRs address different lint rules and different code paths, so they should not be merged or treated as one change.", + "canonical_pr_reason": "No single canonical PR for this cluster: the PRs target different rules and different code paths, so they should not be merged as one change.", "best_issue_reason": null, - "best_pr_reason": "PR #45720 is the cleaner representative of the cluster if one must be chosen, because it is a focused code-path fix across model forwards; PR #46259 is a broader, different rule addition spanning processors and docs.", + "best_pr_reason": "#45720 is the better representative only in the sense that it is the more concrete code-path change; #46259 is mostly a separate linter-rule/docs rollout. Even so, they are not duplicates.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "Different underlying tasks: tuple/return_dict forward cleanup vs do_* flag consumption in image/video processors. Shared filenames are incidental and not evidence of the same fix." + "reason": "Different underlying changes: #45720 patches model forward signatures/tuple-return behavior for TRF003, while #46259 introduces a separate processor-flag consumption rule and associated edits. Shared files are broad-model overlap, not the same bug or change." } ] }, @@ -21495,7 +21621,7 @@ "soft_pairs": [ "pull_request:45932|pull_request:46100" ], - "prepared_review_unit_hash": "701eba2c86a20efc1e8a909c4ba2bc1bf500df8228552da7ef88b42407d7c6ab", + "prepared_review_unit_hash": "6789b4734ef1f5eb556c9269da7ccbd818681d8ea14195de3bdc51dcd99bcf68", "component_budget": { "node_count": 2, "item_count": 2, @@ -21517,18 +21643,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates. PR #45932 is a broad model/config/workflow change, while PR #46100 is a tiny test-only change that intentionally adds failing assertions and targets a different code path. They should not be clustered as the same underlying change.", - "confidence": 0.98, + "summary": "These two PRs are both non-production \u201cdo not merge\u201d changes, but they do not fix the same underlying problem: one is a mixed CI/model/config tweak PR, the other is a test-file migration that intentionally introduces failing assertions. I would not cluster them as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45932 is the better canonical PR because it contains the substantive code changes across model implementations and workflow/secrets plumbing, whereas PR #46100 is just a test-migration stub with deliberate failures.", + "canonical_pr_reason": "PR #45932 is the better representative of the cluster only in the loose sense that it contains the broader change set; PR #46100 is just a minimal test-only placeholder and does not share the same concrete code-path change.", "best_issue_reason": null, - "best_pr_reason": "PR #45932 is the most representative and materially meaningful change in the set; PR #46100 is a narrow test-only edit and is not a good canonical representative.", + "best_pr_reason": "PR #45932 is the more substantive artifact and the closest thing to a cluster representative, but it still should not be treated as a duplicate of #46100.", "soft_edge_verdicts": [ { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "Different purpose and code paths: #45932 changes multiple model implementations and CI secrets, while #46100 only edits a ViT test file and adds failing assertions. They do not appear to fix the same bug or implement the same feature." + "reason": "Reject: the PRs touch different things and do not appear to fix the same bug or implement the same change. #45932 changes CI/workflow and several model/config files; #46100 only edits a ViT test file and adds obvious failing assertions. They are not plausibly mergeable into one PR as a single code-path fix." } ] }, @@ -21561,12 +21687,12 @@ "pull_request:46073|pull_request:46196", "pull_request:46154|pull_request:46196" ], - "prepared_review_unit_hash": "f605969488e985877dc33290309e829e6f53dd1992380ea53f75f1975535c6e4", + "prepared_review_unit_hash": "436d1e25452d526b165362ad851896c0b80e80561ac70dfbe36cfda7a78c8e64", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 8083, + "serialized_chars": 8081, "estimated_input_tokens": 2021, "estimated_eval_tokens": 4298 }, @@ -21574,7 +21700,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 8083, + "serialized_chars": 8081, "estimated_input_tokens": 2021, "estimated_eval_tokens": 4298 }, @@ -21583,30 +21709,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate PRs adding SDPA dispatch support to different model families (CTRL, T5, and wav2vec2/SeamlessM4T). They share a broad theme but not the same concrete bug or code path, so they should not be clustered as duplicates.", + "summary": "These PRs are all SDPA-dispatch enablements, but they target different model families and code paths (CTRL, T5, wav2vec2/SeamlessM4T), so they are related feature work rather than duplicates.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No PR is a true duplicate of the others. If a representative must be chosen, 46154 is the cleanest single-model PR and the most self-contained of the set.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "46154 is the most focused and least broad: it targets one model family (T5) with a clear SDPA dispatch change, making it the best representative of the cluster, though not a duplicate of the others.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:46073", "right": "pull_request:46154", "accept": false, - "reason": "Both are SDPA-related, but they modify different model implementations (CTRL vs T5) and different code paths. Same theme, not the same underlying fix." + "reason": "Both add SDPA support, but one is for CTRL and the other for T5; they touch different model implementations and are not the same underlying fix." }, { "left": "pull_request:46073", "right": "pull_request:46196", "accept": false, - "reason": "CTRL SDPA support is unrelated to the wav2vec2_bert / wav2vec2_conformer / SeamlessM4T changes. The PRs touch different architectures and cannot plausibly be merged as one fix." + "reason": "Shared SDPA theme only; CTRL is a separate architecture from wav2vec2/SeamlessM4T, so these are distinct model-specific changes." }, { "left": "pull_request:46154", "right": "pull_request:46196", "accept": false, - "reason": "T5 dispatch support and the multi-model wav2vec2/SeamlessM4T dispatch work address different model-specific implementations, so they are not duplicates." + "reason": "T5 SDPA dispatch and wav2vec2/SeamlessM4T SDPA dispatch affect different code paths and models, so they could not realistically be merged as one PR." } ] }, @@ -21636,12 +21762,12 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "41468e419c5c5cb67fcf6a981df82e5667fca5841c45d993f3c5e15f21d04f15", + "prepared_review_unit_hash": "8f64c7ecbe2543b106c0906a48a6af0bf91c1a3a318653f5d406c0c6f80f3ac1", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5571, + "serialized_chars": 5572, "estimated_input_tokens": 1393, "estimated_eval_tokens": 3042 }, @@ -21649,7 +21775,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5571, + "serialized_chars": 5572, "estimated_input_tokens": 1393, "estimated_eval_tokens": 3042 }, @@ -21658,18 +21784,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two pull requests overlap on distributed/tensor-parallel files, 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 touching many unrelated areas. This is not a duplicate pair.", - "confidence": 0.95, + "summary": "The two PRs are both in distributed-model territory, but they are not the same change: 46126 is a focused EP/router/trainer integration, while 46269 is a broad distributed branch-base PR spanning many subsystems. Shared filenames alone are not enough to treat them as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46126 is the more focused, self-contained change set centered on EP and Trainer integration; PR 46269 is a large umbrella branch with many unrelated distributed updates, so it is a poorer canonical duplicate target.", + "canonical_pr_reason": "46126 is the more focused, self-contained change; 46269 is a broad integration/base branch PR with many unrelated distributed edits.", "best_issue_reason": null, - "best_pr_reason": "PR 46126 is the best representative of a concrete code-path change. PR 46269 is too broad and mixed to serve as a clean duplicate anchor.", + "best_pr_reason": "46126 is the clearest representative artifact in this cluster because it describes a specific feature change rather than an umbrella branch-base sweep.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "Shared filenames and soft similarity are not enough here: 46126 implements a specific EP/Trainer integration, while 46269 is a large distributed branch base spanning many subsystems. They do not look like the same underlying bug or mergeable single PR." + "reason": "Different scope and likely different underlying changes: 46126 adds EP/router/trainer integration, while 46269 is a large distributed base branch touching many areas beyond that feature." } ] }, @@ -21699,7 +21825,7 @@ "soft_pairs": [ "pull_request:46152|pull_request:46207" ], - "prepared_review_unit_hash": "b06bec4a63189d5d5108cb1eec88b0f437d7b18ec7e8b55b375ab80503d5f269", + "prepared_review_unit_hash": "ba1d5d1577ba511732fc1fa5c5b21dd354bed918c4051dfc7c8cfdd077b23b0f", "component_budget": { "node_count": 2, "item_count": 2, @@ -21721,18 +21847,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are related only at a high level: both stop using direct file-open source inspection, but they fix different code paths in different modules with different mechanisms. They should not be merged as duplicates.", - "confidence": 0.89, + "summary": "These are related in theme (removing unsafe/source-file open() access), but they fix different code paths in different modules and should not be merged as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46207 is the better representative of the cluster because it targets one concrete source-reading check in `modeling_utils.py` and includes tests, whereas PR 46152 is a broader import-utils refactor with a different implementation strategy.", + "canonical_pr_reason": "PR 46152 is the broader, more central change: it adds a reusable resource-based source resolution path in import_utils, which addresses the underlying file-open pattern at the utility level.", "best_issue_reason": null, - "best_pr_reason": "PR 46207 is the strongest standalone candidate: it addresses a specific bug-prone source inspection path and is narrowly scoped enough to evaluate on its own.", + "best_pr_reason": "PR 46152 is the best representative for the cluster because it is the more general source-access refactor; PR 46207 is a separate, narrower caller-specific replacement in modeling_utils.", "soft_edge_verdicts": [ { "left": "pull_request:46152", "right": "pull_request:46207", "accept": false, - "reason": "Different underlying changes: 46152 refactors internal source-path resolution in `import_utils.py`, while 46207 changes `_can_set_attn_implementation` in `modeling_utils.py` from file-open to `inspect.getsource`. Same broad theme, but not the same concrete bug or patch." + "reason": "Both avoid open() on source files, but they touch different modules and different behaviors: import_utils helper/source resolution vs modeling_utils attention-implementation detection. They do not appear to fix the same concrete bug or be mergeable into one PR." } ] }, @@ -21767,58 +21893,58 @@ "pull_request:46362|pull_request:46368", "pull_request:46362|pull_request:46366" ], - "prepared_review_unit_hash": "daf9670b7a981f8840e7eed5a9545e310174299c0652fcb79e09e0a428fa375b", + "prepared_review_unit_hash": "4d7adc4aad4138c855ae2e57f161c2ec1192d96093a529e17bfb311eca758327", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 8227, - "estimated_input_tokens": 2057, - "estimated_eval_tokens": 4370 + "serialized_chars": 8271, + "estimated_input_tokens": 2068, + "estimated_eval_tokens": 4392 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 4, - "serialized_chars": 8227, - "estimated_input_tokens": 2057, - "estimated_eval_tokens": 4370 + "serialized_chars": 8271, + "estimated_input_tokens": 2068, + "estimated_eval_tokens": 4392 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are mostly unrelated model-specific test adjustments. The only close soft link is between the two DeepSeek V4 PRs, but they fix different problems: one skips an incompatible quant-cache test, the other adds XPU/CUDA backend selection for distributed testing. I would not merge any of the soft pairs as duplicates.", - "confidence": 0.92, + "summary": "These PRs are all model-test maintenance changes, but they address different underlying failures: one Cohere2MoE device-map test setup fix, two model-specific quant-cache skips, and one DeepSeek V4 XPU/distributed-worker backend adjustment. None of the soft pairs look like the same concrete bug or change.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46366 is the most substantive standalone change in the set: it addresses distributed-test backend selection and is more representative of an actual code-path fix than the model-specific skip-only PRs.", + "canonical_pr_reason": "PR 46362 is the clearest narrow fix: it adds a model-specific skip with an explicit incompatibility reason for DeepSeek V4 quant-cache testing.", "best_issue_reason": null, - "best_pr_reason": "PR 46366 is the strongest standalone candidate to keep because it fixes environment/backend handling rather than merely skipping a failing test, and it is not tied to a single model-specific incompatibility.", + "best_pr_reason": "PR 46362 is the most representative single change in this set because it is a small, self-contained test adjustment with a concrete failure explanation.", "soft_edge_verdicts": [ { "left": "pull_request:46337", "right": "pull_request:46368", "accept": false, - "reason": "Both are model-test maintenance PRs, but they target different models and different failure modes: Cohere2 MoE test setup vs NemotronH quant-cache skipping." + "reason": "Both are test-only model PRs, but they fix different models and different failures: Cohere2MoE device-map/offload setup vs NemotronH quant-cache incompatibility." }, { "left": "pull_request:46337", "right": "pull_request:46362", "accept": false, - "reason": "Different models and different changes. One adjusts Cohere2 MoE test configuration; the other skips a DeepSeek V4 quant-cache test." + "reason": "Unrelated concrete issues: Cohere2MoE test configuration for hf_device_map vs DeepSeek V4 skipping an incompatible quant-cache test." }, { "left": "pull_request:46362", "right": "pull_request:46368", "accept": false, - "reason": "Superficially similar skips, but they are for different model code paths and different cache implementations; not the same underlying bug/change." + "reason": "Both skip generate_with_quant_cache, but they target different model implementations with separate cache incompatibility details; they are not the same underlying bug or a single mergeable change." }, { "left": "pull_request:46362", "right": "pull_request:46366", "accept": false, - "reason": "Same file, but the changes are unrelated: one skips a quant-cache test, the other adds XPU fallback logic for a distributed worker." + "reason": "DeepSeek V4 quant-cache test skip and DeepSeek V4 XPU/distributed-worker backend support affect different code paths and fix different failures, so they should not be merged as duplicates." } ] }, @@ -21850,7 +21976,7 @@ "pull_request:46376|pull_request:46417", "pull_request:46376|pull_request:46425" ], - "prepared_review_unit_hash": "18e3dad1947182c71e5c4dd657bec9ca5de2262d5b8e1c24c9c5a379958ead4d", + "prepared_review_unit_hash": "dd890f8eec3ef7fba556d07989f187b1690d7dc4d2839899646a1455034a7ceb", "component_budget": { "node_count": 3, "item_count": 3, @@ -21872,24 +21998,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster groups three unrelated PRs: multimodal continuous batching infrastructure, a new HunYuan VL model integration, and RADIO model support. They touch different codepaths and represent separate feature additions, not duplicate fixes.", + "summary": "The soft similarities are superficial: one PR adds multimodal continuous-batching runtime support, while the others add two different model families (HunYuan VL and RADIO). They are not the same underlying change or bug.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No clear canonical duplicate PR: the three PRs target distinct subsystems and model families, and the overlap is only broad 'add support' language.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "46376 is the broadest infrastructure PR, but it is still not a duplicate of the standalone HunYuan VL or RADIO model-addition PRs.", "soft_edge_verdicts": [ { "left": "pull_request:46376", "right": "pull_request:46417", "accept": false, - "reason": "Different changesets and goals: continuous batching multimodal support vs. adding a new HunYuan VL model package and docs. No shared concrete bug or mergeable duplicate implementation." + "reason": "Continuous batching multimodal runtime support is unrelated to adding the HunYuan VL model package, docs, and tests; different code paths and deliverables." }, { "left": "pull_request:46376", "right": "pull_request:46425", "accept": false, - "reason": "These are unrelated feature PRs: one modifies continuous batching internals, the other adds RADIO model support across model files and auto mappings. They do not fix the same underlying problem." + "reason": "A multimodal continuous-batching runtime change is not the same as RADIO model integration; they touch different model families and code paths." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index 747ae15037201a21cfc9bf523707c0aa0c44247c..b341e18df568790426e03e26afe5cae86b777bb7 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260605t000034z", + "analysis_id": "hybrid-model-20260605t060128z", "archived_artifacts": { - "hybrid": "snapshots/20260605T000034Z/analysis-runs/hybrid-model-20260605t000034z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260605T000034Z/analysis-runs/hybrid-model-20260605t000034z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-06-05T00:09:30Z", + "published_at": "2026-06-05T06:10:47Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260605T000034Z", + "snapshot_id": "20260605T060128Z", "variant": "hybrid" } diff --git a/snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.json b/snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..e6b0d9164f38c6c332471cea0b352bee6a71c8c4 --- /dev/null +++ b/snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.json @@ -0,0 +1,1995 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260605T060128Z", + "generated_at": "2026-06-05T06:10:45Z", + "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": 0.55, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference" + ], + "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-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, + "file_overlap": 1.0, + "area_overlap": 0.425, + "patch_similarity": 0.866, + "shared_filenames": [ + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "left_ranges": [ + [ + 26, + 46 + ], + [ + 138, + 143 + ], + [ + 182, + 187 + ], + [ + 283, + 295 + ], + [ + 327, + 341 + ], + [ + 348, + 354 + ], + [ + 360, + 369 + ], + [ + 399, + 406 + ], + [ + 411, + 419 + ], + [ + 428, + 433 + ], + [ + 472, + 492 + ], + [ + 507, + 513 + ], + [ + 519, + 528 + ], + [ + 541, + 559 + ], + [ + 562, + 567 + ], + [ + 595, + 601 + ], + [ + 607, + 614 + ], + [ + 627, + 643 + ], + [ + 685, + 690 + ], + [ + 708, + 714 + ], + [ + 720, + 727 + ], + [ + 740, + 757 + ], + [ + 761, + 766 + ], + [ + 780, + 786 + ], + [ + 791, + 798 + ], + [ + 807, + 822 + ], + [ + 840, + 845 + ] + ], + "right_ranges": [ + [ + 26, + 31 + ], + [ + 34, + 44 + ], + [ + 136, + 141 + ], + [ + 180, + 185 + ], + [ + 281, + 293 + ], + [ + 325, + 339 + ], + [ + 346, + 352 + ], + [ + 358, + 367 + ], + [ + 397, + 404 + ], + [ + 409, + 417 + ], + [ + 426, + 434 + ], + [ + 463, + 483 + ], + [ + 498, + 504 + ], + [ + 510, + 519 + ], + [ + 532, + 550 + ], + [ + 553, + 564 + ], + [ + 586, + 592 + ], + [ + 598, + 605 + ], + [ + 618, + 634 + ], + [ + 676, + 688 + ], + [ + 699, + 705 + ], + [ + 711, + 718 + ], + [ + 731, + 748 + ], + [ + 752, + 762 + ], + [ + 771, + 777 + ], + [ + 782, + 789 + ], + [ + 798, + 813 + ], + [ + 831, + 836 + ] + ] + }, + { + "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", + "left_ranges": [ + [ + 458, + 464 + ] + ], + "right_ranges": [ + [ + 458, + 464 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-41115-3", + "canonical_pr_number": 45613, + "duplicate_pr_numbers": [ + 41116 + ], + "target_issue_number": 41115, + "reason": "PRs in cluster-41115-3 are treated as duplicates because they converge on issue #41115 with closing_reference evidence." + }, + { + "cluster_id": "cluster-41211-3", + "canonical_pr_number": 44339, + "duplicate_pr_numbers": [ + 41356 + ], + "target_issue_number": 41211, + "reason": "PRs in cluster-41211-3 are treated as duplicates because they converge on issue #41211 with closing_reference evidence." + }, + { + "cluster_id": "cluster-43240-3", + "canonical_pr_number": 43251, + "duplicate_pr_numbers": [ + 43254 + ], + "target_issue_number": 43240, + "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45561-3", + "canonical_pr_number": 45645, + "duplicate_pr_numbers": [ + 45639 + ], + "target_issue_number": 45561, + "reason": "PRs in cluster-45561-3 are treated as duplicates because they converge on issue #45561 with closing_reference 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/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..50895c2be33219e55b1d7d6c50d8fccb69def6db --- /dev/null +++ b/snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,22030 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260605T060128Z", + "generated_at": "2026-06-05T06:10:45Z", + "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-10000-451", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43296", + "issue:43425", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43827", + "issue:43828", + "issue:43866", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:46015", + "issue:46165", + "issue:46294", + "issue:46304", + "issue:46402" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44918|issue:46165", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:46294", + "issue:44805|issue:45161", + "issue:44560|issue:46165", + "issue:43596|issue:46165", + "issue:43572|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:45874|issue:46294", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:43828|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:45685", + "issue:44918|issue:45161", + "issue:44292|issue:44918", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43828|issue:45685", + "issue:44918|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue: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": "1ca6031677298406b7f90b52eb84cd1dccebc39d81e66856e238e260ccb6c383", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 64, + "serialized_chars": 46879, + "estimated_input_tokens": 11720, + "estimated_eval_tokens": 23696 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous cluster of unrelated bug reports. The only clear duplicate pocket is the DeepSpeed ZeRO-3 / BERT index-out-of-bounds reports; most other soft links are just broad topical similarity and should not be merged.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43638 is the best canonical issue: it most clearly states the shared DeepSpeed ZeRO-3 + non-pretrained BERT failure, and it cleanly absorbs the near-duplicate issue:43596.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43638 is the strongest representative for the cluster\u2019s only true duplicate pair and is specific enough to anchor triage.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different Qwen3.5 failures: missing config attribute vs streaming chat/completions API error." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different models and backends: AMD GPU failure for GPT-OSS vs MPS histogram op not implemented for Int." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Both are model/runtime issues, but the model, hardware, and code path are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Different root causes: autocast dtype mismatch in Phi-tiny-MoE vs GPT-OSS running on AMD GPUs." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Both touch MoE-ish models, but the reported failures and code paths are different." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Separate model/runtime failures with no clear shared bug mechanism." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "Unrelated failures: DeepSpeed collective mismatch vs ONNX export error." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "Different model families and different failure modes during export vs FP8 inference." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ONNX export error is not the same as Qwen3 auto device mapping cuda assert." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "One is a task crash, the other is stale RAG documentation; not the same issue." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs cleanup, but they target different removed v5 references and are not the same concrete change." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "Different breakages: ONNX export vs import failure for GenerationMixin." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "Different training bugs: TRL SFT unpacking vs resume-with-local-model/trust-remote-code." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Both are tensor/device errors, but the models and failure mechanisms differ." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "ZeRO-3/BERT index error is not the same as the generic mask shape mismatch." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Same as above: related runtime style, but different bug and repro." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Different model/input pipeline failures; no shared underlying bug." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Generic shape error vs GPT-OSS MoE tensor-parallel issue; not the same defect." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video chat-template StopIteration is unrelated to resume-training / trust-remote-code." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "DeepSpeed/BERT zero-sized tensor error is unrelated to resume-training with a local model." + }, + { + "left": "issue:43572", + "right": "issue:46165", + "accept": false, + "reason": "Config field regression vs resume-training behavior; different problem." + }, + { + "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": "Collective mismatch under Jamba/ZeRO-3 is not the same as BERT zero-sized tensor indexing." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same underlying DeepSpeed ZeRO-3/BERT empty-tensor index-out-of-bounds bug; wording differs but the repro is the same class of failure." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Both involve model execution, but the Liger Kernel cross-entropy call and GPT-OSS TP issue are different bugs." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Incorrect detokenization vs LoRA merge/save/load mismatch; separate issues." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen3 NVFP4 runtime error is unrelated to the LoRA merge/save problem." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Different distributed-training bugs: checkpointing/FSDP mismatch vs generic mask shape mismatch." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "DeepSpeed training weight-shape error vs collective mismatch; not the same root cause." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Same component family, but CPU-offload incompatibility and meta->to_empty materialization are different bugs." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "BERT zero-sized tensor index error is unrelated to GPT-OSS tensor-parallel behavior." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "Activation-checkpoint/FSDP recompute mismatch is unrelated to CustomTrainer compute_loss signature regression." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "Different model families and different distributed/runtime failures." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different models and failure modes: dtype mismatch in Mistral-3 vs invalid method call in GraniteMoEHybrid." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Gemma4 reverse embedding lookup overhead is unrelated to Qwen3.5 video token/template mismatch." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3/vLLM attribute regression is unrelated to the Mistral tokenizer regex crash." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Device-map forward mismatch vs expensive reverse embedding lookup are different Gemma4 issues." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch-version incompatibility is not the same as Qwen-3 NVFP4 runtime failure." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "FSDP activation-checkpointing mismatch and Qwen3 auto device map cuda assert are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Different models and different bugs: autocast dtype mismatch vs GPT-OSS tensor-parallel issue." + }, + { + "left": "issue:43541", + "right": "issue:43716", + "accept": false, + "reason": "Mixtral grouped_mm tracing error is unrelated to the Mistral-3 image preprocessor dtype mismatch." + }, + { + "left": "issue:43828", + "right": "issue:44292", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Qwen-3 NVFP4 runtime failure." + }, + { + "left": "issue:43716", + "right": "issue:45685", + "accept": false, + "reason": "Mistral-3 image dtype issue is unrelated to an MPS histogram operator limitation." + }, + { + "left": "issue:44918", + "right": "issue:45161", + "accept": false, + "reason": "TRL SFT embedding unpacking bug is unrelated to GPT-OSS TP failure." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Different models and different training/runtime paths; not a duplicate." + }, + { + "left": "issue:45507", + "right": "issue:45685", + "accept": false, + "reason": "Both are MoE-related only at a broad level; the concrete failures are different." + }, + { + "left": "issue:43716", + "right": "issue:44292", + "accept": false, + "reason": "Different model-specific runtime errors with no shared code path." + }, + { + "left": "issue:43039", + "right": "issue:44918", + "accept": false, + "reason": "Cross-entropy/Liger Kernel issue is unrelated to TRL SFT input-embedding unpacking." + }, + { + "left": "issue:41762", + "right": "issue:43638", + "accept": false, + "reason": "Both involve DeepSpeed ZeRO-3, but one is Gemma3 loading and the other is BERT index-out-of-bounds; not the same bug." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are v5 docs cleanup, but they concern different removed references and are not the same concrete documentation change." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Mistral-3 image dtype mismatch is unrelated to Qwen3.5 video StopIteration." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Different backends and errors: autocast dtype mismatch vs MPS Int histogram limitation." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "TRL trainer embedding issue is unrelated to MPS MoE backend failure." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different models and code paths; only broad training/runtime similarity." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "Collective mismatch under Jamba/ZeRO-3 is unrelated to the trainer signature regression." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "FSDP checkpointing mismatch is unrelated to the GenerationMixin import failure." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "AMD GPU model-runtime failure vs outdated RAG examples; not the same issue." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross-entropy call is not the same as the generic mask/tensor shape mismatch." + }, + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Corrupted checkpoint vs missing text encoder weights are different model-loading problems." + }, + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "Both mention version incompatibility, but the affected versions and failure conditions differ." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "PaddleOCR-VL/vLLM load failure is unrelated to USE_DISTRIBUTED=0 support regression." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Same broad ZeRO-3 area, but Gemma3 loading and BERT zero-sized tensor errors are different bugs." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3/vLLM config regression is unrelated to Qwen3 auto device mapping cuda assert." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The only accepted duplicate pair (issue:43596 / issue:43638) is well supported by the titles and cluster signals, and the rest of the soft-edge rejections stay cautious rather than overclaiming sameness. The summary accurately characterizes the cluster as mostly heterogeneous." + }, + "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-10000-451", + "nodes": [ + "issue:36010", + "issue:39290", + "issue:41093", + "issue:41762", + "issue:43039", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43792", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43901", + "issue:43975", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44512", + "issue:44560", + "issue:44589", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45161", + "issue:45237", + "issue:45362", + "issue:45464", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304" + ], + "soft_pairs": [ + "issue:44589|issue:46304", + "issue:45823|issue:46294", + "issue: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:43792|issue:45901", + "issue:43854|issue:43866", + "issue:43425|issue:45070", + "issue:43596|issue:45800", + "issue:43975|issue:45901", + "issue:43792|issue:45070", + "issue:45362|issue:45464", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:43404|issue:43866", + "issue:43541|issue:43828", + "issue:44841|issue:45084", + "issue:36010|issue:39290", + "issue:45800|issue:45823", + "issue:41093|issue:41762", + "issue:43296|issue:43366", + "issue:43572|issue:45070", + "issue:44805|issue:44918", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:43531|issue:45070" + ], + "prepared_review_unit_hash": "f256caf58c2a293616cbc0c532214f0e1390e12d4b2d822f8df0cfc2f1cebee5", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39793, + "estimated_input_tokens": 9949, + "estimated_eval_tokens": 20154 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The tokenizer regression pair is a clear duplicate; the ZeRO-3/Bert index-error pair is related but not a safe duplicate under a conservative standard. Most other links are only broad textual similarity.", + "confidence": 0.68, + "canonical_issue_reason": "issue:45920 is the best canonical issue because it is open, broader in scope, and explicitly aggregates the v5 AutoTokenizer wrong-token-ID regression across multiple model families.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best global issue target for the tokenizer duplicate cluster because it is the most current umbrella report for the same concrete regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Different failure classes: float8 storage lookup vs USE_DISTRIBUTED=0 compatibility." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Both are model-input/runtime bugs, but the symptoms and affected paths differ." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both mention docs, but they concern different APIs and different missing/changed behavior." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "Different model/backend failure vs table-question-answering crash." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Different model-specific loading/processor problems." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Unrelated kernel/loss-path issue vs video StopIteration error." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "Tokenizer detokenization bug vs RAG docs/example drift." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same underlying v5 tokenizer wrong-token-ID regression across models; 45920 is the broader umbrella report." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "Checkpoint-loading/model implementation failure vs documentation examples issue." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Remote-code regressions are not the same as PyTorch USE_DISTRIBUTED=0 support breakage." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Corrupted checkpoint vs unsupported checkpoint-loading implementation; not the same bug." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes constructor incompatibility is unrelated to distributed build/runtime support." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Model loading/type error vs docs/examples issue." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but they involve different models and error modes." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Related ZeRO-3/index-error family, but the concrete failures differ: mask/tensor shape mismatch versus out-of-bounds loading a non-pretrained Bert model; too weak for a clear duplicate." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper runtime failure is unrelated to table-question-answering." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Different model-loading issues with no shared failure mode." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch version incompatibility is not the same as a pydantic PretrainedConfig regression." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "ZeRO-3/Bert loading index error is unrelated to torch/transformers compatibility." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Different feature areas: detokenization vs table QA crash." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper runtime failure is unrelated to PretrainedConfig pydantic breakage." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Same general model family, but different APIs and concrete failures." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "Model load/type error vs table QA crash." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Unrelated sliding-window config issue vs Gemma4 device-mismatch forward error." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Streaming chat API failure is unrelated to MPS histogram/int issues." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Different model/backend problems; no shared code path." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Weight-tying bug is unrelated to checkpoint corruption." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Different runtime failures and model families; not mergeable as one fix." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor/model bug vs template compilation error." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to sliding_window config errors." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers compatibility issue is unrelated to Gemma4 device placement." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are ZeRO-3/index errors, but the concrete symptoms and triggers differ enough to avoid merging." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL load failure is unrelated to GGUF gpt-oss support." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "StableLm config field regression is unrelated to the pydantic PretrainedConfig issue." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Both are index/mask-related, but they hit different training paths and likely different bugs." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Int4 OOM regression is unrelated to tied-weights translation-model issues." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Different MoE/tracing versus TP-only GPT-OSS issues." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch is unrelated to streaming chat API failure." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument is a different loading-path bug than the pydantic PretrainedConfig regression." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF architecture support and Qwen streaming failure are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Sliding-window model issue is unrelated to the pydantic config-field regression." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The tokenizer duplicate call (45812 \u2194 45920) is well grounded, but the ZeRO-3 pair (41093 \u2194 43638) is overstated. The packet only shows broad textual similarity plus different concrete symptoms: one is a mask/tensor shape mismatch, the other is an out-of-bounds error when loading a non-pretrained Bert model under ZeRO-3. That is related, but not enough to call it a clear duplicate under a conservative standard. The summary also overstates this by saying these are the only clear duplicates. Please downgrade or mark the ZeRO-3 link as non-duplicate/uncertain." + }, + "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-10000-451", + "nodes": [ + "issue:36331", + "issue:38175", + "issue:39186", + "issue:39692", + "issue:40990", + "issue:41093", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43404", + "issue:43541", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43866", + "issue:43872", + "issue:43873", + "issue:43950", + "issue:43975", + "issue:43994", + "issue:44292", + "issue:44360", + "issue:44403", + "issue:44479", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44589", + "issue:44863", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45399", + "issue:45507", + "issue:45542", + "issue:45632", + "issue:45698", + "issue:45701", + "issue:45759", + "issue:45800", + "issue:45902", + "issue:46294" + ], + "soft_pairs": [ + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:43541|issue:44589", + "issue:44479|issue:46294", + "issue:41093|issue:43596", + "issue:43054|issue:43994", + "issue:45759|issue:45902", + "issue:45542|issue:45800", + "issue:39186|issue:44918", + "issue:43716|issue:44918", + "issue:40990|issue:44960", + "issue:43645|issue:45399", + "issue:45632|issue:45698", + "issue:43404|issue:45005", + "issue:42915|issue:45399", + "issue:44530|issue:45005", + "issue:43873|issue:45005", + "issue:44863|issue:45005", + "issue:44360|issue:44512", + "issue:44918|issue:45507", + "issue:44492|issue:44512", + "issue:43975|issue:44292", + "issue:44292|issue:45507", + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue: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": "8846a2b1cf07cc9c99b4d712033ebfc09c60e5745fd087512abca6e621de1c60", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36896, + "estimated_input_tokens": 9224, + "estimated_eval_tokens": 18704 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a collection of unrelated issue threads, with a few small duplicate clusters around non-persistent buffer loading, remote-code/custom-module loading, SigLIP2 regressions, and Qwen video-input regressions.", + "confidence": 0.72, + "canonical_issue_reason": "Issue 43950 is the clearest representative of the shared non-persistent-buffer regression in v5, and it has the strongest duplicate linkage in this set (notably 44534, and likely 45902).", + "canonical_pr_reason": null, + "best_issue_reason": "43950 is the best single issue to anchor the cluster because it states the regression clearly, is broad enough to cover closely related reports, and is more canonical than the more model-specific variants.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Different bugs: one is an image preprocessor/model dtype mismatch, the other is a video chat-template token/feature mismatch." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "Both are compatibility complaints, but the reported failures and root causes differ: remote-code breakage vs torch/transformers version incompatibility." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated failures: grouped_mm tracing error vs missing Float8 storage type." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": true, + "reason": "Same Qwen video-input regression family; both point to a video template / token-feature mismatch affecting Qwen3.5-class models." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are IndexErrors, but the symptoms and code paths differ (mask shape mismatch vs zero-size tensor during ZeRO-3 init)." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 regression area; both describe nonsensical/poor outputs from siglip2-base-patch16-224 and likely stem from the same loading/processor issue." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "Different problems: multimodal config unwrapping vs uninitialized rotary-buffer memory after materialization." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "No meaningful overlap beyond general environment/setup incompatibility language." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "Different failures and models: FSDP 2-D weight error vs TRL embedding unpacking failure." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Different model families and different failure modes." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "Unrelated issues: perplexity evaluation on gpt-oss-20b vs a GLM5-specific report." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Different issues: notebook custom-model initialization vs flash-attn fallback gating." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": true, + "reason": "Both concern trust_remote_code/custom-module resolution and cache/path collisions leading to loading the wrong code module." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Related only at a high level ('tied weights'); the model families and concrete bug paths are different." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Unrelated: FineGrainedFP8Config failure vs flash-attn fallback checks." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "Different code paths: paged attention cache/group type error vs tied-weights handling." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior is not the same bug as tied-weights regression." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "Checkpoint loading failure for NemotronH is unrelated to tied-weights translation-model issues." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA/indexing discussion is unrelated to a docs typo about the removed run command." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "Different models and failures: TRL embedding unpacking vs invalid method call in GraniteMoEHybrid." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "Unrelated documentation nits." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Different models and unrelated inference failures." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Different root causes and subsystems." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same underlying regression: non-persistent buffers are being corrupted/junk-filled during v5 loading." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Both mention SigLIP2, but one is a documentation example error and the other is a model-quality regression." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different numerical issues in different training/inference paths; not the same bug." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "Different errors and code paths: custom trainer API mismatch vs DeepSpeed ZeRO-3 loading failure." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Different storage/type incompatibility reports with no shared concrete bug." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": true, + "reason": "Same non-persistent-buffer materialization/loading regression; 45902 looks like a model-specific manifestation of the broader 43950 bug." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "SigLIP2 embedding quality is unrelated to a tokenizer-version change report." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Checkpoint corruption is unrelated to noise/logging when loading a transformer." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Both are SigLIP2-related, but one is a probabilities anomaly and the other is a text-embedding quality complaint; not clearly the same underlying bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays appropriately conservative. The canonical choice of 43950 is well supported by the explicit duplicate-style match to 44534, and the other accepted soft pairs are phrased cautiously enough to fit the issue titles. The rejected pairs also avoid overclaiming beyond 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": 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-10000-451", + "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:45663|issue:45902", + "issue:44910|issue:45859", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:41553|issue:43881", + "issue:44403|issue:44451", + "issue:43975|issue: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": "dd0b94d6366d79d893300215d9388062470d33f14c0ee9dfcc2150a4ab20febd", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36796, + "estimated_input_tokens": 9199, + "estimated_eval_tokens": 18654 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of false-positive similarity matches across unrelated issues. Only two pairs look like true duplicates or near-duplicates: the v5 docs/pipeline-removal reports and the `_is_hf_initialized` loading regression.", + "confidence": 0.93, + "canonical_issue_reason": "No single cluster-wide canonical issue: the items span many unrelated bugs. If one duplicate target is needed, 43827 is the cleanest canonical issue because 44509 restates the same v5 docs/pipeline-removal regression.", + "canonical_pr_reason": null, + "best_issue_reason": "43827 is the best issue to keep: it is specific, actionable, and has a near-identical duplicate in 44509.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Different models and different failure modes: tokenizer error messaging vs missing `all_tied_weights_keys` attribute." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Both involve Gemma/Qwen3.5 MoE memory behavior, but the concrete bugs differ: FSDP2 `shared_kv_states` rebuild vs uninitialized `inv_freq` after `to_empty`." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Different code paths: Flash Attention packed-sequence misinterpretation vs CPU-offload incompatibility in rotary embedding." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Unrelated regressions: LoRA gradient checkpointing vs int4 quantization memory reservation." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Both concern SigLIP2, but one is a quality comparison and the other is an HF-vs-JAX implementation discrepancy; not the same bug." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Voxtral AutoTokenizer error handling is unrelated to glm-4v-9b loading failure." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Generic loading-noise complaint vs a specific model load failure; not the same underlying issue." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Tokenizer detokenization bug and Python 3.13 flash_attn import issue are unrelated." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Both mention loading problems, but they are different models/issues with no shared concrete bug." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "`AutoConfig` missing fields with `trust_remote_code` is not the same as broad remote_code loading breakage." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different issues: PEFT gradient checkpointing vs MoE training memory inefficiency." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Gemma3 ZeRO-3 load failure and Qwen3 MoE FP8 failure are unrelated model-loading bugs." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS histogram/int handling in MoE is unrelated to Gemma4 device-map forward mismatch." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3VL weight_only loading error and GLM5 issue are different model-specific problems." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "Importing Qwen2.5-VL and bitsandbytes `_is_hf_initialized` are separate failures." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading failure and flash-attn2 fallback gating are unrelated." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different models and root causes: glm-4v-9b loading vs missing `all_tied_weights_keys` on Molmo." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Remote_code regression and Gemma4 device-map mismatch are not the same bug." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3 MoE LoRA compatibility across versions is unrelated to a tie-word-embeddings warning in Qwen3.5 fine-tuning." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "`add-new-model-like` token-mapping failure is a different code path than remote_code loading regressions." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Autocast dtype mismatch in Phi-tiny-MoE is unrelated to Gemma4 device placement." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Mistral tokenizer regex patch crash and flash-attn2 fallback blocking are unrelated." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance regressions, but they hit different functions (`convert_ids_to_tokens` vs `convert_tokens_to_ids`) and are not the same bug." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Broad remote_code breakage and Kimi-K2.5 tokenizer codec regression are different concrete issues." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "MobileLLM loading key error is unrelated to flash-attn2 fallback gating." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "lm_head weight tying in Mistral3 and non-persistent buffer corruption are distinct loading/state bugs." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same underlying docs regression: v5 still references removed pipeline tasks in summarization/translation docs." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying `_is_hf_initialized` regression during model loading/init-empty-weights; both report the same TypeError from the same code path." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Qwen3 MoE FP8 failure is unrelated to the `_is_hf_initialized` loading regression." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Whisper loading/run failure and `_torch_extract_fbank_features()` issue are not clearly the same underlying bug." + }, + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Different MoE/deepspeed training problems: weight conversion vs gradient averaging." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative on the soft-edge pairs. The two accepted pairs are plausibly the same underlying regressions by title/code path, and the rest are reasonably rejected as distinct bugs. No evidence of overstatement." + }, + "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-10000-451", + "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:41553|issue:44991", + "issue:43940|issue:45313", + "issue:44928|issue:45588", + "issue:43854|issue:44863", + "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:43366|issue:44589", + "issue:42915|issue:43957", + "issue:43645|issue:43646", + "issue:43550|issue:44910", + "issue:44560|issue:44918", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:44661|issue:45823", + "issue:44560|issue:44805", + "issue:43632|issue:45834", + "issue:38175|issue:43994", + "issue:43638|issue:44661" + ], + "prepared_review_unit_hash": "b65625dd76676f389c7a50ef81f36ba7ad7145bd99d817a56753ce81a6e1598e", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37484, + "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 collection of unrelated Transformers bug reports across different models and subsystems. The only clear duplicate-like pair is the two custom model initialization regressions in Transformers 5.0.0; the rest are too model- or path-specific to merge safely.", + "confidence": 0.84, + "canonical_issue_reason": "issue:43646 is the broader report of the custom model initialization regression, while issue:43645 is the notebook-specific variant of the same underlying bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best representative issue because it states the core regression cleanly and covers the underlying problem without narrowing to a specific notebook scenario.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Different failure modes: a Trainer signature mismatch versus a tensor/mask shape mismatch." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is an effectiveness issue and the other is a missing flag on a trust_remote_code variant." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer-loading regressions affecting different models; same broad area but not the same bug." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 weight-loading failures, but they involve different Qwen models and different symptoms." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "One is an RLHF numerical stability issue; the other is a flash-attention crash on a null auxiliary state." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different model-loading problems for different architectures; no clear shared root cause." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": false, + "reason": "The error text is similar, but there is not enough evidence that these reports share the same concrete code path." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both concern Qwen video regressions, but the described failures are different enough that they likely need separate fixes." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "One is a checkpoint-compatibility problem for NemotronH; the other is missing text-encoder weight loading for Sam3Model." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading errors, but they concern different models and different error behavior." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Unrelated errors: DeepSpeed training vs image processor segmentation-map handling." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttention cache parsing and checkpoint loading are unrelated code paths." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Different Qwen3.5 bugs: rope-index scaling versus uninitialized memory after meta-to-empty materialization." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF architecture support and autocast dtype mismatch are different problems." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Different Gemma4 failures: FSDP2 shared_kv_states handling versus device_map placement mismatch." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "Support for a new GGUF architecture and a sliding_window issue in Qwen3-MoE are not the same bug." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "Audio batching leakage and Whisper empty-transcription handling are separate defects." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer corruption versus long-OCR index errors; different subsystems and symptoms." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Qwen3-VL video input regression and MPS histogram support are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Importing a missing class and meta-device loading breakage are different loading issues." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF model architecture support is unrelated to the float8 storage import/storage error." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Fine-grained FP8 config failure and meta-device loading regressions are distinct bugs." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "These look like the same underlying Transformers 5.0.0 custom-model-initialization regression; one report is just notebook-specific." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Torch.compile SDPA failure on Bamba and flash-attention illegal memory access on Qwen3.5 are different code-path problems." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Both mention Qwen3-related video/embedding flows, but one is a StopIteration video regression and the other is a TRL SFT unpacking failure." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "A script execution failure and GGUF architecture support are not the same issue." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "An import error for GenerationMixin and a Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "Tokenizer mapping-name handling and Gemma4 device placement are different bugs." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different generic runtime failures with no evidence of a shared root cause." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but one is zero probabilities and the other is nonsensical results under AutoModel/pipeline; too little evidence for a single bug." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "ZeRO-3 training on a non-pretrained Bert and tokenizer-mapping-name handling are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative. The only accepted duplicate-like pair, issue:43645/issue:43646, is plausibly the same Transformers 5.0.0 custom-model initialization regression, and the other soft-edge rejections do not overclaim 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": 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-10000-451", + "nodes": [ + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43708", + "issue:43720", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44075", + "issue:44164", + "issue:44291", + "issue:44387", + "issue:44451", + "issue:44488", + "issue:44534", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44877", + "issue:44910", + "issue:44960", + "issue:44977", + "issue:44991", + "issue:45030", + "issue:45276", + "issue:45292", + "issue:45313", + "issue:45335", + "issue:45357", + "issue:45362", + "issue:45406", + "issue:45488", + "issue:45588", + "issue:45812", + "issue:46050" + ], + "soft_pairs": [ + "issue:44877|issue:45030", + "issue:43596|issue:44661", + "issue:44661|issue:44805", + "issue:43317|issue:44756", + "issue:43644|issue:44534", + "issue:38175|issue:39692", + "issue:42491|issue:43720", + "issue:41720|issue:42915", + "issue:44960|issue:44977", + "issue: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:44960|issue:46050", + "issue:45362|issue:45406", + "issue:42491|issue:45292", + "issue:42491|issue:43931", + "issue:44164|issue:44291", + "issue:44910|issue:45588", + "issue:45276|issue:45335", + "issue:43425|issue:43957", + "issue:43404|issue:43454" + ], + "prepared_review_unit_hash": "7beb130493026c84956e6ecc2026a701760085860d5c6f139d70e3d1a5a62643", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36977, + "estimated_input_tokens": 9245, + "estimated_eval_tokens": 18746 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous collection of Transformers v5/model-loading regressions across different model families. The only clear duplicate groups are the non-persistent-buffer junk reports and the resize_token_embeddings propagation reports; the two generic BERT load failures also look like the same underlying load-regression.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43644 is the cleanest canonical anchor: it is the earlier, more descriptive version of the exact duplicate pair with issue:44534 about non-persistent buffers being filled with junk in Transformers 5.0.0.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 is the best overall issue representative among the clear overlaps because it states the bug broadly and cleanly, making it the strongest duplicate-triage anchor in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Different models and different symptoms; both mention config/loading failures, but there is no evidence they share the same concrete bug." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "One is a DeepSpeed/zero3 BertModel index error, the other is an add-new-model-like tokenizer mapping failure; unrelated code paths." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer-mapping validation vs a tensor mask shape mismatch; not the same underlying issue." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Both are loading/offload problems, but one is dequantized model offload failure and the other is mmap/OOM on Strix Halo; different bugs." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Near-verbatim duplicate about Transformers v5 filling non-persistent buffers with junk; same bug and same fix target." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "SigLIP2 zero-probabilities bug vs a documentation example error; different root causes." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "LoRA/qwen3_moe compatibility issue vs BitNet packed-weight loading; unrelated failures." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both involve Qwen3, but one is auto device mapping/cuda assert and the other is FineGrainedFP8Config; not the same concrete problem." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "Different model families and different flash-attention failures; not mergeable into one fix." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3Moe FP8 loading vs Strix Halo mmap/OOM; unrelated issues." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": true, + "reason": "Both report the current/latest version failing to load BERT-family models; likely the same model-loading regression affecting multiple checkpoints." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "DeepSpeed ZeRO-3 language_model weight loading vs save_pretrained visual encoder key regression; different code paths." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "LoRA/qwen3_moe compatibility bug is unrelated to resize_token_embeddings behavior." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer resume/max_steps calculation vs optimizer argument handling; separate trainer/optimizer issues." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/load regressions, but they affect different models and likely different tokenizer paths." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Non-persistent buffer junk vs notebook custom-model initialization; both are v5 regressions but not the same bug." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Docs example/model-processor mismatch vs incorrect inference results in a model; distinct problems." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Different tokenizer regressions: one is LlamaTokenizer pre-tokenizer override, the other is wrong token IDs for Granite models." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both mention memory/offloading, but one is Qwen3 MoE training efficiency and the other is quantization offloading behavior; not the same bug." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different model loading failures: unit-test load issue for GLM-4.7 vs shape mismatch for Qwen3-VL-30B." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Accelerate+deepspeed Qwen3 MoE conversion issue vs GLM-4.7 unit-test loading; unrelated." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "One is training-memory inefficiency, the other is int4 quantization reserved-memory/OOM regression; same broad theme but different code paths." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 issue vs sinq save/load regression; no shared underlying bug." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Qwen3.5 chat crash vs Gemma4Processor missing _tokenizer in transformers serve; different failures." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "Qwen3 MoE hf4.x/hf5.x compatibility issue is unrelated to resize_token_embeddings propagation." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MoE LoRA compatibility issue vs Qwen3-VL weight-shape mismatch; different bugs." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "save/from_pretrained extra_state handling vs init_empty_weights unexpected argument; distinct loading/saving regressions." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both crash in attention-related code, but one is packed sequence misinterpretation and the other is s_aux=None AttributeError; separate issues." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Same resize_token_embeddings regression: embeddings are not propagated to the output/decoder embeddings across model variants." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch 2.10 compatibility vs meta-device loading breakage; not the same bug." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both are weight-tying bugs, but they affect different model classes and are not clearly the same code-path problem." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The duplicate triage is generally grounded in the packet and the soft-edge calls are mostly conservative. The two confirmed duplicate-like groups (43644/44534 and 45276/45335) are well supported by the issue titles, and the BERT loading pair is phrased cautiously enough as a likely shared regression. No mergeability problems 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": 7, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42947", + "issue:43335", + "issue:43381", + "issue:43425", + "issue:43611", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43716", + "issue:43828", + "issue:43854", + "issue:43873", + "issue:43957", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44488", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44877", + "issue:44912", + "issue:44928", + "issue:44960", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45588", + "issue:45663", + "issue:45684", + "issue:45698", + "issue:45701", + "issue:45704", + "issue:45759", + "issue:45812", + "issue:45823", + "issue:45859", + "issue:46315" + ], + "soft_pairs": [ + "issue:41720|issue:44155", + "issue:39692|issue:44960", + "issue:44402|issue:44488", + "issue:42947|issue:45704", + "issue:36010|issue:39404", + "issue:43873|issue:44387", + "issue:45663|issue:45759", + "issue:44291|issue:44756", + "issue:45663|issue:45859", + "issue:44155|issue:45823", + "issue:44387|issue:44756", + "issue:44292|issue:44912", + "issue: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: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": "db9e4ece86bb3656228be1bd28d85c6ce338ef07e4a0535471306dc39bc240b7", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38213, + "estimated_input_tokens": 9554, + "estimated_eval_tokens": 19364 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly noisy and heterogeneous, spanning several unrelated Transformers 5.x regressions (model loading, quantization/offload, gradient checkpointing, multimodal configs, and tokenizer behavior). The only clear duplicate family here is the resize_token_embeddings regression affecting output/decoder embeddings.", + "confidence": 0.57, + "canonical_issue_reason": "Issue 45292 is the cleanest, most general anchor for the resize_token_embeddings bug family; the Gemma4 and t5gemma reports read like model-specific manifestations of the same missing embedding propagation.", + "canonical_pr_reason": null, + "best_issue_reason": "45292 is the strongest duplicate target in the cluster because it states the shared failure mode without model-specific noise and can subsume the closely related embedding-update reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Different failures: Qwen3 device-map CUDA assert vs AudioFlamingo3 batched inference token/embedding leakage." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Doc example errors for SigLIP2 are unrelated to the GLM5 loading issue." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading complaints, but they affect different models and likely different root causes." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/LoRA behavior is unrelated to the T5 apex RMSNorm memory leak." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Import path breakage for GenerationMixin is not the same as Whisper pipeline return_language regression." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both involve quantization/memory, but one is offloading behavior and the other is increased reserved memory/OOM under int4." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Gemma4 FSDP2 shared_kv_states rebuilding is unrelated to Qwen multimodal text_config unwrapping." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "Unexpected init_empty_weights argument handling and Strix Halo mmap OOM are different code paths." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Gemma4 attention/FSDP state handling is unrelated to Qwen3_5Moe rotary embedding CPU offload." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo3 track leakage is unrelated to Gemma4 PLE device-mapping mismatch." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "int4 quantization OOM and Strix Halo mmap OOM are both memory issues but not the same bug." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantization/model-loading failures for different models and backends." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug is unrelated to Gemma4 FSDP attention state handling." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device model loading breakage is a different problem from disabling mmap to avoid OOM." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective under PEFT LoRA and the other is blocked in eval mode." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same resize_token_embeddings regression; 45276 is a Gemma4-specific manifestation and 45292 states the general output_embeddings failure." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "save/from_pretrained extra_state handling is unrelated to mmap-related OOM on Strix Halo." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Gradient checkpointing in eval mode and RLHF NaN from 3D position_ids are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Unit-test model loading for GLM-4.7 is unrelated to Qwen3.5 save_pretrained regression." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression for Qwen VL models is unrelated to strict config blocking granite_speech loading." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Both concern custom-model save/load flows, but one is a permissions side effect and the other is wrong-module loading." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention sink handling crash is unrelated to Gemma4 FSDP shared_kv_states rebuilding." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization breakage and Trainer resume_from_checkpoint max_steps miscalculation are different bugs." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Jupyter custom-model initialization and extra_state save/load handling do not describe the same issue." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same underlying resize_token_embeddings propagation bug; both report embeddings not updating after resize, with different affected model classes." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer wrong IDs regression and convert_tokens_to_ids slowdown are separate tokenizer issues." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is not the same as meta-device loading failures." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility and autocast dtype mismatch are different failures." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Different model-loading failures affecting different checkpoints." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to disabling mmap on Strix Halo." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper return_language pipeline regression is unrelated to tokenizer-version behavior changes." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex_attention on torch >= 2.9 and Gemma4 device mismatch are separate failures." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both are Transformers 5.0 custom-model loading regressions, but the reported breakage and code paths are not the same." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch 2.10 compatibility and Gemma4 device-map forward mismatch are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch incompatibility and Mistral-3 dtype mismatch are different issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet and stays conservative overall. The only clearly supported duplicate family is the resize_token_embeddings regression around 45292/45276/45335, and the remaining soft-edge rejections are appropriately cautious given the titles available." + }, + "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-10000-451", + "nodes": [ + "issue:32090", + "issue:36032", + "issue:39404", + "issue:40990", + "issue:42915", + "issue:43299", + "issue:43421", + "issue:43425", + "issue:43452", + "issue:43582", + "issue:43596", + "issue:43611", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43716", + "issue:43856", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44265", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44661", + "issue:44671", + "issue:44683", + "issue:44737", + "issue:44756", + "issue:44805", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45759", + "issue:45859", + "issue:46050", + "issue:46285", + "issue:46326", + "issue:46402" + ], + "soft_pairs": [ + "issue:44530|issue:45084", + "issue:43906|issue:44530", + "issue:39404|issue:44912", + "issue:44265|issue:44898", + "issue:44479|issue:44805", + "issue:44589|issue:45084", + "issue:43906|issue:46326", + "issue:43425|issue:44918", + "issue:43901|issue:44509", + "issue:46050|issue:46326", + "issue: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:43716|issue:45084", + "issue:46285|issue:46402", + "issue:40990|issue:43421", + "issue:36032|issue:42915", + "issue:43643|issue:44530" + ], + "prepared_review_unit_hash": "5a0321d764168ee413e0582e4cd6a15f958791cd7dfdd564ca5f122665207511", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "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 cluster is mostly a grab-bag of distinct Transformers regressions and model-specific loading/runtime bugs, many around v5 changes, tokenizers, or quantization. None of the soft pairs look like the same underlying bug/change closely enough to merge.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43611 is the broadest and most representative report here: a v5 model-loading regression tied to `base_model_prefix`, which matches the cluster\u2019s dominant theme of loading breakages.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43611 is the best single issue to anchor this cluster because it is general, high-impact, and clearly about a core regression rather than a model-specific edge case.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Both are errors, but one is Qwen3.5 PagedAttentionCache/linear_attention and the other is a template compilation TypeError; different code paths." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "A reproduction of issue 38071 is unrelated to Qwen3.5 cache crashes." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper `return_language` pipeline regression is unrelated to gpt-oss MXFP4 quantization fallback." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "`torch.export` with `torch_compilable_check` is a different path than Perceiver image interpolation failures." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Both are regressions, but one is video input on Qwen VL variants and the other is a mask shape `IndexError`; no shared bug." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to non-template node compilation errors." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "A reproduction of issue 38071 is not the same as the mistral_common `continue_final_message` bug." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch version incompatibility is unrelated to Qwen3.5 embedding unpacking with TRL SFT trainer." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both touch docs, but one is about `return_all_scores` in text classification and the other about removed pipeline tasks; different fixes." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized model save/load corruption is unrelated to tokenizer `continue_final_message` behavior." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both mention Qwen3-VL, but one is a v5 loading regression and the other is a specific weight-shape mismatch; not the same bug." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "Nonsensical SigLIP2 outputs are not the same as `ModelOutput` key assignment when a previous key was `None`." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5Tokenizer special-tokens conflict is unrelated to gguf loading failures for AutoTokenizer/AutoModel." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "Extra state handling in save/from_pretrained is a different problem than unwrapping `text_config` in composite Qwen configs." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression and tokenizer load failure for EMBEDDIA/est-roberta are different failures." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "DeepSpeed ZeRO-3 init `IndexError` is unrelated to compiled flex_attention incompatibility." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer load issues, but one is a generic tokenizer failure and the other is a Mistral regex patch crash; different root causes." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Training memory inefficiency on Qwen3 MoE is not the same as disabling mmap on Strix Halo to avoid OOM." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "Save/load `extra_state` handling is unrelated to video input regressions in Qwen VL models." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline language return issue and CamemBERT MLM prediction regression are unrelated." + }, + { + "left": "issue:44737", + "right": "issue:45859", + "accept": false, + "reason": "Both mention device/offload, but XLNet CPU arange placement and Qwen3_5MoE rotary embedding offload are separate code paths." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Different tokenizer loading failures for different models; no evidence they share the same underlying patch." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon warmup TypeError is unrelated to the `_is_hf_initialized` v5 flag regression." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer mapping/model-addition logic is unrelated to flex_attention compilation failures." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Both are load failures, but for different models and likely different causes; not mergeable as one bug." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 regressions, but one is model loading with `base_model_prefix` and the other is non-persistent buffer initialization; distinct bugs." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer `_gpu_broadcast_one` NoneType error is unrelated to Whisper pipeline `return_language` behavior." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Mistral-3 dtype mismatch and a template compilation TypeError do not share an underlying fix." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "CLIP structure inconsistency and Sam3Model text encoder weight loading are different model-specific issues." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "High perplexity on gpt-oss-20b is unrelated to tokenizer backend post-processor updates." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to compile-time non-template node errors." + }, + { + "left": "issue:44683", + "right": "issue:44661", + "accept": false, + "reason": "Flex attention compilation on torch>=2.9 and `add-new-model-like` tokenizer mapping failure are unrelated." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "These are different model-loading failures; shared version context is not enough to merge." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Shared v5 context only; the failure modes and code paths differ." + } + ] + }, + "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-10000-451", + "nodes": [ + "issue:34567", + "issue:36032", + "issue:40444", + "issue:42175", + "issue:42915", + "issue:43089", + "issue:43262", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43452", + "issue:43482", + "issue:43540", + "issue:43541", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43746", + "issue:43819", + "issue:43883", + "issue:43994", + "issue:44206", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44488", + "issue:44530", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44811", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45325", + "issue:45507", + "issue:45584", + "issue:45987", + "issue:46178", + "issue:46315", + "issue:46326" + ], + "soft_pairs": [ + "issue:34567|issue:40444", + "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:40444|issue:46326", + "issue:42175|issue:46178", + "issue:42915|issue:44530", + "issue:44811|issue:45584", + "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:44403|issue:45081", + "issue:43404|issue:44466", + "issue:43883|issue:44991", + "issue:42915|issue:45084", + "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": "243551020d5c6ff16764e75c7ba540c4340c8f1b8e7f5eb02b0a44d3763f3205", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "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 heterogeneous: most items are unrelated v5 regressions across loading, multimodal, audio, generation, and custom-model paths. I did not find a clear duplicate pair with enough evidence to merge.", + "confidence": 0.84, + "canonical_issue_reason": "No single canonical issue cleanly represents the whole cluster; if one must be chosen, issue 43646 is the broadest v5 regression around custom model initialization.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43646 is the most general and representative of the recurring v5 custom-model/init regressions in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "Unrelated bugs: trainer token-count tracking vs multimodal fine-tuning failure." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are load failures, but for different models and likely different underlying causes." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Different code paths: torch dynamo grouped_mm tracing vs template compilation error." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Different areas: iterable-dataset multimodal finetuning vs tokenizer continue_final_message behavior." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Both mention backend packaging, but one is torch extras and the other is tensorflow_text backend registration." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Different failures in different Qwen model families and different runtime components." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Whisper-related, but one is batch_decode skipping special tokens and the other is empty-transcription generation after alignment." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Related v5 init regressions, but one is non-persistent buffer corruption and the other is custom model initialization; not clearly the same bug." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Both audio-feature extraction related, but the titles point to different concrete failures." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both involve GGUF loading, but one is a from_pretrained/gguf_file API break and the other is a specific Qwen2.5-GGUF regression." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Completely different problems: missing tied-weights metadata vs model loading failure for a specific checkpoint." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer method conflict vs custom model initialization in notebooks." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Different loading problems: device_map/offload dequantization vs GGUF/v5 model loading." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Both generation/cache-adjacent, but one is compiled-forward chunked generation inconsistency and the other is a return-type contract bug." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Different issues entirely: template-node compilation vs GraniteMoEHybrid invalid method call." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are version-gated loading failures, but they affect different models and likely different loaders." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen variants and different bugs: hidden-states output vs rope-index scaling." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT checkpoint loading vs tokenizer continue_final_message logic; no shared bug." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio chat-template sampling-rate default vs MIMI encoder padding-mask mismatch." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions around model initialization, but the symptoms are different and not clearly one code-path bug." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization breakage vs non-persistent buffer junk; related theme, but not the same concrete defect." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Generic loading noise vs Mistral regex patch crash; unrelated." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Both mention tied lm_head weights, but one is an untied model class and the other is device-dependent serialization inconsistency." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Unrelated: missing all_tied_weights_keys vs tokenizer loading regression for a specific model." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Different model/runtime failures; no shared code path." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation performance overhead vs cache length return-type bug." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "Different modalities and components: image/text model output quality vs processor image-size mismatch." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Both multimodal, but one is a segmentation fault on Llava Next load and the other is a video-input ValueError in Qwen3OmniMoe." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "Different audio model bugs: MIMI padding-mask mismatch vs DAC from_latents/forward mismatch." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Different model-loading failures; no evidence they share the same loader bug." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different failures in different tokenizer/model paths; not the same underlying bug." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export export failure vs processor output-size mismatch; unrelated." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer loading regression vs slow convert_tokens_to_ids performance regression; different defects." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Different bugs and different triggers: FineGrainedFP8Config vs init_empty_weights/_is_hf_initialized handling." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio fbank feature extraction issue vs compiled flex_attention 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": 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-10000-451", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42915", + "issue:43296", + "issue:43381", + "issue:43404", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43742", + "issue:43792", + "issue:44062", + "issue:44079", + "issue:44291", + "issue:44403", + "issue:44462", + "issue:44479", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44756", + "issue:44811", + "issue:44829", + "issue:44877", + "issue:44987", + "issue:44991", + "issue:45020", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45479", + "issue:45538", + "issue:45663", + "issue:45812", + "issue:45923", + "issue:46209", + "issue:46326" + ], + "soft_pairs": [ + "issue:45216|issue:45357", + "issue:36032|issue:43742", + "issue:44062|issue:44987", + "issue:42915|issue:44589", + "issue:43632|issue:44756", + "issue:46209|issue:46326", + "issue:44479|issue:45325", + "issue:36032|issue:41720", + "issue:43531|issue:44877", + "issue:43404|issue:45127", + "issue:43296|issue:43531", + "issue:40444|issue:45325", + "issue:43611|issue:44534", + "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": "87b9bb80960ab254c0b4a4c09696543f80dbb9cdeea242217155a600efa7a50c", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35809, + "estimated_input_tokens": 8953, + "estimated_eval_tokens": 18162 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are distinct loading/regression bugs that only share broad transformer-version or model-family similarity. I would not merge any of the soft pairs as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "issue:45020 is the broadest cluster anchor: it describes recent-transformers breakage across models using remote_code, which is the closest high-level umbrella among these otherwise model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best representative issue because it captures the widest symptom pattern in the set; however, the cluster is still too diverse for duplicate consolidation.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45216", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5 save_pretrained regressions, but one is about incorrect checkpoint contents generally and the other specifically about visual encoder key saving; related, not the same bug." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Different tokenizer/model-load failures: T5 tokenizer method-name conflict versus a MobileLLM key error." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "Both mention loading problems, but the first is an AddedToken special-argument error while the second is a recent-version model load failure for a specific repo." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Both involve fp8-related failures, but one is Qwen3MoE with FineGrainedFP8Config and the other is a missing Float8 storage object; different code paths." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "Transformers v5 internal flag breakage versus mmap OOM on Strix Halo; unrelated symptoms and fixes." + }, + { + "left": "issue:46209", + "right": "issue:46326", + "accept": false, + "reason": "FP8 inference on B200 and continue_final_message behavior are unrelated problems." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Both are Qwen2.5-VL/video-related regressions, but one is a broad v5.3 video-input issue and the other is a rope-index scaling bug for still-image positions." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer load conflict versus Qwen3 auto device-map cudaErrorAssert; no shared underlying bug." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Sliding_window behavior in Qwen3-MoE and strict config loading for granite_speech are different issues." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "One is a missing lm_head weight tie in Mistral3 image-text generation, the other is LoRA merge collapse with extended vocab; different mechanisms." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window bug are unrelated." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "IterableDataset multi-image fine-tuning failure is not the same as the Qwen2.5-VL rope-index scaling regression." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Both are Transformers v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffers getting junk values." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch 2.10 incompatibility and loading noise are not the same bug." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Extremely high perplexity on GPT-OSS and Whisper batch_decode skip_special_tokens are unrelated." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper model execution failure versus missing Float8 storage object; different failure modes." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but one concerns _is_hf_initialized and the other buffer initialization; not duplicate." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key handling and Gemma4 FSDP2 shared_kv_states rebuilding are unrelated." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Loading noise and tokenizer-loading failure for est-roberta are different bugs." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Both are Qwen2-VL/Qwen2.5-VL training-related, but one is multi-GPU training and the other is multi-image iterable dataset finetuning; not the same underlying issue." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime special-token post-processor updates and BOS/EOS not being added on add_special_tokens=True are tokenizer bugs, but they are distinct behaviors and fixes." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is a torch/kernel compatibility issue, not the remote_code loading regression." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Gradient checkpointing in eval mode and missing supports_gradient_checkpointing on a trust_remote_code model are related only superficially." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument during init_empty_weights and missing Float8 storage object are separate transformer-v5 failures." + }, + { + "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 iterable finetuning." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer.json being ignored and Granite models getting wrong token IDs are connected only at a high level; they are not the same bug." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "Qwen2.5-GGUF loading under v5 and CLIPTokenizer model_max_length are different tokenizer/loading problems." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Flash-attention-3 training degeneration and single_label_classification zero-loss are both training regressions, but they affect different code paths." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Qwen2.5-GGUF loading failure and missing pad_token_id on Llama4Config are unrelated configuration/loading issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-10000-451", + "nodes": [ + "issue:40444", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43531", + "issue:43582", + "issue:43606", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44877", + "issue:44898", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45491", + "issue:45542", + "issue:45701", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:46402" + ], + "soft_pairs": [ + "issue:43317|issue:43856", + "issue:42175|issue:45542", + "issue:43818|issue:46402", + "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:43644|issue:44079", + "issue:40444|issue:45381", + "issue:44265|issue:45072", + "issue:43749|issue:44079", + "issue:43611|issue:43646", + "issue:45325|issue:45381", + "issue:42175|issue:43296", + "issue:43792|issue:43957", + "issue:43632|issue:45072", + "issue:43994|issue:44898", + "issue:45325|issue:45834", + "issue:43454|issue:44466", + "issue:42915|issue:43482", + "issue:43701|issue:45325", + "issue:44291|issue:44877", + "issue:44479|issue:45823", + "issue:43873|issue:45701", + "issue:43296|issue:45800", + "issue:43606|issue:44265", + "issue:43262|issue:43746" + ], + "prepared_review_unit_hash": "1ab9eefd3714af22a32703d336c1b602266364855862cdd719fab187abb645fa", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35892, + "estimated_input_tokens": 8973, + "estimated_eval_tokens": 18202 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a semantic near-match set, not a true duplicate set: a few items overlap around Transformers v5 loading/initialization regressions and Qwen2.5-VL video/position-id issues, but the individual bugs are materially different. I would not merge any of the soft edges.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43611 is the broadest and most representative anchor in the cluster: a clear Transformers v5 model-loading regression. It sits near several other v5 init/load complaints, even though the cluster overall is too heterogeneous to treat as one duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43611 is the best single representative because it is concrete, broadly scoped, and central to the largest thematic slice of the cluster (v5 loading/init breakages).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Different problems: device-map/offload loading failure vs Qwen3 MoE training memory usage." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both involve TensorFlow/backend packaging, but one is missing TF in the torch extra while the other is an undefined tf backend from tensorboard-only installation." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Unrelated multimodal model bugs: Video-LLaVA visual tower vs Sam3Model text encoder weight loading." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Completely different domains: tokenizer error messaging vs Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "One is ModelOutput key assignment, the other is non-persistent buffer initialization in v5; not the same bug." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Both touch vision models, but one is Video-LLaVA tower/weights and the other is Qwen3.5 save_pretrained key serialization." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Both are v5-related, but sliding-window Qwen3-MoE behavior and init_empty_weights argument handling are distinct bugs." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "No underlying overlap: Qwen3 auto device mapping CUDA assert vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both mention batching/padding behavior, but they affect different models and different failure modes." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Different regressions: buffer serialization/initialization vs ModelOutput key bookkeeping." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL vision issues, but one is finetuning with multiple images and the other is video vision_position_ids." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "No common bug path: torch.export/exportability vs bfloat16 dtype mismatch in inference." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "FSDP CPU-RAM-efficient loading is unrelated to ModelOutput key assignment." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 loading/init regressions, but one targets base_model_prefix loading and the other custom model initialization; too broad to call duplicates." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": false, + "reason": "Adjacent Qwen2.5-VL position-id regressions, but one is still-image temporal scaling and the other is video input vision_position_ids." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "Packaging/backend issue vs a specific PaddleOCR-VL load failure in vLLM; not the same bug." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper runtime failure and meta-device model loading breakage are unrelated." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "Different code paths: _is_hf_initialized flag breakage vs dtype mismatches during bfloat16 inference." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision-model inference issues, but they involve different architectures and different root causes." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Both are multimodal indexing/positioning style bugs, but on different models with different symptoms and likely different fixes." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both mention lm_head/tied weights, but one is a load/tie failure in AyaVision and the other is device-dependent serialization inconsistency." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Different issues: Qwen3Moe FP8 config failure vs Qwen2.5-GGUF loading under transformers v5." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch is unrelated to Qwen2.5-VL position-id scaling." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Both are loading regressions, but one is init_empty_weights argument handling and the other is strict config validation for granite_speech." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Qwen video-input regression and Gemma4 device_map offload mismatch are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to tokenization changes across versions." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "PaddleOCR-VL load failure in vLLM and torch/transformers version incompatibility are not the same underlying bug." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU offload device mismatch and torch.export exportability are different failure modes." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio processor sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43454", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43606", + "issue:43720", + "issue:43749", + "issue:43819", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:44186", + "issue:44265", + "issue: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:44568|issue:45920", + "issue:35707|issue:41720", + "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:40990|issue:42947", + "issue:43377|issue:44898", + "issue:43322|issue:46050", + "issue:43454|issue:43720", + "issue:45325|issue:46315", + "issue:41720|issue:43322", + "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" + ], + "prepared_review_unit_hash": "527a6a29d1ee542da8bafe66e76dc3fbd33db5b67e53f14018ccbc6c08e8def1", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35932, + "estimated_input_tokens": 8983, + "estimated_eval_tokens": 18222 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a loose mix of mostly unrelated Transformers bugs and regressions. A few are superficially similar (tokenizer/version regressions, device/offload failures, multimodal loading issues), but none of the soft pairs look like the same underlying defect well enough to merge.", + "confidence": 0.19, + "canonical_issue_reason": "issue:45701 is the broadest umbrella report in the set: it frames a version-wide tokenization regression with multiple inbound references and can represent the general tokenizer breakage theme better than narrower model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the most actionable standalone bug report: it names concrete affected models, a clear regression, and has enough scope to be a useful reference for tokenization correctness issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Both mention generation quality, but one is perplexity on GPT-OSS and the other is empty-transcription Whisper failure; different models and code paths." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Bark voice preset failure is audio-processing specific, while the other is a generation bug with inputs_embeds/past_key_values; no shared underlying defect." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both concern tokenizer behavior after v5, but one is mdeberta BOS/EOS insertion and the other is wrong token IDs across several models; likely separate tokenizer regressions." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation with cached embeddings is unrelated to Qwen3 auto device-mapping cuda assert; different subsystems and symptoms." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "LayoutLMv2 tokenizer NER/padding crash is unrelated to torch.export failures around torch_compilable_check." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both are device-map/forward failures, but one is a Qwen3 cuda assert and the other is a Gemma4 PLE mismatch; not the same bug." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Both are device/offload-related, but Bark CPU offload mismatch and DAC missing STE behavior are different model-specific issues." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "AyaVision tied-weights garbage text and Bamba torch.compile/SDPA failure are distinct model implementation problems." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 doc example errors and Qwen2.5-VL iterable-dataset multi-image finetuning failure are unrelated examples of different workflows." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "glm-4v-9b loading failure and sleng-bert tokenizer loading failure are separate model/load paths with no clear common root cause." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "One is a MoonshineStreaming sliding-window off-by-one bug, the other is slow convert_tokens_to_ids performance; different failure modes." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI batch-vs-single padding-mask inconsistency is unrelated to OmDet-Turbo processor/model image-size mismatch." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "AyaVision weight tying and SwitchTransformers/TimmWrapper dtype mismatches are separate model correctness issues." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen3 MOE accelerate+deepspeed weight conversion and Nemotron gradient-checkpointing flag are unrelated configuration/load issues." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "StableLm pad_token_idx missing and Qwen 3.5/3.6 from_config text_config unwrapping are different config-layer bugs." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Device-map dequantized-model offload failure and generic offloading-with-quantization report are too broad and lack a concrete shared code path." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on GPT-OSS and gradient-checkpointing ineffective with LoRA are unrelated training/inference problems." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI encoder padding mask bug and Perceiver non-default-resolution failure are distinct vision/audio model issues." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next segfault on load and sinq quantized save/load breakage are not the same failure, even though both involve model loading." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "AyaVision tied weights and BitNet packed-weight unpacking during accelerate loading are different implementation defects." + }, + { + "left": "issue:45325", + "right": "issue:46315", + "accept": false, + "reason": "Qwen2.5-VL rope-index scaling and tokenizer conversion performance regression are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3 auto device-mapping assert and Llava Next segmentation fault are different model-loading failures." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU RAM efficient loading breakage and non-persistent buffer junk in v5 are not the same bug." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Both are version-dependent loading failures, but one is tokenizer loading for a specific model and the other is kernels import requiring revision/version; different root causes." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "3d_parallel.py runtime failure and Qwen3Moe FineGrainedFP8Config failure are unrelated errors." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3 cuda assert on A800 and Qwen3.5 flash-attention generation issues are separate inference paths and symptoms." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both involve save/load serialization around multimodal weights, but one is general lm_head tied-weight serialization and the other is Qwen3.5 visual encoder keys; not clearly the same defect." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels and Molmo tied-weights attribute error are unrelated image processor vs model attribute bugs." + }, + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "GPT-OSS perplexity and broad tokenization changes are too different to treat as one underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "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:43881", + "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:43881|issue:44991", + "issue:36331|issue:42617", + "issue:43065|issue:46402", + "issue:44186|issue:45072", + "issue:44877|issue:44991", + "issue:43577|issue:43708", + "issue:36331|issue:42915", + "issue:43450|issue:45701", + "issue:44008|issue:45823", + "issue:43873|issue:44756", + "issue:43550|issue:43720", + "issue:43404|issue:45357", + "issue:42175|issue:46291", + "issue:41720|issue:42617", + "issue:44423|issue:45406", + "issue:43550|issue:43606", + "issue:44610|issue:44898", + "issue:44186|issue:44610", + "issue:43295|issue:45701", + "issue:42915|issue:43317", + "issue:45632|issue:45684", + "issue:43425|issue:46291", + "issue:36010|issue:42617", + "issue:41720|issue:45161", + "issue:43525|issue:46291", + "issue:43531|issue:43572", + "issue:45081|issue:45812", + "issue:44671|issue:44743", + "issue:43316|issue:44913" + ], + "prepared_review_unit_hash": "1013963e773c095d7dc1d6de61d4455a0bed46a6987c5a331f28966cbdc4b89c", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36769, + "estimated_input_tokens": 9193, + "estimated_eval_tokens": 18642 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "No genuine duplicate cluster emerges here: the soft pairs mostly connect unrelated model-, tokenizer-, trainer-, and serving-specific bugs that only share broad subsystem words like quantization, offloading, or tokenization.", + "confidence": 0.72, + "canonical_issue_reason": "issue:43873 is the broadest and most connected issue in this set (quantization/offloading with inbound references), so it is the best anchor if a single issue must represent the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the strongest global representative because it is active, broadly scoped, and externally referenced; the other issues are narrower point failures.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Both are memory-related, but one is int4 quantization OOM and the other is a T5 apex RMSNorm leak; different code paths and root causes." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Different models and failures: glm-4v-9b loading vs est-roberta tokenizer loading." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "Trainer compute_loss kwarg error is unrelated to a 3d_parallel.py runtime failure." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Same model family, but dummy Conv2d in Sam3PixelDecoder and missing text encoder weights are distinct bugs." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer padding/NER crash is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Config strictness for granite_speech is unrelated to est-roberta tokenizer loading." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Blip2 dtype propagation and Trainer resume/max_steps logic are separate issues." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Unexpected compute_loss kwarg and Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batch-shape bug is unrelated to tokenization changes." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma-related, but one is a variable name collision in forward and the other is a device-map mismatch in a different model." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both mention offloading/memory, but one is quantization offload behavior and the other is mmap OOM on Strix Halo; not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "torch.compile with SDPA and BitNet packed-weight loading are different concrete failures." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Mistral3 lm_head tying and Qwen3.5 visual-encoder key saving are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Packaging/backend dependency selection and kernels version/revision requirements are different installation issues." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 device-mapping CUDA assert is unrelated to running 3d_parallel.py." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both affect serve, but continuous-batching multimodal crash and missing Gemma4Processor tokenizer are separate code paths." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both involve image size mismatches, but they are for different models and different failure mechanisms." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes and OmDet-Turbo resolution mismatch are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Processor.tokenizer regression and version-dependent tokenization changes are not the same specific bug." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "FP8 config failure and dequantized model offload failure are distinct quantization/loading problems." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collisions and read-only permission propagation from save_pretrained are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch incompatibility and kernels dependency-version enforcement are different dependency issues." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error is unrelated to 3d_parallel.py execution." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both touch parallelism/device mapping, but one is an A800 CUDA assert and the other is a GPT-OSS TP issue; not the same bug." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Missing pad_token_id and kernels version/revision dependency errors are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Both mention missing padding-related config fields, but they affect different models and different attribute names." + }, + { + "left": "issue:45081", + "right": "issue:45812", + "accept": false, + "reason": "Mistral regex patch crash and Granite wrong token IDs are different tokenizer regressions." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression and Qwen3.5 recurrent-state reset are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Duplicate of offloading/memory theme only; the underlying problems differ enough that they should stay separate." + }, + { + "left": "issue:43295", + "right": "issue:44991", + "accept": false, + "reason": "Processor API regression and tokenizer loading failure are not the same underlying defect." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "SDPA torch.compile failure and CPU-offload device mismatch are different model/runtime bugs." + }, + { + "left": "issue:44610", + "right": "issue:44186", + "accept": false, + "reason": "Image processor resolution bug and LayoutLMv2 tokenizer crash are unrelated." + }, + { + "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-10000-451", + "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:44186|issue:44361", + "issue:42915|issue:43322", + "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:43873|issue:45704", + "issue:44530|issue:46291", + "issue:45406|issue:45464", + "issue:41720|issue:43746", + "issue:43421|issue:45701", + "issue:45357|issue:46402", + "issue:43606|issue:43653", + "issue:43531|issue:45759", + "issue:34634|issue:41720", + "issue:44492|issue:44509", + "issue:42490|issue:43329", + "issue:34567|issue:35707", + "issue:44625|issue:45479", + "issue:44451|issue:44991", + "issue:44514|issue:44545", + "issue:43531|issue:44589" + ], + "prepared_review_unit_hash": "53c3f91fdce4fcc41534ea0747acace67b9de523390f8016a1a2c5fc4acd462c", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36222, + "estimated_input_tokens": 9056, + "estimated_eval_tokens": 18368 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous. There is one strong duplicate candidate: issues 44514 and 44545 share the same Qwen2_5_VLProcessor.apply_chat_template batched-input/padding=False title and appear to describe the same bug. The remaining pairs look like distinct issues in different models or code paths.", + "confidence": 0.76, + "canonical_issue_reason": "issue:44545 is the best canonical issue for the close duplicate pair because it matches issue:44514 on the same title and has slightly more inbound references/discussion activity.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44545 is the strongest issue representative in this cluster because it is the only accepted duplicate pair and is a direct match for issue:44514.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Different models and failure modes: Qwen3OmniMoe video processing vs Qwen3VL from_pretrained weight_only loading." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Both are tokenizer-related, but the concrete failures differ: LayoutLMv2 NER/batching vs MLuke task AttributeError." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Unrelated loading failures: Qwen3Moe fine-grained FP8 config vs Llava Next segmentation fault." + }, + { + "left": "issue:43421", + "right": "issue:45812", + "accept": false, + "reason": "Both involve tokenizers, but one is runtime post-processor syncing and the other is a Granite token-ID regression." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer loss-signature mismatch and Qwen3 auto device-map CUDA assert are unrelated bugs." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Different subsystems: BeitImageProcessorFast label reduction vs glm-4v model loading failure." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Different tokenizer issues: fast/slow inconsistency vs T5 add_special_tokens name conflict." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS CLIP bus error is unrelated to progressive generation with inputs_embeds/past_key_values." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Evaluation batching bug and SAM3 text/bounding-box behavior are different problems." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Image segmentation-map processing and BarkProcessor voice_preset handling are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "torch.compile SDPA failure and BigBirdTokenizer special-token registration are different bugs." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both involve loading/offload, but the underlying problems and models differ materially." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization offloading behavior and a T5 RMSNorm memory leak are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:46291", + "accept": false, + "reason": "PagedAttentionCache linear_attention crash and kernels import/version validation are different code paths." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming inference failures are not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3 device-mapping CUDA assert is unrelated to GraniteSpeech PEFT local checkpoint loading." + }, + { + "left": "issue:43421", + "right": "issue:45701", + "accept": false, + "reason": "Runtime special-token/post-processor syncing is different from a broad tokenization-version regression." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Different save/load propagation bugs: Qwen3.5 visual encoder keys vs Sam3Model text encoder weights." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to BigBirdTokenizer mask-token registration." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Sliding-window behavior and composite config unwrapping are distinct Qwen-related issues." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping failure are unrelated." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "These are different documentation issues: a typo vs stale pipeline-task references." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Local-path import warning and multimodal token-counting video-branch bug are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token counting and progressive generation cache handling are different problems." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "One is Qwen3.5 config propagation; the other is a general classification loss bug for num_labels=1." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/model-loading regressions, but they affect different models and failure points." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Same title and same reported Qwen2_5_VLProcessor.apply_chat_template batched-input crash with padding=False; this looks like the same bug." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Sliding-window Qwen3-MoE bug and Float8 storage lookup error are unrelated." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The soft-edge rejections are mostly conservative and grounded in the issue titles, but the cluster-level conclusion overstates the evidence. In the packet, 44514 and 44545 have identical titles, which supports a strong duplicate candidate, but the summary goes beyond the packet by claiming they are the \"only true duplicate pair\" and that 44545 has the \"most engagement.\" That is not well supported: several other issues have higher discussion_activity, and the packet does not establish that the two reports describe the exact same crash beyond the matching title. The canonical-issue rationale also adds subjective claims like \"clearer/cleaner report\" that are not grounded in the packet. Because the summary and confidence are too strong for the available evidence, I would reject this cluster result." + }, + "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": 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-10000-451", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:42915", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43645", + "issue:43653", + "issue: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:45759|issue:45859", + "issue:44077|issue:44625", + "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:39692|issue:44610", + "issue:43782|issue:45381", + "issue:43746|issue:46050", + "issue:42886|issue:44843", + "issue:45704|issue:45923", + "issue:43421|issue:43653", + "issue:43381|issue:43844", + "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:43596|issue:44220", + "issue:43540|issue:46294", + "issue:42915|issue:43746", + "issue:45356|issue:45907", + "issue:30064|issue:36010" + ], + "prepared_review_unit_hash": "56c6001ee60254d943de2f43b006f895d733e20c3a9e4ddd3c6a5f34cd5be393", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 37083, + "estimated_input_tokens": 9271, + "estimated_eval_tokens": 18798 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a heterogeneous mix of unrelated Transformers issues. The only clear duplicate-style pair is the offline tokenizer/Mistral-regex regression; most other soft-similarity pairs share only a broad subsystem or model family, not the same concrete bug.", + "confidence": 0.79, + "canonical_issue_reason": "issue:44843 is the clearest root-cause report in the only near-duplicate pair: it pinpoints the offline regression to an unconditional `model_info()` call inside `_patch_mistral_regex`, while issue:42886 describes the same symptom at a higher level.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44843 is the best representative because it names the exact failing code path and bug trigger, making it the strongest canonical anchor for the cluster\u2019s only true duplicate subthread.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Different failures: image interpolation/resolution handling vs bfloat16 dtype mismatch." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Different code paths: config unwrapping for composite Qwen models vs CPU-offload incompatibility in rotary embedding." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Unrelated bugs: optional `post_init` handling vs `num_labels` propagation in Qwen3.5." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both are tokenizer-related, but one is runtime special-token/post-processor updates and the other is save/load token-ID loss." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Different components: flash-attention sink handling vs multimodal config unwrapping." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken constructor error is unrelated to the model-loading failure for ScandiBERT." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "One is a tie-word-embeddings warning during LoRA fine-tuning; the other is incorrect visual encoder key saving." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Different problems: a specific model load regression vs PIL image processors wrongly requiring torchvision." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Different training/eval failures with different triggers and stacks." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Doc-example errors for SigLIP2 vs an OmDet-Turbo processor size mismatch." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Both involve Qwen VL/video, but the reported bugs differ: weight_only loading vs video position IDs." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "Local PEFT checkpoint loading issue vs quantized-model roundtrip corruption after save/load." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": true, + "reason": "Same offline tokenizer regression: both point to `AutoTokenizer.from_pretrained` / `_patch_mistral_regex` breaking when `HF_HUB_OFFLINE=1`; 44843 gives the exact root cause." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "Completely different issues: apex RMSNorm memory leak vs missing gradient-checkpointing flag." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both touch special tokens, but one is runtime post-processor updates and the other is BigBird mask-token registration/decoding." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "One is an eval-mode restriction for gradient checkpointing; the other is a ZeRO-3 training instability." + }, + { + "left": "issue:45663", + "right": "issue:45923", + "accept": false, + "reason": "Different model/runtime bugs: Gemma4 FSDP2 shared KV state rebuilding vs a trust_remote_code flag omission." + }, + { + "left": "issue:45704", + "right": "issue:45859", + "accept": false, + "reason": "Unrelated: T5 apex RMSNorm leak vs Qwen3_5MoE rotary embedding CPU-offload incompatibility." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Different subsystems and symptoms: Qwen3 device-mapping CUDA assert vs T5 RMSNorm leak." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to Jupyter custom-model initialization in Transformers 5.0." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Both are model-quality regressions, but on different models and different failure modes." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Tokenizer special-token update bug is not the same as LayoutLMv2 crashing on NER padding/truncation." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT checkpoint loading and quantization roundtrip differ from generic quantization offloading behavior." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Related to Mistral regex/tokenizer handling, but one is an AttributeError crash and the other is a codec-handling regression; not the same concrete bug." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 doc example errors are unrelated to quantization/offloading behavior." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Different low-level failures: ZeRO-3/BertModel indexing vs fbank feature extraction." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both are Qwen video-input issues, but the concrete failure is different: ValueError in processing vs chat-template token/feature mismatch." + }, + { + "left": "issue:42915", + "right": "issue:43746", + "accept": false, + "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to GraniteSpeech PEFT checkpoint loading." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Mistral-regex/codec regression is unrelated to `list_repo_files` not forwarding `revision`." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Image segmentation-map processing and GenerationMixin import errors are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the set as mostly heterogeneous and identifies issue:44843/issue:42886 as the only clearly duplicate-style pair. The soft-edge verdicts are generally conservative and the accepted pair is well-supported by the titles and stated root cause. No PR-mergeability concerns apply here." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "nodes": [ + "issue:30064", + "issue:36331", + "issue:39692", + "issue:41720", + "issue:43295", + "issue:43322", + "issue:43335", + "issue:43388", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43618", + "issue:43643", + "issue:43653", + "issue:43697", + "issue:43716", + "issue: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:45584", + "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:43295|issue:45834", + "issue:43540|issue:44008", + "issue:43618|issue:43697", + "issue:44062|issue:44488", + "issue:43550|issue:43819", + "issue:45183|issue:45479", + "issue:44186|issue:44898", + "issue:43531|issue:43643", + "issue:43653|issue:43819", + "issue:43322|issue:45685", + "issue:43716|issue:44008", + "issue:43335|issue:43756", + "issue:43388|issue:46032", + "issue:43643|issue:44877", + "issue:43950|issue:45987", + "issue:44843|issue:45081", + "issue:30064|issue:36331", + "issue:44448|issue:44671", + "issue:41720|issue:44991", + "issue:44821|issue:45084", + "issue:43421|issue:44811", + "issue:43577|issue:43818", + "issue:43540|issue:45381", + "issue:39692|issue:45834", + "issue:45584|issue:45834" + ], + "prepared_review_unit_hash": "ec3da15150cec9a46618d9e042ecd6a4fa3f9e01b02e5aac0ef096c609213cff", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36828, + "estimated_input_tokens": 9207, + "estimated_eval_tokens": 18670 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose collection of unrelated Transformers bugs that only overlap lexically around tokenizers/processors, loading, and model-specific regressions. The soft edges look like name-similarity rather than true duplicate bugs, so I would not merge any of them.", + "confidence": 0.32, + "canonical_issue_reason": "Issue 44843 is the most central representative of the tokenizer/loading regression subset: it targets a concrete shared code path (_patch_mistral_regex / AutoTokenizer loading), has nearby similar reports, and is broader than the more specific follow-ups.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44843 is the best single issue representative because it sits at the center of the tokenizer-loading/Mistral-regex cluster and is the most generally actionable code-path bug among the items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Unrelated bugs: BitNet accelerate loading vs DAC STE/forward mismatch. Same broad ML area only." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token decoding bug vs BitNet packed-weight loading bug; no shared code path." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction issue vs tokenizer-mapping add-new-model-like failure; different subsystems." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype/SDPA-ish model behavior, but they are model-specific and different failures." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor/tokenizer regression on v4.57.5 vs Kosmos2.5 OCR index error; not the same bug." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Video-input processing error in Qwen3OmniMoe vs Gemma3n variable name collision; distinct code paths." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIPOutput attribute regression vs RTDetrV2 output-difference regression; unrelated model behaviors." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken keyword conflict vs tokenizer loading failure for a specific model; different root causes." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure vs DAC latent/forward mismatch; no common concrete bug." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Input validation messaging bug vs degenerate loss for single-label classification; different issues." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "LayoutLMv2 tokenizer padding/NER crash vs Perceiver non-default resolution failure; not duplicates." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior vs AutoConfig trust_remote_code field loss; unrelated regressions." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBird special-token decode bug vs DAC latent/forward mismatch; no overlap in bug mechanics." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next loading segfault vs MPS histogram int error in moe code; unrelated symptoms and paths." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Mistral-3 dtype mismatch vs Gemma3n audio mask attribute collision; different model-specific bugs." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction bug vs Smollm3 RoPE layer count mismatch; not the same defect." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Metrics gathering truncation bug vs Mamba2 use_cache incorrect results; distinct runtime paths." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "AutoConfig trust_remote_code missing fields vs strict config blocking granite_speech; both loading-related but not same bug." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "Non-persistent buffer corruption on from_pretrained vs StaticCache.get_seq_length returning wrong type; unrelated cache/model-loading bugs." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Both involve _patch_mistral_regex, but one is offline/network access and the other is an attribute error on tokenizer wrapping; not the same concrete bug." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Image processor void segmentation maps vs CustomTrainer compute_loss signature mismatch; completely different areas." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Pegasus output drift across versions vs CamemBERT masked LM prediction errors; different model regressions." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device mapping CUDA assert vs tokenizer loading failure for est-roberta; not the same bug." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading vs template-node compilation error; no shared code path." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Tokenizer backend post-processor update bug vs Whisper batch_decode skip_special_tokens behavior; related to tokenizers, but different functionality." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype propagation bug vs Video-LLaVA temporal-attention/weight-sharing bug; unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Qwen3OmniMoe video processing error vs Qwen2.5-VL video vision_position_ids issue; both video-related, but distinct model/code-path bugs." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 doc example errors vs Kosmos2.5 OCR index error; documentation/model usage issue vs runtime model bug." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty-transcription generation bug vs Kosmos2.5 long OCR input index error; different models and 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": 17, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:28218", + "issue:29159", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:42175", + "issue:43262", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43644", + "issue:43653", + "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:44912", + "issue:44936", + "issue:44977", + "issue:45003", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45372", + "issue:45406", + "issue:45701", + "issue:45834", + "issue:45987", + "issue:46192", + "issue:46209", + "issue:46285", + "issue:46326" + ], + "soft_pairs": [ + "issue:43782|issue:44977", + "issue: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:45987|issue:46192", + "issue:28218|issue:29159", + "issue:44561|issue:45020", + "issue:44610|issue:45072", + "issue:43720|issue:45072", + "issue:44671|issue:45701", + "issue:43825|issue:43827", + "issue:43782|issue:45325", + "issue:43582|issue:44857", + "issue:43749|issue:44291", + "issue:45081|issue:45406", + "issue:42175|issue:45081", + "issue:43906|issue:46209", + "issue:45372|issue:46326", + "issue:43262|issue:43819", + "issue:34634|issue:36331", + "issue:45003|issue:46285", + "issue:43653|issue:44912" + ], + "prepared_review_unit_hash": "e6b69807cf9287ac92cefb3de1369dffe1dc1bdc9eb75eb6eb96530274c96b14", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36290, + "estimated_input_tokens": 9073, + "estimated_eval_tokens": 18402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous issue cluster with several weak thematic similarities. I found one likely true duplicate pair around `remote_code` breakage, while the tokenizer, loading, docs, and model-specific items appear unrelated enough to keep separate.", + "confidence": 0.66, + "canonical_issue_reason": "Issue 44561 is the clearest root-cause report: it names the concrete v5 regression (`is_torch_fx_available` removal) and the affected surface (`trust_remote_code` models), which makes it the best canonical anchor for the remote_code failure mode also described more broadly in 45020.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44561 is the most actionable and specific representative among the cluster items; it pins down the failure mechanism rather than only describing the downstream symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Different model/runtime failures: Qwen3VL weight-only loading vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "Both are tokenizer-loading bugs, but for different tokenizers and failure modes; not the same underlying defect." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "One is non-persistent buffer initialization junk; the other is FSDP CPU RAM efficient loading. Related area, different bug." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Different problems: Bark CPU offload device mismatch vs dtype mismatch in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Unrelated model families and symptoms: Qwen video regression vs Kosmos2.5 OCR index error." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "Both mention pipeline(), but one is an error-message issue and the other is a model-output correctness bug." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Generic tensor shape/index error vs trainer evaluate-after-train failure; no shared code-path evidence." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset bug and GenerationMixin import error are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "BLIP2 dtype propagation and Apple Silicon allocator warmup TypeError are different subsystems." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE discussion and v5 pipeline docs drift are distinct issues." + }, + { + "left": "issue:45987", + "right": "issue:46192", + "accept": false, + "reason": "Cache sequence-length return type and return_tensors handling are separate API bugs." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": false, + "reason": "Both are tokenizer-related, but the titles point to different symptoms; not enough evidence they are the same whitespace bug." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": true, + "reason": "Both describe the same remote_code regression in recent Transformers versions; 44561 gives the concrete cause (`is_torch_fx_available` removal) and 45020 the broader symptom." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Processor image-size mismatch vs dtype mismatch in inference; unrelated failures." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "BitNet packed-weight loading and dtype mismatch issues are different code paths." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression and a general tokenization-change report are not the same bug." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Error message about removed pipeline tasks vs docs still referencing removed tasks; same theme, but not one concrete duplicate bug." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3VL weight-only loading error and Qwen2.5-VL RoPE position-id scaling are separate regressions." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Allocator warmup TypeError and AMP/CUDA loss crash are unrelated." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "FSDP CPU RAM loading breakage and init_empty_weights argument mismatch are distinct loading bugs." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Mistral tokenizer patch crash and Gemma4 processor `_tokenizer` attribute crash are different failures." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Backend-install packaging issue vs tokenizer regex patch crash; no shared underlying bug." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated reproduction of an older issue is not the same as the B200 FP8 inference failure." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Both involve `mistral_common`, but one is an import failure during Gemma 4 loading and the other is a tokenizer behavior bug." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling-rate default and DAC latent reconstruction mismatch are unrelated." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "BarkProcessor voice preset and custom trainer `compute_loss` signature mismatch are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "Unsafe `sys.modules` access and CLIP text-model structure inconsistency are different regressions." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug and MXFP4 quantization fallback are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The cluster is indeed heterogeneous, and the lone accepted pair (`44561`/`45020`) is a plausible same-family `remote_code` regression grounded in the titles. The other soft-edge rejections are framed cautiously enough 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": 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-10000-451", + "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: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:43742|issue:45020", + "issue:44060|issue:44368", + "issue:45084|issue:45538", + "issue:43526|issue:44554", + "issue:43329|issue:45600", + "issue:44589|issue:45042", + "issue:43650|issue:43906", + "issue:43452|issue:44991", + "issue:43638|issue:44936", + "issue:43746|issue:44811", + "issue:43550|issue:44912", + "issue:44387|issue:44617", + "issue:43819|issue:43927", + "issue:36010|issue:43645", + "issue:43257|issue:43329", + "issue:46421|issue:46424", + "issue:43854|issue:44977", + "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:42490|issue:43257", + "issue:43454|issue:43577", + "issue:43262|issue:44811", + "issue:43257|issue:43696" + ], + "prepared_review_unit_hash": "a4115a465f1614ccf4105a96e1228ef39dfdbfa317eeb0c7277b417e7b4c1498", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35717, + "estimated_input_tokens": 8930, + "estimated_eval_tokens": 18116 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated issue reports across different models, loaders, processors, and training paths. None of the soft pairs look like safe duplicate matches, and there is no single clear canonical issue or PR for the set.", + "confidence": 0.95, + "canonical_issue_reason": "No canonical issue is safe here because the items span heterogeneous bugs rather than one underlying defect cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue in this list is a good global representative; the closest-looking pairs still differ in concrete code path and symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "Different models and failures: BEiT label reduction vs RTDetrV2 output divergence." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "Specific MobileLLM key error is not the same as a broad remote_code loading regression." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention Qwen tied-weight warnings, but they describe different model/config paths and symptoms." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Unrelated: a compile/template-node error versus a tokenizer max_length constant issue." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Image label reduction bug and MPS attention correctness bug are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Multimodal token counting bug is unrelated to stale auto_mappings config references." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to an unwanted torchvision dependency." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "A placeholder issue title is not the same as a reproduction of another bug." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading-related, but one is gguf_file/from_pretrained handling and the other is a specific tokenizer load failure." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Different trainer failures: Deepspeed ZeRO3 BERT indexing vs evaluate-after-train behavior." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT checkpoint loading and Whisper batch_decode token filtering are unrelated code paths." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile/SDPA failure and MXFP4 quantization fallback are different issues." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention OOM, but one is int4 quantization memory regression and the other is a model-specific video OOM." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "DAC latent mismatch and DiaConfig token-ID persistence are unrelated bugs." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Import path breakage is not the same as Jupyter notebook custom-model initialization failure." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is unrelated to multimodal token counting." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both concern cache behavior, but one is recompilation from lazy init and the other is max_cache_length handling for quantized cache only." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms: GLM-4.7-Flash loading tests vs Qwen3.5 flash-attention generation." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "BitNet packed-weight loading and OmDet-Turbo image sizing are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 zero probabilities and OmDet-Turbo processor/model size mismatch are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Documentation example errors for SigLIP2 do not match the MXFP4 quantization loading bug." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "ImportError for GenerationMixin and unexpected compute_loss kwarg are different API breakages." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to the torchvision requirement regression." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both involve exception handling, but they are different functions and different failure mechanisms." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Trailing-slash local path warning is unrelated to Qwen3 MoE weight conversion." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision weight-tying failure and BLIP2 dtype casting are distinct bugs." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sampling-rate defaults are unrelated to Whisper batch_decode skip_special_tokens handling." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43329", + "issue:43450", + "issue:43452", + "issue: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:34567|issue:40990", + "issue:43606|issue:45859", + "issue:43749|issue:44756", + "issue:29159|issue:34567", + "issue:44568|issue:45356", + "issue:43257|issue:43931", + "issue:42915|issue:43742", + "issue:44291|issue:45042", + "issue:42915|issue:44821", + "issue:44987|issue:45538", + "issue:43452|issue:43582", + "issue:43927|issue:44912", + "issue:43697|issue:45357", + "issue:42915|issue:45538", + "issue:44849|issue:45701", + "issue:44568|issue:45488", + "issue:33357|issue:34567", + "issue:41720|issue:43540", + "issue:42175|issue:43824", + "issue:43317|issue:45538", + "issue:44479|issue:45381", + "issue:44062|issue:44991", + "issue:43540|issue:44479", + "issue:28282|issue:34689", + "issue:43540|issue:44560", + "issue:45137|issue:45405", + "issue:43742|issue:44291", + "issue:24643|issue:29159", + "issue:44987|issue:45907" + ], + "prepared_review_unit_hash": "043bb84f30b681794768465644b9dad79f5991e8ee639289f7066ad456c960c2", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37330, + "estimated_input_tokens": 9333, + "estimated_eval_tokens": 18922 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous set of closed Transformer bugs spanning tokenizer regressions, model-loading failures, multimodal/video processing, offload, and quantization. I found one likely true duplicate-style soft edge in the Qwen video regression area; the rest are distinct issues.", + "confidence": 0.84, + "canonical_issue_reason": "issue:45701 is the broadest umbrella item in the cluster and best matches the tokenization-regression theme shared by several nearby reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the most representative issue overall because it is broad, user-facing, and central to the tokenization/version-regression subset.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Different APIs and failures: missing fields from AutoConfig with trust_remote_code vs text_config unwrapping in AutoModelForCausalLM.from_config." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": false, + "reason": "Both are video-related, but one is an undefined-name bug in _get_num_multimodal_tokens and the other is a broader 5.3.0 regression across Qwen video models." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Undefined helper names in video token counting vs batched video processor shape errors; different code paths." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Unrelated topics: TrainerState token counter update vs high perplexity on gpt-oss-20b." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both involve offload, but one is a bark-small device mismatch and the other is a Qwen rotary-embedding CPU-offload incompatibility." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Different problems: FSDP CPU RAM efficient loading vs mmap OOM on Strix Halo." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer inconsistency vs TrainerState token-count tracking; not the same bug." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Separate tokenizer regressions affecting different models and failure modes." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed vs Qwen3-VL weight-shape mismatch on load." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different load-time failures on different models: FineGrainedFP8Config vs a MobileLLM key error." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument during init_empty_weights is unrelated to PIL backend requiring torchvision." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Unrelated: Qwen3Moe FP8 config failure vs AutoImageProcessor URL loading." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Different issues: loading physical-intelligence/fast vs CLIPTokenizer model_max_length being 10**30." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "gguf_file loading breakage is unrelated to the Apple Silicon caching_allocator_warmup TypeError." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Different failure modes: DiaConfig custom token IDs persistence vs MXFP4 quantization fallback on gpt-oss-20b." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetrV2 output mismatch and Qwen3.5 save_pretrained key regression are separate bugs." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "No overlap beyond both being tokenizer/model issues; one is Qwen3Moe FP8, the other CLIPTokenizer max length." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 output_hidden_states bug is distinct from the broad tokenization regression report." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both tokenizer-related, but the BOS/EOS special-token bug and the LlamaTokenizer ByteLevel/Metaspace override are different code paths." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error in CLIP model loading is unrelated to TrainerState token tracking." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different Qwen failures: auto device mapping cudaErrorAssert vs video-input ValueError in Qwen3OmniMoe." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Backend packaging issue for TensorFlow is unrelated to the Qwen2.5-VL import error." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Device-map/offload loading failure is unrelated to CLIPTokenizer max_length behavior." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "These describe the same Qwen video-input regression family around 5.3.0; 45381 is a specific qwen2.5-vl manifestation of the broader multi-model bug in 44479." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer-loading TypeErrors with different roots and models." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Both touch Qwen video inputs, but one is a Qwen3OmniMoe processing error and the other is a broader v5.3.0 regression across several Qwen models." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "ImportError for missing PyTorch is unrelated to Llama 3.2 90B Vision model-loading breakage." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Different video bugs: a ValueError in Qwen3OmniMoe processing vs a StopIteration in Qwen3-VL embedding video handling." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque underflow is unrelated to the PEFT minimum-version bump." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "Different loading failures: MobileLLM key error vs init_empty_weights/_is_hf_initialized TypeError." + }, + { + "left": "issue:24643", + "right": "issue:29159", + "accept": false, + "reason": "Training with DeepSpeed weight-shape error is unrelated to tokenizer slow/fast inconsistency." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Loading physical-intelligence/fast is unrelated to list_repo_files needing to forward revision." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet, and the only accepted soft edge (44479 \u2194 45381) is conservative and supported by the titles. The rejected pairs are generally reasoned at the level of distinct APIs, models, or failure modes, without overstating duplication evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "nodes": [ + "issue:27957", + "issue:28282", + "issue:29127", + "issue:34567", + "issue:36331", + "issue:38175", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43404", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43531", + "issue:43577", + "issue:43582", + "issue: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:28282|issue:29127", + "issue:43577|issue:45072", + "issue:42915|issue:43749", + "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:44291|issue:44821", + "issue:34567|issue:36331", + "issue:43526|issue:45479", + "issue:43404|issue:45987", + "issue:44062|issue:45081", + "issue:27957|issue:28282", + "issue:43756|issue:46399", + "issue:44913|issue:46121", + "issue:44792|issue:44964", + "issue:43906|issue:45381", + "issue:43756|issue:45325", + "issue:45042|issue:45081", + "issue:43742|issue:44589", + "issue:43452|issue:44488", + "issue:45357|issue:45920", + "issue:43531|issue:44821", + "issue:45405|issue:45600", + "issue:43643|issue:45375" + ], + "prepared_review_unit_hash": "a30be84874f22047c547fcb2d59678d4e8d490b1feb05e4dc3017ef795c05bea", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36187, + "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 and not a duplicate cluster. The only strong soft-duplicate pair is the SigLIP2 regression reports (43493/43994); the rest describe different models, different errors, or different code paths.", + "confidence": 0.34, + "canonical_issue_reason": "issue:43493 is the best anchor for the one real duplicate pair because it describes the underlying SigLIP2 implementation discrepancy at a root-cause level, not just a symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43994 is the most actionable issue overall: it names the exact model, the user-facing failure mode, and the affected entry points (AutoModel and pipeline).", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Different bugs: PyTorch missing import vs. LayoutLMv3 error-message clarity." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype, but they are different models and different failures; not the same concrete code-path bug." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "Unrelated: Qwen3 MoE FP8 loading vs. FSDP CPU RAM efficient loading." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Different vision-model failures with different symptoms and likely different model code paths." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Backend dependency packaging issue vs. a model-loading failure for a specific checkpoint." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Different Qwen/Smollm model bugs; one is load/conversion, the other is RoPE-layer dropping." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Apple Silicon allocator TypeError vs. Qwen3VL weight_only load error." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Same underlying SigLIP2 regression: both report bad outputs for google/siglip2-base-patch16-224, one as HF-vs-JAX discrepancy and the other as nonsensical AutoModel/pipeline results." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU placement bug vs. flash-attention sinkless-model AttributeError." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized TypeError is unrelated to AutoImageProcessor URL loading." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "Trainer state token-count tracking vs. custom compute_loss signature breakage; different trainer subsystems." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BEiT image label reduction bug vs. sequence-classification zero-loss bug; not the same issue." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "Mistral3 lm_head tying bug vs. StaticCache seq-length return-type bug." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer AddedToken duplication vs. mistral regex patch crashing on a missing backend_tokenizer attribute." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Flash Attention 2 support for XLM-Roberta vs. missing PyTorch ImportError." + }, + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Different model-family RoPE/kernel bugs with different failure modes." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "GPTNeoX config reload regression vs. rope-validation list TypeError; related area but different bug." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Janus test failure vs. Phi-4 multimodal load failure; different models and symptoms." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen2.5-VL related reports with different reproduction paths and affected behavior." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both involve positional encodings, but they affect different models and different code paths." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Image-processor torchvision dependency regression vs. mistral tokenizer regex crash." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "MobileLLM key error vs. Float8 storage lookup failure are unrelated loading problems." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf_file support breakage for tokenizer/model loading vs. a specific checkpoint load failure." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Qwen3.5 save_pretrained visual-key regression vs. AutoTokenizer wrong token IDs for unrelated models." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding-window bug vs. AutoImageProcessor URL loading issue." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "PEFT version pinning problem vs. broken auto_mappings reference to removed configs." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Generic trust_remote_code config-field loss vs. a specific Qwen3_5MoeVisionConfig field-drop bug; not the same underlying change." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: most pairs are clearly different models, errors, or code paths, and the lone accepted SigLIP2 pair is the only plausible soft duplicate. The verdicts stay conservative overall, with no obvious overclaim beyond what the titles support." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-10000-451", + "nodes": [ + "issue:24643", + "issue:27957", + "issue:28282", + "issue:29127", + "issue:29159", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:38175", + "issue:41762", + "issue:43295", + "issue:43452", + "issue:43577", + "issue: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: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:43452|issue:44843", + "issue:30064|issue:33666", + "issue:43688|issue:43749", + "issue:28282|issue:32090", + "issue:45372|issue:46170", + "issue:43819|issue:44811", + "issue:44936|issue:45823", + "issue:43906|issue:46202", + "issue:28282|issue:33357", + "issue:43295|issue:45356", + "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:44464|issue:44743", + "issue:43577|issue:43697", + "issue:29159|issue:33357" + ], + "prepared_review_unit_hash": "84e9592430723f51acb15c241bc3f875eee529089a6adf01d70b0249ed35b56c", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36660, + "estimated_input_tokens": 9165, + "estimated_eval_tokens": 18586 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous: the soft pairs mostly share broad vocabulary (tokenizers, configs, loading, training, serialization) but do not describe the same underlying bug or change. No duplicate edges should be accepted.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out because the items are not one duplicate set; they span unrelated regressions across tokenization, loading, training, and serialization.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43749 is the strongest standalone report in the set, with clear impact and the richest discussion, but it is not a representative duplicate umbrella for the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-related, but one is Qwen3.5 num_labels propagation and the other is a strict-config load failure for granite_speech; different model families and failure modes." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "BitNet packed-weight loading and Perceiver positional-encoding interpolation are unrelated code paths and bugs." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "One is a v5 output-difference regression in RTDetrV2; the other is broken checkpoint saving for Qwen3.5. Same release window, different underlying issue." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "DeepSpeed training shape error vs. a request to improve a LayoutLMv3 error message; not the same bug." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Gemma3 loading with ZeRO-3 and trainer.evaluate() after train() are different Trainer/DeepSpeed failures." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both involve Qwen2.5-VL, but one is a batched chat-template crash and the other is wrong vision_position_ids for video input; distinct defects." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Unexpected position-id keys during model loading is not the same as sinq quantization breaking save/load roundtrips." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights/serialization, but one is device-dependent lm_head serialization and the other is translation-model tied-weight issues; too broad to merge." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "gguf_file loading failure and BLIP2 dtype propagation are unrelated loading bugs." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Both touch from_pretrained loading, but gguf_file breakage and unconditional model_info() in _patch_mistral_regex are different code paths." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Void segmentation-map processing and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Aux-loss normalization in OLMoE/GPT Oss is unrelated to FSDP CPU-RAM-efficient loading." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "PyTorch import detection failure and a NoneType broadcast error in Trainer are different problems." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Gemma 4 processor loading depends on mistral_common's ReasoningEffort import, while the other is a Windows/ROCm torch.distributed dependency issue." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC STE mismatch and Whisper batch_decode skip_special_tokens behavior are unrelated." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Trainer.evaluate() after train() and Gemma4 PLE device mismatch under device_map='auto' are different failures." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "An isolated reproduction of another issue is not the same as a Gemma pipeline roundtrip bug." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Missing PyTorch import and a MacOS bus error in a CLIP model are unrelated." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "One is a processor/tokenizer regression around custom model code; the other is a Kimi-K2.5 tokenizer codec/mistral-regex regression." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizer failures on different models; not the same underlying tokenizer bug." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "BigBird special-token registration and MLukeTokenizer task failures are unrelated tokenizer bugs." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard-size question and AutoProcessor kwargs forwarding are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Zero-probabilities in SigLIP2 and a processor.tokenizer regression are different model behaviors." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Both are backend-dependency errors, but tensorboard-without-TensorFlow and tensorflow_text mapping failures are different issues." + }, + { + "left": "issue:24643", + "right": "issue:27957", + "accept": false, + "reason": "DeepSpeed weight-shape error and XLMRoberta FlashAttention2 support are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "BigBird mask-token registration and mDeBERTa BOS/EOS add_special_tokens regression are different tokenizer special-token bugs." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Chunked generation with compiled forward and recurrent-state reset with cache are separate generation/cache problems." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "BLIP2 dtype loading and RTDetrV2 output divergence are unrelated." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency and a MacOS CLIP bus error are not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:34567", + "issue:36010", + "issue:41720", + "issue:43262", + "issue:43295", + "issue:43452", + "issue:43504", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43927", + "issue:44361", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44514", + "issue:44554", + "issue:44561", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44843", + "issue:44977", + "issue:44991", + "issue:45290", + "issue:45440", + "issue:45491", + "issue:45588", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46002", + "issue:46050", + "issue:46209", + "issue:46365", + "issue:46402" + ], + "soft_pairs": [ + "issue:44843|issue:44991", + "issue:44514|issue:45290", + "issue:43531|issue:43742", + "issue:43646|issue:43749", + "issue:43756|issue:46402", + "issue:43688|issue:45704", + "issue:43688|issue:45440", + "issue:43653|issue:43927", + "issue:45491|issue:46002", + "issue:43550|issue:43927", + "issue:43688|issue:45701", + "issue:44479|issue:44625", + "issue:44464|issue:44977", + "issue:43550|issue:44361", + "issue:43643|issue:44561", + "issue:41720|issue:44464", + "issue:32090|issue:34567", + "issue:44977|issue:46209", + "issue:43577|issue:44554", + "issue:43577|issue:45588", + "issue:44466|issue:46050", + "issue:32090|issue:33666", + "issue:43504|issue:43720", + "issue:33666|issue:36010", + "issue:43504|issue:43550", + "issue:43504|issue:44610", + "issue:43742|issue:44821", + "issue:44554|issue:46365", + "issue:43720|issue:43927", + "issue:43262|issue:43295", + "issue:43452|issue:44451", + "issue:43606|issue:43927", + "issue:45823|issue:46050" + ], + "prepared_review_unit_hash": "35f52b83821daadacd5f48652e8c6a7ee0a463a68152f764ee0c12155144ddfd", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37603, + "estimated_input_tokens": 9401, + "estimated_eval_tokens": 19058 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most items are unrelated bugs across loading, tokenization, training, generation, and model-specific processor paths. The soft pairs look like surface-level similarity rather than the same underlying fix, so none should be merged.", + "confidence": 0.95, + "canonical_issue_reason": "No clear canonical issue: the items span different subsystems and concrete failure modes, so there is no single bug that plausibly represents the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong best issue anchor: even the closest-looking pairs are distinct enough that picking one representative issue would be misleading.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both involve tokenizer loading, but one is an offline Mistral regex/hub call bug and the other is a specific tokenizer load failure for EMBEDDIA/est-roberta; different code paths." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Both crash in apply_chat_template, but one is batched padding=False handling and the other is assistant tool-call messages with no content; distinct edge cases." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM key-error loading are unrelated failures." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model init breakage and FSDP CPU RAM efficient loading are different failure modes, not one shared bug." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Smollm3 RoPE-layer handling and Sam3 text encoder weight loading are separate model-specific issues." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization in MoE models and a T5 RMSNorm memory leak are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Different model families and different defects: auxiliary-loss normalization vs remote/native implementation divergence." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Tokenizer special-token registration vs config token-ID persistence after save/load are different bugs." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both mention NaNs, but the causes differ: Gemma3 sliding-window attention vs classification head init/checkpoint behavior." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "torch.compile SDPA failure in Bamba and DiaConfig token-ID loss are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "MoE auxiliary-loss normalization and tokenization version changes are not the same issue." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "A video-input regression in Qwen vision-language models and num_labels propagation in Qwen3.5 are different bugs." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Chunked generation inconsistency with compiled forward is not the same as a flash-attention generation regression." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Bamba SDPA/compile failure and MLukeTokenizer task AttributeError are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both involve trust_remote_code, but one is missing fields in AutoConfig output and the other is removal of a helper breaking remote-code models; not one concrete bug." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "CUDA assert on auto device mapping and chunked-generation inconsistency are different code paths and symptoms." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer broadcast TypeError and num_input_tokens_seen not updating are separate Trainer bugs." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Flash-attention generation issues in Qwen3.5 and FP8 inference failures on B200 are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype loading and an MPS attention correctness issue do not share a concrete fix path." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "A dtype-loading bug and a flash-attention s_aux=None crash are distinct." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Serialization of tied lm_head weights and sinq quantized save/load breakage are different bugs." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Generic Trainer _gpu_broadcast TypeError and Qwen2-VL multi-GPU training are too broad to treat as the same bug." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy field loading in Beit and BitNet packed-weight unpacking during accelerate loading are different root causes." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Multi-GPU Qwen2-VL training and a GenerationMixin import error are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Different models and failures: Beit preset loading vs Bamba SDPA/compile behavior." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Beit legacy-field loading and OmDet-Turbo processor size mismatch are unrelated." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "Loading a specific model with a key error and loading AutoImageProcessor from URL are different loading paths." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "An MPS attention correctness bug and a Trainer import issue caused by datasets are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet packed-weight loading and DiaConfig token-ID persistence are not the same bug." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and processor.tokenizer regression are different processor bugs." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "gguf_file loading failures and inability to load a specific tokenizer/model are not the same underlying issue." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU-offload device mismatch in bark-small and DiaConfig token-ID loss are unrelated." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Gemma4 PLE device mismatch and sinq save/load quantization breakage are different defects." + } + ] + }, + "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-10000-451", + "nodes": [ + "issue:28218", + "issue:33357", + "issue:33666", + "issue:36331", + "issue:39692", + "issue:43257", + "issue:43278", + "issue:43295", + "issue:43299", + "issue:43334", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43540", + "issue:43572", + "issue:43606", + "issue:43646", + "issue:43688", + "issue: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:33666|issue:36331", + "issue:43994|issue:46192", + "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:44964|issue:45405", + "issue:39692|issue:43493", + "issue:44849|issue:45357", + "issue:44568|issue:45812", + "issue:44464|issue:46424", + "issue:43493|issue:43844", + "issue:43540|issue:44977", + "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:45759", + "issue:43257|issue:43299", + "issue:44589|issue:45520", + "issue:43824|issue:44589", + "issue:43927|issue:44361", + "issue:43504|issue:43606", + "issue:45229|issue:45823", + "issue:43572|issue:44291" + ], + "prepared_review_unit_hash": "516889d22131dbf2fa51054fe07d354c468648fb8134919635bd66506cbff2d0", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37122, + "estimated_input_tokens": 9281, + "estimated_eval_tokens": 18818 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These candidates are not a duplicate cluster: they span unrelated tokenizer, model-loading, training, config, inference, and documentation problems across many model families. I would not merge any soft pair.", + "confidence": 0.96, + "canonical_issue_reason": "No canonical issue: the items are heterogeneous and do not share one underlying bug or change.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is a good global representative because the set is broad and mostly unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Both are training-related, but one is Qwen2-VL multi-GPU usage and the other is a CustomTrainer signature regression; different bugs and code paths." + }, + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "One is a SigLIP2 model output-quality issue, the other is a preprocessing argument being ignored; no shared underlying failure." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both mention missing pad-token fields, but they affect different config classes and different attributes (`pad_token_id` vs `pad_token_idx`)." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Qwen3.5 hidden-state behavior and FP8 inference on B200 are unrelated issues." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Embedding dtype drift at eval and auxiliary-loss normalization are separate model-behavior bugs." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Packed-weight loading in BitNet and a tokenizer/task AttributeError are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "One is config propagation for Qwen3.5, the other is NaNs in classification heads; different symptoms and paths." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and a pipeline error-message regression are unrelated." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing behavior and a MacOS bus error in a CLIP model are not the same bug." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Model loading failure and a dependency-version bump are not duplicates." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "One is a docs/example page with multiple errors, the other is a core SigLIP2 implementation mismatch; not one concrete fix." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Hidden-states output bug and incorrect save_pretrained key serialization are different regressions." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and different tokenization behaviors." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Chunked generation with compiled forward and cache-length handling for quantized cache are unrelated." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "A model-implementation discrepancy and a ZeRO-3 gradient anomaly are separate issues." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Video-input processing in Qwen3OmniMoe and Qwen3.5 flash-attention generation bugs do not share a concrete code path." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Text-only mm_token_type_ids handling and Gemma4 multi-GPU OOM are unrelated." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Model load failure for a specific checkpoint and an offline-mode bug in AutoTokenizer are different problems." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Docs/example mistakes and a regression around `processor.tokenizer` access are not the same underlying bug." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "A dtype change during evaluation and abnormal gradient growth under ZeRO-3 are distinct training issues." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Both involve missing pad-token config fields, but they are model-specific config regressions with different attributes and loaders." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Custom model initialization breakage and device-dependent serialization of `lm_head.weight` are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "A bus error on MacOS and Qwen2-VL multi-GPU training are not the same failure mode." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Missing `pad_token_id` in Qwen3-VL and failure to unwrap `text_config` in composite Qwen 3.5/3.6 configs are different bugs." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Different Qwen3 MOE loading regressions with different triggers and symptoms." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "An unsupported float8 storage type and a flash_attn import KeyError on Python 3.13 are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Importing Qwen2.5-VL and float8 storage lookup failures are completely separate." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Config save/load token-ID loss and MLukeTokenizer task AttributeErrors are not the same bug." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Legacy field handling in Beit loading and a CPU offload device mismatch in bark-small are unrelated." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 OOM and Gemma4 device-map mismatch are both in the same model family but are different concrete failures." + }, + { + "left": "issue:45357", + "right": "issue:45759", + "accept": false, + "reason": "Incorrect visual-encoder keys on save and missing `text_config` unwrapping on from_config are different composite-model regressions." + }, + { + "left": "issue:43844", + "right": "issue:43688", + "accept": false, + "reason": "Gradient explosion under ZeRO-3 and auxiliary-loss normalization are distinct bugs." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "A missing config field after v5 migration and an unexpected `_is_hf_initialized` argument in init_empty_weights 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": 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-10000-451", + "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:43575|issue:45405", + "issue:44493|issue:46285", + "issue:44792|issue:45405", + "issue:43504|issue:43653", + "issue:43606|issue:44112", + "issue:43994|issue:44493", + "issue:43873|issue:46050", + "issue:45588|issue:46050", + "issue:43065|issue:46369", + "issue:44423|issue:44734", + "issue:29942|issue:30333", + "issue:42490|issue:43696", + "issue:45406|issue:46121", + "issue:43844|issue:44964", + "issue:28218|issue:34567", + "issue:29159|issue:32090", + "issue:45084|issue:45520", + "issue:44554|issue:45446", + "issue:43582|issue:44977", + "issue:15354|issue:38175", + "issue:43540|issue:45200", + "issue:45072|issue:45198", + "issue:45200|issue:45823", + "issue:44521|issue:45290", + "issue:37428|issue:43957" + ], + "prepared_review_unit_hash": "0878fbbede52490904ac9fa4ad05df24b3c95b20651a9e412256ffd205a0913d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35579, + "estimated_input_tokens": 8895, + "estimated_eval_tokens": 18046 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it spans unrelated tokenizer, model-loading, serving, attention, quantization, and CI issues. None of the soft pairs look like true duplicates of the same underlying bug/change, so all soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue stands out; the cluster is too broad and the soft pairs do not share a common root cause.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the strongest standalone bug report here because it is open, has the most inbound attention, and describes a concrete, reproducible multimodal chat-template failure.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Different problems: embedding dtype drift during eval vs resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "OOM during tensor-parallel model load is unrelated to a PEFT version bump." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Unexpected position-id key behavior is not the same as CLIP text-model structural inconsistency." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure and PEFT minimum-version issue are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy preset loading for BEiT segmentation is a different bug from BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "Both mention device behavior, but one is a CPU-offload mismatch and the other is a CI test regression." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "Siglip2 bad outputs and unexpected position-id keys are different model-path failures." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantization/offloading behavior is not the same as broken save_pretrained/from_pretrained after sinq quantization." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention crash on sink-less models is unrelated to quantized model serialization." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "Dummy Conv2d placeholder in Sam3PixelDecoder is a different issue from a memory overflow in sam3-litetext." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Two separate serve crashes with different symptoms and code paths: string .to() error vs KV-cache tensor indexing." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures and MLflow reporting on failed jobs are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Import warning for trailing-slash model paths is unrelated to GPT-oss CUDA OOM." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor missing _tokenizer is not the same as rope-parameter type conversion failing on a JSON list." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient explosion and Phi-4 multimodal loading failure are different bugs." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Added-token spacing in tokenizers is unrelated to TrainerState token-count tracking." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Slow/fast tokenizer inconsistency does not match a Trainer broadcast NoneType error." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Non-template-node compilation failure is unrelated to flash_attn import key errors on Python 3.13." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness and a flex_attention PyTorch version check are distinct issues." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError is unrelated to Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export error and Siglip2 zero-probability outputs are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video-input processing failure in Qwen3OmniMoe is different from Gemma4 mm_token_type_ids defaults." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "BF16 dtype mismatches in inference are not the same as Wav2Vec2 save/tokenization failures." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Gemma4 token-type-id defaults and Gemma4 PLE device mismatch are separate model bugs." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both touch apply_chat_template, but one is all-zero assistant masks for multimodal inputs while the other is a crash on tool-call messages without content." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Flash-support import error and meta-device loading regressions are different issues with different triggers." + } + ] + }, + "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-10000-451", + "nodes": [ + "issue:28218", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43950", + "issue:43957", + "issue:43975", + "issue:44186", + "issue:44466", + "issue:44554", + "issue:44610", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44987", + "issue:45127", + "issue:45200", + "issue:45325", + "issue:45356", + "issue:45381", + "issue:45412", + "issue:45520", + "issue:45563", + "issue:45823", + "issue:46178", + "issue:46192", + "issue:46209", + "issue:46285" + ], + "soft_pairs": [ + "issue:43493|issue:43701", + "issue:44186|issue:46192", + "issue:43232|issue:44849", + "issue:45823|issue:46192", + "issue:43295|issue:43450", + "issue:45381|issue:46209", + "issue:43295|issue:44610", + "issue:43572|issue:44987", + "issue:39692|issue:46192", + "issue:43295|issue:43873", + "issue:44554|issue:45356", + "issue:43572|issue:43957", + "issue:45520|issue:46178", + "issue:37428|issue:43824", + "issue:44466|issue:45127", + "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:42490|issue:43493", + "issue:42915|issue:43824", + "issue:43824|issue:43975", + "issue:44792|issue:45563", + "issue:43388|issue:43526", + "issue:28218|issue:35141", + "issue:38175|issue:43493", + "issue:43278|issue:43856" + ], + "prepared_review_unit_hash": "bad98418b19254e26236040eecb911b778cb7d52f2b350f553baa8137f62a8c5", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36717, + "estimated_input_tokens": 9180, + "estimated_eval_tokens": 18616 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose mix of unrelated Transformers bug reports across tokenization, multimodal processing, model loading, quantization, and generation. None of the soft pairs look like true duplicates; they mostly share broad subsystem vocabulary rather than the same concrete bug. The most representative issue is issue:43295 because it is a central regression report with several nearby similarity links, but it is only a cluster representative, not a duplicate of the others.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43295 is the most central representative in this set: it is a clear regression report, touches a user-facing multimodal code path, and has multiple similarity links to other processor/model-loading issues. It is still not a duplicate of the rest, just the best cluster anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43295 is the best single issue representative because it sits near several soft-linked reports and describes a concrete regression affecting custom multimodal model code.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "Different bugs: SigLIP2 implementation discrepancy vs resume_from_checkpoint key mismatch." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "Different failures: LayoutLMv2 tokenizer crash on NER/padding vs return_tensors being ignored when text_kwargs is passed." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Different generation bugs: sync_gpus kwargs handling vs Qwen3.5 output_hidden_states behavior." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Different subsystems: Gemma4 device_map/PLE device mismatch vs tokenizer argument handling in processors." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Both involve multimodal code, but one is processor.tokenizer regression and the other is batched video processor shape output; not the same bug." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Different issues: Qwen2.5-VL vision_position_ids for video input vs FP8 inference failure on B200." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Custom processor/tokenizer regression is unrelated to OmDet-Turbo processor producing the wrong image size." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Both are load-time regressions, but one is missing pad_token_idx in StableLmConfig and the other is a model-loading failure for physical-intelligence/fast." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "SigLIP2 doc/example errors are unrelated to the tokenizer return_tensors bug." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Custom processor/tokenizer breakage is unrelated to quantization offloading not working as expected." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "Different model-specific correctness bugs: MPS attention value-head issue vs Kimi-K2.5 tokenizer regression." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Both are loading-related, but StableLmConfig pad token metadata and meta-device loading failures are distinct problems." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "Different import/backend problems: flash_attn import on Python 3.13 vs tensorflow_text backend mapping error." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for completely different missing symbols and code paths." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Different weight-serialization bug vs LoRA merge collapse with extended vocabulary." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption in from_pretrained is unrelated to CLIP class structure inconsistency." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver non-default resolution failure is unrelated to RT-DETR memory release behavior." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "Different training/runtime issues: torch.compile SDPA failure on Bamba vs gradient explosion under ZeRO-3." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Different multimodal position/defaulting bugs in different models; one is mm_token_type_ids defaults, the other is still-image temporal position_ids scaling." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Same symptom style, but different model families and code paths: Qwen3-VL vs Llama4Config pad_token_id." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Video processor shape bug and Qwen2-57B-A14B-instruct OOM are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Import warning for local paths ending with slash is unrelated to SigLIP2 implementation discrepancy." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Qwen3 MoE fine-grained FP8 failure is unrelated to Qwen2.5-VL import error." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "ImportError for a missing class and detokenization regression are different bugs." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Failed janus image-generation test is unrelated to a stale generate() warning." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both concern label handling, but one is gather_for_metrics truncation with tuple labels and the other is BeitImageProcessorFast reduce_labels behavior." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer extra-space detokenization bug is unrelated to embedding reinitialization after resize." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is not the same as the HF-vs-JAX discrepancy report, even if related to the same model family." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype changing to FP32 in eval is unrelated to Qwen3 MoE memory inefficiency." + } + ] + }, + "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": 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-10000-451", + "nodes": [ + "issue:29942", + "issue:32090", + "issue:37428", + "issue:38175", + "issue:41720", + "issue:43262", + "issue:43335", + "issue:43425", + "issue:43450", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43827", + "issue:43901", + "issue:44112", + "issue:44206", + "issue:44351", + "issue:44442", + "issue:44448", + "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:46209", + "issue:46234", + "issue:46285", + "issue:46326", + "issue:46402" + ], + "soft_pairs": [ + "issue:41720|issue:46209", + "issue:43653|issue:44112", + "issue:43526|issue:46192", + "issue:44442|issue:45198", + "issue:38175|issue:45412", + "issue:45290|issue:46326", + "issue:43262|issue:43450", + "issue:45520|issue:46170", + "issue:45759|issue:46285", + "issue:44610|issue:44857", + "issue:44561|issue:46234", + "issue:44514|issue:44521", + "issue:43819|issue:44112", + "issue:43827|issue:43901", + "issue:44206|issue:45584", + "issue:43819|issue:45083", + "issue:44857|issue:45593", + "issue:43335|issue:43550", + "issue:43611|issue:46402", + "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:44855|issue:45083", + "issue:37428|issue:44351", + "issue:43550|issue:44112" + ], + "prepared_review_unit_hash": "8df2be93658c732dfdb26259f0f020e777f698f619a69b91cd80bc51eaa08c08", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36621, + "estimated_input_tokens": 9156, + "estimated_eval_tokens": 18568 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a broad false-positive mix of unrelated Transformers issues: chat templating, tokenizer/model loading, vision/audio processors, import/platform regressions, and a few model-specific training/inference bugs. None of the soft pairs look like the same underlying bug or a mergeable PR-equivalent change.", + "confidence": 0.95, + "canonical_issue_reason": "No safe canonical issue: the apparent similarities are mostly superficial (shared APIs or broad subsystem names) and do not converge on one underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the strongest standalone report in the set: it has the most discussion/inbound references and a clear, user-facing repro for multimodal assistant masks. It is representative, but it should not absorb the other issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "Both are GPU-related, but one is an auto device-mapping cuda assert on A800 and the other is FP8 inference failure on B200; different code paths and root causes." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token decode bug vs a stale GraniteSpeech CI device-override test; different components and failure modes." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug vs return_tensors being ignored when text_kwargs is passed; unrelated processor behaviors." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "AutoTokenizer load failure for FastSpeech2ConformerTokenizer vs wav2vec2 save_pretrained/tokenization failure; different tokenizer families and bugs." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 vs RT-DETR memory not released on delete; not the same underlying issue." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both involve apply_chat_template-like flows, but one crashes on assistant messages with tool calls and no content while the other is about continue_final_message for mistral_common tokenizers." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio processor sampling-rate default bug vs video processor batched-shape bug; different modalities and code paths." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Python 3.13 flash_attn import KeyError vs Windows ROCm torch.distributed dependency issue; unrelated platform/import regressions." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Qwen multimodal config unwrapping bug vs CLIP text-model structure inconsistency; different model families and different fixes." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "OmDet processor size mismatch vs LwDetrImageLoss AMP float16 crash; both are vision bugs but not the same defect." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Removal of is_torch_fx_available breaking trust_remote_code models vs missing is_falcon_mamba_ssm_available export; different missing symbols." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Batched apply_chat_template crash with padding=False vs all-zero assistant_masks for multimodal inputs; related API area but different symptom and trigger." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC.from_latents / STE mismatch vs GraniteSpeech stale device-override CI failure; unrelated bugs." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues, but they cover different deprecated APIs (pipeline removal vs return_all_scores); not duplicates." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Unsupported center arg in LasrFeatureExtractor vs Whisper empty-transcription behavior after align_special_tokens; different audio bugs." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC latent reconstruction mismatch vs qwen3_omni_moe feature-length helper behavior; no shared underlying code path." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Float16 AMP crash in LwDetrImageLoss vs D-FINE not using auxiliary losses when denoising is off; different training logic." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformersConfig sparse-layer creation bug vs Bamba torch.compile SDPA failure; different models and fixes." + }, + { + "left": "issue:43611", + "right": "issue:46402", + "accept": false, + "reason": "Base-model-prefix loading break in Transformers 5.0 vs Sam3Model text encoder weights not loading; both are loading issues but not the same one." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "Missing _flash_supports_window_size import vs broad Torch 2.10 incompatibility; one specific import regression, one version-compatibility report." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU-offload device mismatch in suno/bark-small vs qwen3_omni_moe output-length helper bug; unrelated subsystems." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures vs Trainer _gpu_broadcast_one NoneType TypeError; different failures and code paths." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output difference vs Qwen3.5 recurrent-state reset with cache; different model families and root causes." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Gemma 4 mm_token_type_ids defaulting bug vs flash_attention integration crash on s_aux=None; not the same issue." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA merge collapse with extended vocab/tied embeddings vs a category-count limit error; unrelated failure modes." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug vs BigBirdTokenizer special-token registration bug; different components entirely." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 @torch.jit.script parsing error in DebertaV2Model vs qwen3_omni_moe helper behavior; no relation." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Two import errors, but for different missing symbols (_flash_supports_window_size vs HybridCache); not duplicates." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba torch.compile SDPA failure vs GraniteSpeech stale CI device override; unrelated bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-10000-451", + "nodes": [ + "issue:28218", + "issue:29127", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:38175", + "issue:39692", + "issue:40444", + "issue:42490", + "issue:43278", + "issue:43335", + "issue:43504", + "issue:43526", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43746", + "issue:43819", + "issue:43825", + "issue:43854", + "issue:44112", + "issue:44206", + "issue:44360", + "issue:44448", + "issue:44493", + "issue:44561", + "issue:44779", + "issue: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:43819|issue:44493", + "issue:44206|issue:45823", + "issue:43335|issue:43504", + "issue:43278|issue:43825", + "issue:45588|issue:45859", + "issue:29127|issue:32090", + "issue:39692|issue:43526", + "issue:45290|issue:45381", + "issue:45310|issue:45478", + "issue:43335|issue:45083", + "issue:45459|issue:46234", + "issue:39692|issue:44493", + "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:44857|issue:45072", + "issue:45920|issue:46192", + "issue:38175|issue:44493", + "issue:28218|issue:32090", + "issue:44448|issue:44977", + "issue:40444|issue:43526", + "issue:44977|issue:45229", + "issue:43701|issue:45083", + "issue:43746|issue:46209", + "issue:29942|issue:35141", + "issue:32090|issue:33357" + ], + "prepared_review_unit_hash": "0c1f514c07c6bc608ae1e0ed8c7b5621ed678a62927dd7635426790f190be709", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36464, + "estimated_input_tokens": 9116, + "estimated_eval_tokens": 18488 + }, + "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 single duplicate family: it mixes tokenizer regressions, model-loading bugs, CI failures, config issues, and offload/dtype problems. The only clear duplicate pair is the two identical Qwen3.5 MoE from_pretrained reports (issue:45310 and issue:45478).", + "confidence": 0.9, + "canonical_issue_reason": "issue:45920 is the strongest representative issue: it is open, has the broadest impact, and describes a multi-model AutoTokenizer regression rather than a narrow one-off failure.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best issue to keep as the cluster representative because it covers a widespread v5 tokenizer regression across several models and has inbound references.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "Different bugs: DAC.from_latents/forward mismatch vs unexpected position_id keys across models." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Different components and failure modes: feature extractor center arg crash vs Gemma4 device mismatch." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Unrelated model/config bug vs legacy-field pretrained loading bug." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Embedding dtype drift in train/eval is unrelated to a pipeline error-message regression." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Both are offload-related, but one is a flash_attention aux-state crash and the other is a rotary-embedding CPU-offload incompatibility." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "Different errors in different subsystems: LayoutLMv3 box validation vs Trainer GPU broadcast NoneType." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "Documentation example problems do not match a BeitImageProcessorFast reduce_labels bug." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash is unrelated to Qwen2.5-VL video position-id computation." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same underlying bug: identical Qwen3.5 MoE from_pretrained failure on transformers>=5.4.0." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug is unrelated to qwen3_omni_moe output-length helper behavior." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer protobuf decode error hiding vs a missing import_utils symbol are different issues." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 docs example errors are unrelated to a position-id key mismatch." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "Packed-weights loading bug and GraniteSpeech CI stale override failure are different problems." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Both are version regressions, but they affect different models and code paths." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "SwitchTransformers config creation bug is unrelated to GraniteSpeech CI device override behavior." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "Sparse-layer config bug is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch is not the same as an AMP/CUDA loss crash." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Local-path import warning is unrelated to the GLM-4.7-Flash test loading failure." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Model-loading breakage and embedding reinitialization after resize are different bugs." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexer architecture issue is unrelated to the removal of is_torch_fx_available." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different failures: LwDetrImageLoss AMP crash vs dtype mismatch in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "AutoTokenizer wrong token IDs is not the same as return_tensors being ignored when text_kwargs is passed." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to position-id key warnings." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing after added token is unrelated to Trainer broadcast errors." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus v4/v5 output drift is a different issue from Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Multi-image IterableDataset finetuning failure is unrelated to BeitImageProcessorFast label reduction." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Flash-attention generation bug and multi-GPU OOM are distinct failure modes." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint key mismatch is unrelated to qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "PEFT local checkpoint loading and FP8 inference on B200 are different issues." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures do not match an embedding resize reinitialization bug." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer NoneType broadcast and MacOS bus error with CLIP are different code paths." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The cluster summary is conservative and matches the packet: it is mostly heterogeneous, and the only clearly identical duplicate pair is issue:45310 / issue:45478. The choice of issue:45920 as representative is reasonable given its broad multi-model v5 tokenizer regression title and inbound references." + }, + "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": 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-10000-451", + "nodes": [ + "issue:10000", + "issue:15354", + "issue:25251", + "issue:27957", + "issue:32090", + "issue:33290", + "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:44186", + "issue:44206", + "issue:44448", + "issue:44466", + "issue:44479", + "issue:44561", + "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:45491", + "issue:45563", + "issue:45588", + "issue:45698", + "issue:45803", + "issue:45823", + "issue:46399" + ], + "soft_pairs": [ + "issue:44855|issue:46399", + "issue:43550|issue:45083", + "issue:42490|issue:43575", + "issue:10000|issue:15354", + "issue:43927|issue:45698", + "issue:44561|issue:44779", + "issue:37428|issue:44987", + "issue:43278|issue:43329", + "issue:43873|issue:45229", + "issue:44857|issue:45588", + "issue:27957|issue:34689", + "issue:41720|issue:45229", + "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:44448|issue:44466", + "issue:45372|issue:45406", + "issue:43976|issue:45440", + "issue:45198|issue:45803", + "issue:25251|issue:32090", + "issue:45245|issue:45563", + "issue:43653|issue:44743", + "issue:32090|issue:33290", + "issue:44206|issue:44479", + "issue:44871|issue:44977", + "issue:44186|issue:45198" + ], + "prepared_review_unit_hash": "19b71d1e9e6c434daf4b3f30015cd676b2391e6268153198c8aa4338a6d96f22", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36702, + "estimated_input_tokens": 9176, + "estimated_eval_tokens": 18608 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Loose overcluster of mostly unrelated bug reports; the soft pairs share superficial keywords but not the same concrete failure or fix path. No soft edge should be merged.", + "confidence": 0.97, + "canonical_issue_reason": "No single issue is a clear duplicate hub; the cluster is heterogeneous.", + "canonical_pr_reason": null, + "best_issue_reason": "If an anchor is required, issue:45698 is the broadest and most discussion-heavy loading/serialization report, but it still does not represent the rest of the cluster well.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44855", + "right": "issue:46399", + "accept": false, + "reason": "Both mention decorator/script-related failures, but one is a Python 3.13 import/parser issue and the other is a Qwen3.5 registration error; different code paths." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Torch.compile/SDPA crash in Bamba is unrelated to a feature-length helper behavior change in qwen3_omni_moe." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Local-path trailing slash warning and multi-GPU TP OOM are different problems with no shared fix path." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "German DistilBERT error and ViT torch.jit GeneratorExp issue are unrelated model/export failures." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load, but one loses custom token IDs while the other loads the wrong custom module; different serialization bugs." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Removal of is_torch_fx_available breaks trust_remote_code, while DeepSeek tokenizer regression is a separate tokenizer behavior bug." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "Missing flash-support import and loading physical-intelligence/fast may both touch model loading, but the titles indicate different failures and root causes." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift in eval and undefined variables in the video token-count branch are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization/offloading behavior and Gemma4 multi-GPU CUDA OOM are not the same concrete bug." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "float16 AMP loss crash and flash_attention s_aux=None crash are different subsystems and failures." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "Flash Attention 2 issue for XLM-R and Llama 3.2 model-loading regression are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "cudaErrorAssert on auto device mapping is a different failure mode from an OOM during Gemma4 inference." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Qwen3.5 flash-attention generation bug and a stale warning for num_return_sequences are unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer BOS/EOS omission and Gemma3 NaN embeddings under sliding-window attention are different bugs." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 processor import failure and flash_attention s_aux handling are separate code paths." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "BF16-to-FP32 dtype change and gradient-checkpointing-in-eval restriction are unrelated." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export guard/erfinv support problem is unrelated to the flash_attention None handling crash." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "Both are Gemma4-related, but one is a flash-attention sink-less-model crash and the other is a device-mismatch forward bug; not the same fix." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Pegasus v4/v5 output regression and tied-weight serialization inconsistency are distinct issues." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Gemma4 processor import error and serve-time missing _tokenizer attribute are different root causes." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility issue and DeepSeekV3MoE divergence are unrelated." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failure and Fuyu's bare except swallowing errors are separate bugs." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting behavior and Trainer _gpu_broadcast_one NoneType error do not share a concrete code path." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category-limit runtime error and stale warning for num_return_sequences are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug and Qwen3.5 recurrent-state reset bug are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer broadcast NoneType error and AdaFactor/DeepSpeed OOM are different training failures." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported center arg in LasrFeatureExtractor and Qwen VL video-input regression are different components." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma eos_token_id mismatch and Qwen3.5 flash-attention generation failure are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "LayoutLMv2 tokenizer crash on NER inputs and Wav2Vec2 save/tokenization failure are separate tokenizer/model bugs." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Same as above: runtime category limit vs stale warning are not the same underlying issue." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer decode bug and recurrent-state reset under cache 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": 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-10000-451", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:29942", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:35141", + "issue:39692", + "issue:42490", + "issue:42915", + "issue:42994", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43381", + "issue:43421", + "issue:43526", + "issue:43575", + "issue:43611", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43824", + "issue:43873", + "issue:43994", + "issue:44186", + "issue:44265", + "issue:44279", + "issue:44361", + "issue:44442", + "issue:44462", + "issue:44464", + "issue:44625", + "issue:45072", + "issue:45198", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45584", + "issue:45593", + "issue:45712", + "issue:45834", + "issue:46121", + "issue:46369", + "issue:46399" + ], + "soft_pairs": [ + "issue:43611|issue:46121", + "issue:42490|issue:43873", + "issue:44361|issue:45198", + "issue:43065|issue:45412", + "issue:29127|issue:34567", + "issue:33357|issue:39692", + "issue:45459|issue:45712", + "issue:43065|issue:43994", + "issue:42915|issue:42994", + "issue:43066|issue:44462", + "issue:43278|issue:43575", + "issue:45372|issue:45459", + "issue:29942|issue:30064", + "issue:45412|issue:46369", + "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:45440|issue:45593", + "issue:43994|issue:46369", + "issue:43526|issue:45834", + "issue:44464|issue:45584", + "issue:46121|issue:46399", + "issue:43526|issue:44186", + "issue:29127|issue:35141", + "issue:44625|issue:45584", + "issue:43575|issue:43701", + "issue:44442|issue:45072" + ], + "prepared_review_unit_hash": "a553b2db2c508fc39388a5d308db8f6c9b997dc38b43d690cdb59bafb2d4c5b1", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36897, + "estimated_input_tokens": 9225, + "estimated_eval_tokens": 18706 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose semantic grouping of mostly unrelated Transformers bug reports across tokenizers, quantization/offloading, model loading, vision processors, training state, and generation. I don\u2019t see any soft pair that looks like the same underlying bug or change, so no duplicate merges are justified.", + "confidence": 0.34, + "canonical_issue_reason": "issue:43873 is the strongest representative here: it has substantial discussion/inbound references, a concrete reproducible runtime bug, and broad user impact around quantization/offloading.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best single issue to anchor this cluster because it is well-scoped, actively discussed, and broadly relevant compared with the more niche reports in the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Different failures: model loading with base_model_prefix vs rope-parameter validation on a JSON-loaded list." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Trailing-slash local path import warning is unrelated to quantization/offloading behavior." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizer/model code paths: MLukeTokenizer AttributeError vs Wav2Vec2 save/tokenization failure." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and RT-DETR memory release are distinct bugs in different models." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 box validation messaging and TrainerState token counting are unrelated." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error with a community CLIP model is not the same as SigLIP2 docs/model mismatch." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer protobuf exception masking and leftover dummy classes are unrelated maintenance issues." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d is unrelated to SigLIP2 producing bad outputs with AutoModel/pipeline." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Quantization config failure during load and quantized model saving failure are different stages with different root causes." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 is a different bug from ignoring tokenizer.json in the repo." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "BF16/FP32 dtype drift in eval is unrelated to Qwen2-57B TP loading OOM." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma processor import failure from ReasoningEffort and hidden tokenizer errors from protobuf absence are different problems." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void-segmentation-map processing failures are unrelated." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "Both mention memory, but RT-DETR GC behavior and SAM3 model memory overflow are different bugs." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 incorrect outputs and RT-DETR memory leak are not the same code-path problem." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failures with torch_compilable_check are unrelated to Wav2Vec2 save/tokenization errors." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Eval-mode gradient checkpointing and resume_from_checkpoint key mismatch are separate training issues." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE layer-count mismatch and DeepSeekV3 implementation divergence are different model-specific bugs." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer issues." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token post-processor updates and Gemma3 NaN embeddings are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:29942", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support and failing Flash Attention 2 tests are not the same concrete bug." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error messaging and CLIP MacOS bus error are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Qwen2.5-VL import error and generic dependency issue do not share a concrete root cause." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepSeekV3 implementation divergence and D-FINE auxiliary-loss behavior are unrelated model bugs." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "SigLIP2 bad outputs and SAM3-LiteText memory overflow are distinct issues." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and Kosmos2.5 long-OCR index error are different 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 TypeError from a loaded list and missing-decorator crash on rotary-pos embedding registration are different bugs." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels and LayoutLMv2Tokenizer NER/padding crashes are unrelated tokenizer/processor bugs." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 box error handling and output embedding reinitialization in post_init are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty transcription after token alignment are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "TP loading OOM for Qwen2-57B and resume_from_checkpoint key mismatch are distinct training/runtime issues." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "FastSpeech2ConformerTokenizer loading and bf16 dtype mismatches in SwitchTransformers/TimmWrapperModel 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-10000-451", + "nodes": [ + "issue:27957", + "issue:29127", + "issue:33290", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:42175", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43262", + "issue:43316", + "issue:43334", + "issue:43381", + "issue:43388", + "issue:43452", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43575", + "issue:43611", + "issue:43646", + "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:45245", + "issue:45440", + "issue:45479", + "issue:45491", + "issue:45588", + "issue:45859", + "issue:45907", + "issue:46234" + ], + "soft_pairs": [ + "issue:43761|issue:45859", + "issue:42915|issue:43931", + "issue:42490|issue:43381", + "issue:43575|issue:44387", + "issue:43526|issue:44857", + "issue:43334|issue:43531", + "issue:44977|issue:45245", + "issue:44464|issue:45491", + "issue:43262|issue:44206", + "issue:27957|issue:29127", + "issue:43452|issue:45588", + "issue:44743|issue:44811", + "issue:27957|issue:33290", + "issue:43646|issue:44077", + "issue:33290|issue:34689", + "issue:44062|issue:45907", + "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", + "issue:43232|issue:43388" + ], + "prepared_review_unit_hash": "117e02343e060122143bce5c461caeb75962850f7654602043f0975d3f4b169b", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36424, + "estimated_input_tokens": 9106, + "estimated_eval_tokens": 18468 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous set of Transformers issues spanning unrelated bugs (loading, generation, quantization, processors, config/API regressions). I don\u2019t see any convincing duplicate pairs; the soft edges are mostly same-subsystem, not same underlying defect.", + "confidence": 0.98, + "canonical_issue_reason": "Issue #43873 is the most discussion-heavy and broadly user-impacting item here, so it\u2019s the best representative if one issue must stand in for the set.", + "canonical_pr_reason": null, + "best_issue_reason": "#43873 has the highest activity and inbound references, and its quantization/offloading regression is a prominent, high-impact user problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "Both are model-related regressions, but one is CLIP hidden-states behavior and the other is CPU-offload compatibility in Qwen3_5Moe; different code paths and fixes." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Same Qwen3 family, but one is a FineGrainedFP8Config failure and the other is a model shape-mismatch loading error for Qwen3-VL; not the same bug." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "One is a local-path import warning, the other is a gradient-checkpointing eval-mode limitation; unrelated behaviors." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both mention OOM, but the causes differ: tensor-parallel loading of a specific model vs increased reserved memory under int4 quantization." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "Different image-processing/model-loss bugs with no shared concrete failure mode." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Both involve Qwen3 variants, but one is a missing config attribute during load and the other is a sliding-window state-reset bug." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Completely different failures: flash-attention generation bug vs category-count overflow." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Different generation/attention issues in different models; no shared code-path problem." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Both are audio processor regressions, but one is a chat-template sampling-rate default and the other is an unsupported `center` argument in a feature extractor." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "Unrelated: XLM-RoBERTa flash-attention support vs a LayoutLMv3 error-message clarity issue." + }, + { + "left": "issue:43452", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer/model loading with gguf is unrelated to a flash-attention `s_aux=None` crash." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Different subsystems and symptoms: Qwen3.5 cache state reset vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "No concrete overlap beyond being user-facing bugs; flash-attention support and deepspeed Adafactor OOM are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Both touch model initialization, but one is a broad v5 custom-model breakage and the other is a specific `patchtsmixer` post_init policy issue." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Different problems and models; shared OOM/loading wording is too broad to merge." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "Tokenization `AddedToken` argument duplication is unrelated to `list_repo_files` revision forwarding." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "OOM under deepspeed Adafactor vs trainer token-count bookkeeping; no shared bug." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Both mention training/model internals, but one is optimizer OOM and the other is embedding reinitialization in `post_init`." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Both are training/runtime behavior issues, but gradient-checkpointing-in-eval and quantization offloading are distinct code paths." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure vs image-classification resolution/interpolation failure; unrelated." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe `sys.modules` access and DeepSeekV3 implementation divergence are different defects." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer-loading regression and single-label classification zero-loss bug are unrelated." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIP hidden-states output regression is unrelated to `ModelOutput` key assignment semantics." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Import-path warning and Qwen recurrent-state reset are unrelated." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both concern config API differences, but they are different model families and different missing attributes." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Packaging/backend dependency issue vs missing utility export; not the same underlying change." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both are model-loading regressions, but they affect different versions and different failure mechanisms." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading under accelerate is unrelated to tokenizer loading for FastSpeech2ConformerTokenizer." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs synchronization and metric-gathering label truncation are different pipeline 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": 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-10000-451", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:27957", + "issue:28282", + "issue:32090", + "issue:42617", + "issue:42907", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43493", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43643", + "issue:43650", + "issue:43701", + "issue:43742", + "issue:43749", + "issue:43761", + "issue:43825", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:44077", + "issue:44188", + "issue:44190", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44496", + "issue: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:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:43931|issue:45216", + "issue:43761|issue:44188", + "issue:44077|issue:44479", + "issue:44987|issue:46178", + "issue:43493|issue:43827", + "issue:44464|issue:46421", + "issue:44442|issue:45803", + "issue:43873|issue:44871", + "issue:42907|issue:45698", + "issue:43749|issue:43931", + "issue:43643|issue:44913", + "issue:44479|issue:45290", + "issue:43316|issue:45759", + "issue:44855|issue:45137", + "issue:44190|issue:44496", + "issue:43650|issue:46202", + "issue:45632|issue:46424", + "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:43065|issue:44617", + "issue:27957|issue:32090" + ], + "prepared_review_unit_hash": "2cc4cc5c15341c228229963ce08a4c0f1edff785f61851a42f560018888c91c1", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35305, + "estimated_input_tokens": 8827, + "estimated_eval_tokens": 17910 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the paired items span unrelated model-loading, configuration, pipeline, cache, compilation, and docs issues. I do not see any true duplicate groups here; all soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": "No true canonical duplicate target stands out. If forced to pick a representative, issue 45698 has the most discussion/inbound references, but it is not a duplicate hub for the rest of this cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45698 is the strongest representative only by activity and recency, not because it matches the others. It is the most visible issue in an otherwise unrelated set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Different failures: Qwen3-VL weight-shape mismatch on load vs a Qwen3.5 save_pretrained regression." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Different code paths: CLIPVisionModel hidden_states bug vs attention-kernel divergence under torch.compile." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "Unrelated: patchtsmixer post_init validation vs a v5.3.0 video-input regression in Qwen/VL models." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Different problems: loading a specific model on transformers>=5.1.0 vs an unsupported tensorflow_text backend mapping error." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "One is a SigLIP2 implementation mismatch; the other is stale docs referencing pipeline() after removals." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Both touch generation/compile, but one is incorrect outputs and the other is spurious recompilation from cache initialization." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure vs tensor-conversion exception handling; not the same bug." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Different features: quantization/offloading cache behavior vs Gemma eos_token_id config mismatch." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Saving a dequantized model is different from loading the wrong custom module after save_pretrained." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU-RAM-efficient loading failure vs Qwen3-VL weight-shape mismatch." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Missing fields with trust_remote_code on AutoConfig is not the same as rotary_pct resetting on reload." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video-input regression across models vs chat-template crashes on assistant tool-call messages." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Both involve config APIs, but Gemma3TextConfig discrepancy and Qwen multimodal text_config unwrapping are distinct bugs." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 import parsing error vs a DeepSpeed ZeRO3 deque IndexError; unrelated failures." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "Unrelated; one is a vague data request, the other is a Gemma pipeline roundtrip failure." + }, + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Cache-path collision for trust_remote_code is unrelated to max_cache_length behavior in quantized cache." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k output-shape change is unrelated to a missing PyTorch dependency import error." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Processor/tokenizer regression is different from an incorrect translation-support error message after pipeline removals." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile + SDPA failure in Bamba is unrelated to a resume_from_checkpoint key mismatch." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel.py runtime failure and Qwen2 TP OOM are different loading/execution problems." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to checkpoint key mismatch." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "TP OOM while loading Qwen2-57B is not the same as a key error loading MobileLLM-125M." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "Model_max_length question for DeBERTa-V3 is unrelated to a transformers.onnx import error." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is a different bug from Sam3Video CUDA OOM." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "XLMRoberta flash-attention support and a Trainer _gpu_broadcast_one NoneType error 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": 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-10000-451", + "nodes": [ + "issue:41950", + "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:43646", + "issue:43673", + "issue:43688", + "issue:43716", + "issue:43761", + "issue:44077", + "issue:44162", + "issue:44242", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44743", + "issue:44792", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:44977", + "issue:45127", + "issue:45200", + "issue:45406", + "issue:45542", + "issue:45584", + "issue:45698", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46234", + "issue:46365", + "issue:46369", + "issue:46421" + ], + "soft_pairs": [ + "issue:43316|issue:43716", + "issue:43540|issue:44514", + "issue:43295|issue:44871", + "issue:44871|issue:45584", + "issue:43316|issue:44877", + "issue:44610|issue:46369", + "issue:44162|issue:44792", + "issue:44077|issue:44534", + "issue:43404|issue:43441", + "issue:45803|issue:46365", + "issue:45803|issue:46170", + "issue:44871|issue:46050", + "issue:41950|issue:43475", + "issue:43452|issue:44496", + "issue:43646|issue:44913", + "issue:43761|issue:44479", + "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", + "issue:43525|issue:43531", + "issue:44792|issue:44977", + "issue:43761|issue:44242" + ], + "prepared_review_unit_hash": "edd3bfea7e291fa14a077a231fa7d0d0f4e0a1b684b4c754056bccc052cacaa8", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36170, + "estimated_input_tokens": 9043, + "estimated_eval_tokens": 18342 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most candidates are related by broad Transformers/multimodal regression themes, but they describe distinct model- or API-specific bugs. The only strong duplicate-like pair is the chunked-prefill cache regression (43673/46421).", + "confidence": 0.86, + "canonical_issue_reason": "issue:46421 is the clearest root-cause-focused description of the shared chunked-prefill cache lazy-initialization regression; issue:43673 reads like an earlier symptom report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46421 is the best representative because it names the underlying cache-lazy-init problem and the triggering path more precisely than the sibling report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Related config/model mismatch theme, but different models and different failure modes." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both are Qwen video/processor bugs, but the concrete failures differ." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Processor-tokenizer regression vs Gemma eos-token-id config mismatch; not the same bug." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated model/config issue vs Whisper empty-transcription failure." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Both involve config handling, but they affect different model families and code paths." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "Different subsystems: image-size mismatch vs SAM3 litetext memory overflow." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "Different models and symptoms; one is ESM2 breakage, the other a Janus image-generation test failure." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Unrelated bugs: optional post_init enforcement vs non-persistent buffer junk." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Different Mistral/Ministral issues and different failure mechanisms." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Both affect imports, but one is Fuyu tensor conversion and the other is Trainer/datasets import interference." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Bare except swallowing errors is unrelated to the Windows/ROCm torch.distributed import problem." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Gemma token-id config mismatch vs quantized-model save/load corruption." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Both are video/SAM3-adjacent, but one is processor selection and the other is a missing encoder attribute." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Different load-time failures: gguf path handling vs missing model_type in config.json." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Broad v5 initialization breakage vs a specific GPTNeoX config reload bug." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "Different regression surfaces: CLIP hidden_states output vs Qwen video-input generation." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Different training bugs: metric gathering truncation vs auxiliary-loss normalization." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "Both involve Gemma-family fine-tuning, but the underlying bugs are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache/recurrent-state behavior, but different models and distinct code paths." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "StableLm pad-token config regression is unrelated to the tensorflow_text backend registration error." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "CUDA OOM in SAM3 video is not the same as Whisper empty-output generation failure." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Pretrained preset loading with a legacy field is different from loading the wrong custom module after save_pretrained." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": true, + "reason": "Both describe the same chunked-prefill cache initialization regression; one reports missing cache, the other recompilation from lazy init." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Serve KV-cache indexing bug and Gemma4Processor missing _tokenizer are different failures." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "TensorBoard/TensorFlow backend setup issue vs missing import_utils availability helper." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Llama4 pad_token_id absence and Qwen3-MoE sliding_window behavior are unrelated." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure vs Qwen3.5 flash-attention generation issue." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIPVision hidden_states regression is unrelated to load-balancing loss logic." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the soft-edge verdicts stay conservative overall. The 43673/46421 pair is the only clearly duplicate-like match and is supported by the issue titles, while the other rejects are appropriately cautious given 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": 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-10000-451", + "nodes": [ + "issue:25251", + "issue:27957", + "issue:35141", + "issue:41720", + "issue:42490", + "issue:43065", + "issue:43116", + "issue:43335", + "issue:43540", + "issue:43550", + "issue:43644", + "issue:43827", + "issue:43931", + "issue:43976", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44112", + "issue:44190", + "issue:44462", + "issue:44464", + "issue:44479", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44734", + "issue:44737", + "issue:44792", + "issue:44855", + "issue:44877", + "issue:44913", + "issue:44933", + "issue:44964", + "issue:45127", + "issue:45245", + "issue:45357", + "issue:45362", + "issue:45381", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45600", + "issue:45632", + "issue:45803", + "issue:45907", + "issue:46234", + "issue:46369" + ], + "soft_pairs": [ + "issue:45381|issue:46369", + "issue:45563|issue:45584", + "issue:44792|issue:45600", + "issue:43931|issue:44060", + "issue:44534|issue:44913", + "issue:44190|issue:44964", + "issue:35141|issue:42490", + "issue:44855|issue:45588", + "issue:44479|issue:44734", + "issue:43827|issue:43994", + "issue:44617|issue:46369", + "issue:41720|issue:45245", + "issue:43065|issue:44610", + "issue:44464|issue:45563", + "issue:44521|issue:45381", + "issue:44933|issue:45362", + "issue:43644|issue:44913", + "issue:43976|issue:44855", + "issue:44077|issue:44877", + "issue:43550|issue:44190", + "issue:45127|issue:45357", + "issue:44462|issue:45907", + "issue:43540|issue:44734", + "issue:25251|issue:27957", + "issue:45632|issue:46234", + "issue:44112|issue:45600", + "issue:43335|issue:44190", + "issue:44737|issue:45803", + "issue:43116|issue:44190" + ], + "prepared_review_unit_hash": "2ab07b41f6262870ff8e13c79b76bc2dbc0b745cd1313cf5c4a45d3ca28cf724", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36351, + "estimated_input_tokens": 9088, + "estimated_eval_tokens": 18432 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All candidate soft edges are false positives: the items span unrelated bugs, configs, model families, and CI/doc issues rather than one duplicate cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the set spans unrelated model bugs, config problems, docs, and CI failures, and none of the soft pairs look like the same underlying defect.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue; there is no coherent duplicate cluster here, so no issue is a strong canonical representative.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Different models and failures: Qwen2.5-VL video position ids vs Sam3LiteText memory overflow." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "Both involve generation, but one is a stale warning in paged generate and the other is Whisper failing on empty transcription." + }, + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Janus image-generation test failure and Sam3LiteText auto_mappings CI break are unrelated code paths." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Different Qwen variants with different symptoms: weight-shape mismatch vs an incorrect tied-weights warning." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk fill is unrelated to GPTNeoX rotary_pct reverting on reload." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading script failure and Phi-4 multimodal model loading error are distinct problems." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Embedding reinitialization after resize is unrelated to the import warning from a local path ending with a slash." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python 3.13 JIT parsing/IndentationError is unrelated to a flash_attention AttributeError on s_aux=None." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Qwen VL video regression and serve KV-cache continuation indexing are different bugs in different flows." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing pipeline after removals is not the same as nonsensical SigLIP2 outputs with pipeline/AutoModel." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Both mention memory, but Sam3Video OOM and Sam3LiteText memory overflow are different model-specific issues." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3 device-mapping cudaErrorAssert and a categories-limit RuntimeError are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and OmDet-Turbo processor size mismatch are different model/component bugs." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Compiled-forward chunked generation inconsistency is unrelated to a stale warning about num_return_sequences." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "All-zero assistant masks in multimodal chat templates is a different bug from wrong Qwen2.5-VL video position ids." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A nonexistent import from image_utils is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk fill and GPTNeoX rotary_pct reload behavior are separate config/runtime issues." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "These are different Python-version failures with different root causes and models." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "PatchTSTMixer post_init policy and strict granite_speech config loading are unrelated config issues." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to loading a local dataset in an example script." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge with extended vocab causing collapse is not the same as Qwen3.5 save_pretrained writing wrong visual encoder keys." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and list_repo_files not forwarding revision are unrelated hub/tokenizer behaviors." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and serve KV-cache tensor indexing are different failures." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "Pipeline top_k output nesting and XLMRoberta Flash Attention 2 are unrelated bugs." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Local trust_remote_code cache collisions and a missing import_utils helper are unrelated." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "A stale GraniteSpeech test and Sam3LiteText auto_mappings references are different CI breakages." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformersConfig sparse-layer creation and local dataset loading are unrelated." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet device placement in torch.arange and a bare except swallowing errors are unrelated code defects." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Multi-label classification returning empty results and local dataset loading are different example-script 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": 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-10000-451", + "nodes": [ + "issue:27957", + "issue:35141", + "issue:36246", + "issue:38617", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42907", + "issue:43295", + "issue:43404", + "issue:43441", + "issue:43479", + "issue:43526", + "issue:43531", + "issue:43643", + "issue:43650", + "issue:43673", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44112", + "issue:44186", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44496", + "issue:44683", + "issue:44734", + "issue:44938", + "issue:44964", + "issue:44987", + "issue:45003", + "issue:45216", + "issue:45245", + "issue:45276", + "issue:45375", + "issue:45446", + "issue:45542", + "issue:45584", + "issue:45600", + "issue:45632", + "issue:46082", + "issue:46121", + "issue:46123", + "issue:46369", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:45276|issue:45584", + "issue: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:43673|issue:46424", + "issue:44186|issue:44734", + "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", + "issue:43531|issue:45600", + "issue:42907|issue:43441", + "issue:43526|issue:43901" + ], + "prepared_review_unit_hash": "687c135be617a9cb4509f5c1829160c40ff3eb313049d13bf2ee2ed6811a5456", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36064, + "estimated_input_tokens": 9016, + "estimated_eval_tokens": 18288 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated Transformers issues with only superficial overlap around model loading, config handling, cache, or attention subsystems. I don\u2019t see a true duplicate cluster here, so there is no good canonical issue/PR anchor.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Different bugs: Gemma4 resize_token_embeddings vs Whisper empty-transcription generation after align_special_tokens." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Different code paths: processor.tokenizer/image-input regression vs Gemma4 embedding resizing/output embedding behavior." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Unrelated failures: Python 3.14 import/load regression vs missing TensorFlow backend when installing torch extras." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Different config validation problems: Llama head_dim rules vs JSON-loaded rope ignore-keys type handling." + }, + { + "left": "issue:45632", + "right": "issue:46421", + "accept": false, + "reason": "Both mention cache, but one is a trust_remote_code cache-path collision and the other is lazy cache init causing recompilation." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Different issues: category-count limit runtime error vs Whisper empty transcription generation failure." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different symbols and modules: Qwen2_5_VLImageProcessor vs AutoImageProcessor." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "One is an extras/backend packaging complaint; the other is a vague dependency issue without the same concrete failure." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Different bugs: untied lm_head weights in Mistral3 vs incorrect checkpoint saving for Qwen3.5." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "No relation: placeholder 'ADD THE DATA' issue vs _torch_extract_fbank_features implementation bug." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Different concerns: import slowdowns vs unsafe sys.modules access in modeling_utils." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Both are import errors, but they refer to different removed/missing names and separate fixes." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Pipeline bug vs notebook syntax error; no shared underlying code path." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache-related, but one is missing GenerationMixin cache in chunked prefill and the other is max_cache_length handling for quantized cache." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer padding/truncation crash vs serve KV-cache continuation tensor indexing bug; unrelated paths." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Different model-config omissions: Phi4 multimodal defaults vs Qwen3_5MoeVisionConfig missing field." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Both touch flex_attention, but one is a torch>=2.9 compiled failure and the other is a version-check/import guard bug." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Different load failures: missing fields with trust_remote_code vs unrecognized model/config for Olmo-Hybrid." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Different model-loading regressions: Phi-4 multimodal loading vs stale auto_mappings reference to removed Sam3LiteText configs." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Different models and failures; both loading-related but not the same underlying bug." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error and audio feature extraction bug are unrelated." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI test flake in GraniteSpeech vs Phi-4 multimodal model-loading issue; no common code path." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Different problems: final partial batch dropping results in MaskGenerationPipeline vs memory overflow in sam3-litetext." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Unrelated: XLMRoberta FlashAttention2 support vs output embedding reinitialization after resizing token embeddings." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Environment/version compatibility issue vs a specific model loading failure; not the same bug." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Different models and symptoms: Qwen3-MoE sliding_window vs Sam3LiteText auto_mappings cleanup." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Same model family only at a broad level, but one is save_pretrained on dequantized Ministrals/Devstrals and the other is FlashAttention failure in Ministral-3." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "Completely different areas: BeitImageProcessorFast label reduction vs TextClassificationPipeline docs mismatch." + } + ] + }, + "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-10000-451", + "nodes": [ + "issue:28282", + "issue:35141", + "issue:38175", + "issue:41950", + "issue:43295", + "issue:43299", + "issue:43381", + "issue:43441", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44242", + "issue:44279", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44496", + "issue:44617", + "issue:44779", + "issue:44792", + "issue:44933", + "issue:45020", + "issue:45042", + "issue:45103", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45335", + "issue:45357", + "issue:45372", + "issue:45405", + "issue:45584", + "issue:45588", + "issue:45593", + "issue:45759", + "issue:45812", + "issue:46082", + "issue:46209", + "issue:46234", + "issue:46399", + "issue:46424" + ], + "soft_pairs": [ + "issue:43295|issue:45335", + "issue:44792|issue:46424", + "issue:28282|issue:35141", + "issue:44496|issue:45042", + "issue:46082|issue:46399", + "issue:45372|issue:46234", + "issue:43441|issue:45588", + "issue:44016|issue:44062", + "issue:43299|issue:45216", + "issue:43901|issue:44496", + "issue:43673|issue:44336", + "issue:43824|issue:44933", + "issue:43381|issue:45593", + "issue:43976|issue:44279", + "issue:44496|issue:45759", + "issue:44075|issue:44368", + "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:44279|issue:45405", + "issue:43299|issue:45357", + "issue:41950|issue:45020", + "issue:43295|issue:44779", + "issue:44779|issue:45812", + "issue:44351|issue:45278" + ], + "prepared_review_unit_hash": "bb3d9240456b978427133d49c400dfcf933abfae65aae03791035c2613023802", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35586, + "estimated_input_tokens": 8897, + "estimated_eval_tokens": 18050 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose collection of unrelated Transformers v5 regression reports, centered on model loading, tokenizers, processors, and import/export breakages. I do not see any true duplicate pairs among the soft edges.", + "confidence": 0.41, + "canonical_issue_reason": "issue:45020 is the broadest umbrella-style report in the set: it explicitly describes recent-version breakage affecting models that use `remote_code`, which best matches the overall theme of the cluster better than the more specialized reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the most representative issue because it frames the widest user-visible regression class and could serve as the closest cluster anchor, even though the items are not duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Different failures: processor/tokenizer image handling regression vs token embedding resize not updating decoder weights." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "Unrelated bugs: janus image-generation test failure vs cache length handling for quantized cache." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "ImportError due to missing PyTorch vs embedding reinitialization after resize; different code paths." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model config vs PIL image processor requiring torchvision; unrelated issues." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "LlamaConfig head_dim validation vs Qwen3.5 kernel registration error; not the same bug." + }, + { + "left": "issue:45372", + "right": "issue:46234", + "accept": false, + "reason": "Both are missing-import reports, but for different symbols and subsystems; not a duplicate bug." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but one is a model-specific failure and the other is a sink-less-model AttributeError in a helper module." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error vs AddedToken keyword argument collision; unrelated." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Qwen3VL MoE loading breakage vs incorrect Qwen3.5 save_pretrained output; different failure modes." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Documentation mismatch for `return_all_scores` vs model config import error; not the same underlying problem." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Generation cache regression vs ANSI color output in loading_report; unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Different import errors involving different missing symbols and modules." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing eval-mode behavior vs D-FINE auxiliary-loss logic; unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python version compatibility report vs vague dependency issue; not the same concrete bug." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Generic unrecognized model error vs composite Qwen config unwrapping bug; different code paths." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "SGD optimizer args ignored vs tie_word_embeddings warning during LoRA fine-tuning; unrelated." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Token embedding resize bug vs Whisper generation failure after align_special_tokens; different models and paths." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "Placeholder/data request vs FP8 inference failure on B200; unrelated." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "BLIP2 dtype propagation bug vs Qwen3.5 checkpoint save regression; not a duplicate." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "Auto-docstring crash vs missing mistral_common symbol; different tooling and failure modes." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "MoE load-balancing loss bug vs Gemma 4 mm_token_type_ids defaulting issue; unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Zero-probability model output vs CUDA OOM; not the same bug." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "General dependency issue vs minimum PEFT version bump; unrelated." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen3.5, but one is loading failure and the other is incorrect visual-encoder keys on save; different bugs." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification pipeline processor lookup bug vs broad remote_code regression report; not the same concrete issue." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Processor/tokenizer image handling regression vs Deepseek tokenizer output regression; different affected paths." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer correctness regressions in different model families; too broad to treat as the same bug." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "Single missing import for HybridCache vs broad 'many import errors' report; related theme but not a single identical bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "nodes": [ + "issue:33290", + "issue:39692", + "issue:41950", + "issue:42175", + "issue:42994", + "issue:43299", + "issue:43317", + "issue:43441", + "issue:43450", + "issue:43479", + "issue:43525", + "issue:43650", + "issue:43653", + "issue:43723", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44291", + "issue:44297", + "issue:44361", + "issue:44462", + "issue:44466", + "issue:44521", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:45081", + "issue:45276", + "issue:45325", + "issue:45397", + "issue:45405", + "issue:45542", + "issue:45593", + "issue:45600", + "issue:45656", + "issue:45712", + "issue:45823", + "issue:45834", + "issue:45907", + "issue:46132" + ], + "soft_pairs": [ + "issue:44521|issue:45325", + "issue:39692|issue:43901", + "issue:43479|issue:43525", + "issue:41950|issue:45823", + "issue:44683|issue:45542", + "issue:43976|issue:45405", + "issue:44462|issue:45081", + "issue:43299|issue:43854", + "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", + "issue:42175|issue:43976", + "issue:43723|issue:44162", + "issue:43441|issue:44683", + "issue:43901|issue:43994" + ], + "prepared_review_unit_hash": "b6eea5077641f64c4eab9484265128502509fdf786476ee64ee23e7ec3ade8d3", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36297, + "estimated_input_tokens": 9075, + "estimated_eval_tokens": 18406 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is highly heterogeneous overall; most soft pairs are only thematically similar at a broad subsystem level, not true duplicates. The strongest hub among the issues is the tokenizer-loading regression around AutoTokenizer/tokenizer.json, but even that does not unify the cluster cleanly.", + "confidence": 0.22, + "canonical_issue_reason": "issue:44462 is the clearest concrete regression in the mostly tokenizer/loading-heavy subset, with a specific failure mode and enough breadth to be a plausible anchor if a single issue must represent the group.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44462 is the best global representative because it is specific, user-facing, and closer than the docs/config-only or unrelated model/task bugs; it also sits near other tokenizer-related reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Both are multimodal-related, but one is about assistant masks in chat templating and the other is about temporal position IDs in Qwen2.5-VL; different code paths and failures." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "One is a SigLIP2 docs example with model/processor mismatch and quantization failure; the other is a TextClassificationPipeline docs mismatch. Both are documentation issues, but not the same bug." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different model configs and different failures: Phi4Multimodal default sub-config initialization vs Llama4Config missing pad_token_id." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Both touch video/multimodal areas, but the bugs are unrelated: video-classification choosing image processors vs Gemma4 device mismatch in forward." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Compiled flex_attention on torch>=2.9 is unrelated to the TensorFlow backend install issue." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "One is a Python version compatibility bug; the other is a dependency version bump to unreleased peft. No shared code path." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "Both concern tokenizer loading, but one is AutoTokenizer ignoring tokenizer.json and the other is Mistral regex patching crashing on backend_tokenizer access; different defects." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Both are model-loading regressions, but for different models and causes: Qwen3VL MoE loading vs GLM-4.7-Flash unit-test loading." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Both are tokenizer-related, but save_pretrained writing the wrong tokenizer_class is not the same bug as AutoTokenizer ignoring tokenizer.json." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor batch-shape bug is unrelated to a generic dependency issue with transformers." + }, + { + "left": "issue:43650", + "right": "issue:46132", + "accept": false, + "reason": "A vague data request is unrelated to AttentionInterface.register changing function behavior." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regex crash and list_repo_files not forwarding revision are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Both involve serialization/model-shape mechanics, but one is inconsistent lm_head.weight saving and the other is resize_token_embeddings not propagating in Gemma4." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error and init_empty_weights TypeError are unrelated failures." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "Ministral-3 FlashAttention failure and Kosmos2.5 OCR long-input index error are different model bugs." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized model saving failure is not the same bug as device_map=auto failing to load a dequantized model with GPU+CPU offload." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "BigBirdTokenizer special-token registration bug is unrelated to D-FINE auxiliary losses being skipped." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "Processor save_pretrained missing files and Gemma4 zero3 from_pretrained are different code paths and symptoms." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor save_pretrained missing files is unrelated to the performance regression after upgrading Transformers." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Broken auto_mappings references and leftover dummy_pt_objects classes are separate repository-maintenance issues." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "Both involve DeepSpeed and optimizers, but one is an OOM with Adafactor and the other is optimizer.step being called twice; not the same bug." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "Different GLM model internals: missing rms_norm_eps in config vs torch.split return values in GlmMoeDsaIndexer." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "MLukeTokenizer AttributeError on tasks is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "Both are output regressions in v5, but CamemBERT masked LM predictions and Deepseek tokenizer results are different models and failures." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error and Gemma4 device mismatch are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "TensorFlow backend installation and Python version compatibility are different dependency issues." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer loading in v5 and ESM2 being broken are both broad regressions, but they do not point to the same underlying defect." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "FlashAttention failure in Ministral-3 and compiled flex_attention failure on torch>=2.9 are distinct attention-path bugs." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs mismatch for return_all_scores and incorrect SigLIP2 outputs with AutoModel/pipeline 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": 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-10000-451", + "nodes": [ + "issue:36246", + "issue:41628", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:42886", + "issue:43262", + "issue:43299", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43643", + "issue:43673", + "issue:43761", + "issue:43784", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:44188", + "issue:44220", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44297", + "issue:44336", + "issue:44351", + "issue:44373", + "issue:44485", + "issue:44496", + "issue:44521", + "issue:44589", + "issue:44734", + "issue:44779", + "issue:44871", + "issue:45003", + "issue:45103", + "issue:45292", + "issue:45362", + "issue:45397", + "issue:45542", + "issue:45561", + "issue:45563", + "issue:45593", + "issue:46132", + "issue:46399", + "issue:46424" + ], + "soft_pairs": [ + "issue:43673|issue:43761", + "issue:43262|issue:44521", + "issue:44297|issue:45397", + "issue:43824|issue:43976", + "issue:43525|issue:43881", + "issue:43901|issue:44246", + "issue:43525|issue:43643", + "issue:44589|issue:45362", + "issue:43784|issue:45542", + "issue:41628|issue:43976", + "issue:44188|issue:44336", + "issue:45103|issue:46399", + "issue: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:44246|issue:44871", + "issue:42886|issue:46132", + "issue:44734|issue:46424", + "issue:36246|issue:42757", + "issue:44336|issue:44485", + "issue:36246|issue:44351", + "issue:45003|issue:45561", + "issue:44779|issue:45292", + "issue:44336|issue:46424" + ], + "prepared_review_unit_hash": "16670fed138d61d2210802fd79a806f14883d4749ca3aff70367b24f394449f2", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36041, + "estimated_input_tokens": 9011, + "estimated_eval_tokens": 18278 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is heterogeneous: it mixes import/export errors, model-loading regressions, cache/generation bugs, and unrelated docs/tests issues. None of the soft pairs look like the same underlying bug or a PR that could plausibly merge into one change set, so I would not accept any duplicate edges.", + "confidence": 0.95, + "canonical_issue_reason": "issue:43824 is the clearest representative of the cluster\u2019s concrete import/export regression pattern: a user-facing top-level import failure with multiple related references.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 is the most actionable and recognizable failure mode in the set, but only as a rough representative; the cluster is too mixed to treat as one duplicate bucket.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Different regressions: chunked-prefill cache behavior vs CLIPVisionModel hidden_states output." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Both involve chat templates, but one is an audio sampling-rate default and the other is multimodal assistant masks." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save metadata mismatch is unrelated to gemma-4 zero3 from_pretrained loading." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Importing Qwen2.5-VL symbols is a different failure than Python version compatibility for 5.1.0." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Different model families and different loading failures: Llama4Config pad_token_id vs glm-4v-9b load failure." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs for return_all_scores are unrelated to intermittent import performance." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Both are config-related, but missing pad_token_id and trust_remote_code stripping fields are distinct bugs." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Float8 storage lookup error is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Different backend/import problems: sentence-transformers nn import vs tensorboard/tf backend handling." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "AutoImageProcessor import failure is not the same as Python 3.9/3.10 support mismatch." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention kernel divergence on torch.compile is unrelated to ANSI codes in loading_report." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "Future-annotations docstring crash and Qwen3.5 kernel registration error are different code paths." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "xdist race on captured_info.txt is unrelated to a stale num_return_sequences warning." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Terminal ANSI output bug is unrelated to a wrong docstring for position_ids." + }, + { + "left": "issue:44242", + "right": "issue:44496", + "accept": false, + "reason": "Missing load-balancing loss and unrecognized-model config errors do not share a code path." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "position_ids buffer access bug is unrelated to gemma-4 zero3 loading." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Different training-loss issues: router-logits gating vs D-FINE auxiliary losses." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both mention Qwen MoE, but one is LoRA compatibility across branches and the other is a Qwen3VL-MoE loading regression." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Sliding_window behavior in Qwen3-MoE is not the same as Qwen3.5 chat crashes." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and unsafe sys.modules access are unrelated." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline processor selection is unrelated to _torch_extract_fbank_features()." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowness and Gemma-3 eos_token_id inconsistency are different subsystems." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline cache loading bug is unrelated to AttentionInterface.register changing function behavior." + }, + { + "left": "issue:44734", + "right": "issue:46424", + "accept": false, + "reason": "KV cache continuation crash in serve is not the same as max_cache_length being honored only for Quantized Cache." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they involve different symbols in different packages and different causes." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI output formatting and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Missing Qwen2_5_VLImageProcessor and missing HybridCache are separate export/import regressions." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access and xdist file races are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "DeepSeek tokenizer regression and resize_token_embeddings/output_embeddings mismatch are different bugs." + }, + { + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "ANSI code emission and cache length handling 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": 38, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36246", + "issue:43257", + "issue:43262", + "issue:43352", + "issue:43475", + "issue:43540", + "issue:43596", + "issue:43673", + "issue:43901", + "issue:44162", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44466", + "issue:44479", + "issue:44496", + "issue:44561", + "issue:44749", + "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:45584", + "issue:45600", + "issue:45656", + "issue:45685", + "issue:45701", + "issue:45759", + "issue:45859", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46170", + "issue:46291", + "issue:46369", + "issue:46399", + "issue:46431" + ], + "soft_pairs": [ + "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:43901|issue:45003", + "issue:43352|issue:45250", + "issue:43673|issue:44561", + "issue:44162|issue:45200", + "issue:44496|issue:45003", + "issue:43257|issue:45137", + "issue:43262|issue:44246", + "issue:45292|issue:45584", + "issue:45375|issue:45759", + "issue:45412|issue:46123", + "issue:45292|issue:45701", + "issue:44561|issue:46399", + "issue:44877|issue:45375", + "issue:45685|issue:46431", + "issue:44987|issue:45459", + "issue:44995|issue:45200", + "issue:44466|issue:45335", + "issue:44479|issue:46170", + "issue:45397|issue:45656", + "issue:45137|issue:45600", + "issue:45987|issue:46082", + "issue:43475|issue:43540" + ], + "prepared_review_unit_hash": "d408a70f253fb6ad906b14c38a9a0caa99aa7f57e53035b1f9b0e0b4c3be4d16", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35608, + "estimated_input_tokens": 8902, + "estimated_eval_tokens": 18060 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous cluster of mostly unrelated Transformers issues. The soft pairs share broad keywords like Qwen, deepspeed, config, or generation, but they describe different models, symptoms, and code paths, so none of the candidate merges look like true duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "issue:45003 is the closest thing to a cluster representative because it is a core library/runtime bug in modeling_utils and sits near several import/config-related regressions, even though the overall cluster is not a single coherent bug.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45003 is the strongest global anchor: it touches core library behavior, is broadly impactful, and is more representative than the model-specific or docs-only reports in this mixed set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "Different SAM 3 video failures: missing attribute on vision encoder output vs memory overflow in sam3-litetext." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "MoE load-balancing loss omission and RT-DETR AMP crash are unrelated bugs in different code paths." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Position_ids buffer access error is not the same as the reported slowdown after upgrading Transformers." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "CPU offload incompatibility in rotary embeddings and a missing decorator on apply_rotary_pos_emb are distinct Qwen3.5 issues." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Kernel version/revision requirements for evaluate are unrelated to the Qwen3.5 rotary registration crash." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Qwen2.5 VL image processor import failure and BertModel Zero3 index error are different failures with no shared code path." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "One is a documentation mismatch; the other is a runtime bug in sys.modules access." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "Nemotron flash-attn unsupported error is model-specific; the other is just a generic flash attention topic." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked-prefill cache handling and removal of is_torch_fx_available are separate incompatibilities." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are unrelated model-specific bugs." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading and unsafe sys.modules access are different loader failures." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Qwen3 MoE weights not converted under accelerate+deepspeed is different from the ZeRO3 deque/index error." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio chat-template sampling-rate default and import slowdowns are unrelated issues." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "resize_token_embeddings output mismatch and Whisper empty-transcription generation failure do not share the same root cause." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Strict config dropping a field and from_config not unwrapping text_config are both config-handling bugs, but they affect different fields and paths." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "RT-DETR memory leak and MaskGenerationPipeline dropping final partial batches are unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "Embedding resize behavior and version-dependent tokenization changes are different problems." + }, + { + "left": "issue:44561", + "right": "issue:46399", + "accept": false, + "reason": "Removed helper breaking trust_remote_code is unrelated to the Qwen3.5 apply_rotary_pos_emb registration crash." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both involve config strictness, but they concern different models and different missing fields." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Kernel revision requirements and tokenizer error masking when protobuf is absent are distinct issues." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "GlmMoeDsa stale cache on second forward and Gemma 4 zero defaults are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Tied-weight serialization inconsistency and decoder.embed_tokens not updating after resize are related only loosely by embeddings, not the same bug." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Qwen video input regression and Windows/ROCm torch.distributed import failure are unrelated." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Gemma-4 Zero3 from_pretrained and optimizer.step being called twice under deepspeed are different deepspeed problems." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 index error and stale auto_mappings references are not the same underlying bug." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache seq-length type mismatch and LlamaConfig head_dim validation are unrelated." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MPS histogram failure in MoE and missing chat template error are different features and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:41628", + "issue:42886", + "issue:42994", + "issue:43065", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43482", + "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:44933", + "issue:44977", + "issue:44995", + "issue:45229", + "issue:45230", + "issue:45356", + "issue:45381", + "issue:45397", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45656", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202", + "issue:46421" + ], + "soft_pairs": [ + "issue:44483|issue:46202", + "issue:43653|issue:44162", + "issue:44060|issue:44077", + "issue:42886|issue:45229", + "issue:44368|issue:44977", + "issue:44995|issue:45588", + "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: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", + "issue:28282|issue:30990", + "issue:43475|issue:45381", + "issue:41628|issue:44933" + ], + "prepared_review_unit_hash": "d19114d88199de63fbcc0ee69bbd2a227997372927e62fa30b19a4d081771000", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35881, + "estimated_input_tokens": 8971, + "estimated_eval_tokens": 18198 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: the items span unrelated import errors, tokenizer/processor save issues, model-specific regressions, generation/cache bugs, and SAM/Qwen/Whisper-specific failures. I did not find any soft-edge pair that clearly describes the same underlying bug or change, so all soft links are rejected.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 45229 is the most prominent single report in the set: it has the highest discussion activity, an inbound reference, and a concrete high-impact runtime failure. If one issue must represent the cluster, this is the best anchor despite the cluster\u2019s overall heterogeneity.", + "canonical_pr_reason": "No PRs are present in this set.", + "best_issue_reason": "45229 is the strongest standalone issue candidate because it is concrete, heavily discussed, and clearly actionable. None of the other issues are close enough duplicates to displace it.", + "best_pr_reason": "No PRs are present in this set.", + "soft_edge_verdicts": [ + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "Both involve generation/output behavior, but one is an API acceptance regression and the other is a Gemma pipeline roundtrip/tokenization issue; they are not the same bug." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Both are model/tokenizer-related, but BigBird tokenizer decode behavior and ESM2 breakage are distinct failures with no shared code-path evidence." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Different models and failures: one is a tied-weights warning, the other is a post_init policy issue for patchtsmixer." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline tokenizer cache loading and Gemma multi-GPU CUDA OOM are unrelated problems." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "Both mention Qwen3.5, but one is a weight-tying warning while the other is a flash-attention generation bug; not the same underlying defect." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "A stale indexer cache on second forward and a flash-attention None-attribute crash are different code paths and symptoms." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "One is a config warning about tied embeddings; the other is a stale warning in paged generate for num_return_sequences." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Both are SAM3-related, but one concerns a dummy Conv2d in pixel decoding and the other a model-type mismatch warning during instantiation." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "torch.compile attention-kernel divergence and Gemma pipeline roundtrip failure are unrelated bugs." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Strict config loading for granite_speech and a Windows ROCm torch.distributed import/dependency problem are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "A Qwen3-VL weight-shape mismatch and a Qwen3.5 tie_word_embeddings warning are different loading/config issues." + }, + { + "left": "issue:45584", + "right": "issue:46170", + "accept": false, + "reason": "Whisper empty-transcription generation failure and a Windows ROCm import error do not share the same failure mode." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Both touch cache/compiled generation, but one is about diverging attention kernels under torch.compile and the other about lazy cache initialization causing recompilation." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "SAM3 pixel-decoder configuration and MaskGenerationPipeline partial-batch result dropping are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Both are save_pretrained-related, but tokenizer_class mismatch and missing processor files are different bugs and likely different save paths." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer AttributeError and a regression causing slower filtering after upgrade are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenization/decoding regressions, but Whisper timestamp decoding and Kimi-K2.5 codec handling are distinct issues." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Missing processor save files and a deepspeed optimizer double-step bug are entirely different subsystems." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "A generic Transformers 5.0 custom-model initialization breakage is not the same as the ESM2-specific breakage report." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Performance regression in filtering after upgrade and double optimizer stepping with deepspeed are unrelated." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "resume_from_checkpoint key mismatch and slower filtering after upgrade do not describe the same defect." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "Both concern generation internals, but one is an _update_model_kwargs_for_generation/sync_gpus bug and the other is missing cache in chunked_prefill." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior changes are unrelated." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "The first item is too generic to match the specific gemma-4 zero3 from_pretrained problem." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are different quantization/loading paths." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe breakage and missing processor save files are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "A missing field in Sam3VisionEncoderOutput is not the same as a trust_remote_code AutoConfig missing-fields issue." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch-missing ImportError for AutoModel and Sentence Transformers loading hang are different problems." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal vision issues, but they affect different models and different attributes/position-id logic." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but one is a missing top-level AutoImageProcessor import and the other is a non-existent image_utils import; insufficient evidence they stem from the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-10000-451", + "nodes": [ + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42371", + "issue:42673", + "issue:42831", + "issue:42898", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43377", + "issue:43504", + "issue:43525", + "issue:43526", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43873", + "issue:43976", + "issue:44038", + "issue:44164", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44351", + "issue:44496", + "issue:44568", + "issue:44792", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45081", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45310", + "issue:45375", + "issue:45406", + "issue:45478", + "issue:45561", + "issue:45712", + "issue:45753", + "issue:46077", + "issue:46123", + "issue:46202", + "issue:46326", + "issue:46399" + ], + "soft_pairs": [ + "issue:42371|issue:45092", + "issue:45406|issue:46202", + "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:42673|issue:43704", + "issue:46077|issue:46123", + "issue:41628|issue:44351", + "issue:44964|issue:45137", + "issue:43232|issue:44188", + "issue:44222|issue:45310", + "issue:44945|issue:45753", + "issue:42994|issue:44164", + "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", + "issue:42898|issue:44568", + "issue:43976|issue:44933", + "issue:43278|issue:45137" + ], + "prepared_review_unit_hash": "e58a46f4e8f4b179d359ac3a179e12a52a1f41acd727052a9a788db60a848092", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36023, + "estimated_input_tokens": 9006, + "estimated_eval_tokens": 18268 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose bundle of unrelated Transformers bug reports, with one clear duplicate pair around the Qwen3.5 MoE from_pretrained failure. Most other soft pairs share only broad subsystem terms (imports, saving/loading, generation, Qwen models) but not the same concrete bug.", + "confidence": 0.95, + "canonical_issue_reason": "issue:45310 is the better canonical issue: it cleanly represents the Qwen3.5 MoE from_pretrained regression, and issue:45478 is an obvious duplicate with the same title and bug scope.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the strongest global representative because it has an exact duplicate in the set and the problem description is specific and central to its subcluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "Different problems: TF32 backend configuration vs. an old InternVL2 meta-initialization incompatibility." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Both involve multimodal/image-ish behavior, but one is a processor `_tokenizer` attribute error and the other is a Gemma roundtrip/tokenization issue." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Related Qwen3.5 family, but one is a missing config field silently dropped by strict, the other is a from_pretrained loading error." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Different code paths: quantized model save failure vs. missing `pad_token_id` on `Llama4Config`." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer loading regression vs. `continue_final_message` behavior for mistral_common tokenizers; not the same bug." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Import-time `nn` NameError in sentence-transformers integration is unrelated to Qwen3.5 kernel registration." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Same model family, but different failures: from_pretrained error vs. config field dropping." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Both are AttributeErrors, but one is Mistral tokenizer regex patching and the other is an auto-docstring kwargs processing crash." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization/offloading bug is unrelated to pipeline-parallelism output corruption." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": false, + "reason": "Same Qwen3 model label, but the reports describe different VRAM leak contexts and are not clearly the same code-path." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Both touch generation attention/pipeline logic, but one is bidirectional attention masking and the other is partial-batch result dropping." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different symbols (`AutoImageProcessor` vs `HybridCache`) and different failure modes." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Model loading failure for Phi-4 multimodal is unrelated to the DeepSpeed ZeRO3 deque error." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Different generation/attention issues: kwargs update after `sync_gpus` vs. diverging attention kernels under `torch.compile`." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Both mention MoE/Qwen, but one is FP8 save_pretrained and the other is a from_pretrained loading error; not the same bug." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Pipeline parallelism incorrect output is unrelated to the Qwen3.5 specific-image infinite loop." + }, + { + "left": "issue:42994", + "right": "issue:44164", + "accept": false, + "reason": "Both involve save/load, but one is quantized model saving and the other is extra_state handling; different concrete failures." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Same broad Qwen MoE area, but different model variants and generations of transformers; insufficient evidence of the same underlying bug." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Different missing exports: `GenerationMixin` vs `AutoImageProcessor`." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation kwargs handling is unrelated to the janus image-generation test failure." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "FP8 accuracy regression is not the same as FP8 save_pretrained failure, even if both mention FP8." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "Backend dependency packaging issue is unrelated to leftover dummy classes leaking into `dir(transformers)`." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Model config loading failure is unrelated to the pytest-xdist captured_info.txt race." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug is unrelated to the xdist file race." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Pretrained preset loading with a legacy field is unrelated to the testing-utils race condition." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Package backend selection is unrelated to missing image_utils imports." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI padding-mask batching bug is unrelated to load-balancing loss not being added." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer cleanup behavior change is unrelated to mdeberta BOS/EOS token insertion regression." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python-version compatibility issue is unrelated to the image_utils import problem." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift during evaluation is unrelated to the ZeRO3 deque underflow." + } + ] + }, + "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": 41, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28282", + "issue:30333", + "issue:30990", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:42222", + "issue:42831", + "issue:42898", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44079", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44448", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44869", + "issue:44933", + "issue:45071", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45200", + "issue:45310", + "issue:45356", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:45812", + "issue:46082", + "issue:46132", + "issue:46209", + "issue:46231", + "issue:46431" + ], + "soft_pairs": [ + "issue:30990|issue:35141", + "issue:44623|issue:45310", + "issue:30333|issue:33290", + "issue:42222|issue:45200", + "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:44016|issue:46231", + "issue:44734|issue:45103", + "issue:42898|issue:45356", + "issue:43874|issue:45092", + "issue:44295|issue:45656", + "issue:44075|issue:46231", + "issue:44060|issue:45125", + "issue:42898|issue:44448", + "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:46209|issue:46431", + "issue:30333|issue:36331", + "issue:30333|issue:35141", + "issue:22355|issue:28282", + "issue:45071|issue:46082" + ], + "prepared_review_unit_hash": "a7dffda18a604c305c18b8d02b6655c61d626629a4be5703c00975060d8b6217", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37640, + "estimated_input_tokens": 9410, + "estimated_eval_tokens": 19076 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated bug reports spanning different subsystems, models, and release regressions. I found no soft-edge pair that clearly describes the same underlying issue, so I would keep them separate.", + "confidence": 0.61, + "canonical_issue_reason": "issue:45812 is a broad, concrete regression report with clear user impact: AutoTokenizer returning wrong token IDs across an entire model family. It is specific enough to anchor a cluster while still representing a common v5 tokenizer regression theme.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45812 is the strongest standalone issue here because it is severe, clearly described, and affects multiple Granite models with a silent correctness regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Different problems: loading/stall behavior in Sentence Transformers vs embedding reinitialization after resize. No shared code path." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Different artifacts and symptoms: processor.save_pretrained file omission vs Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow reporting on training failure is unrelated to a deepspeed OOM during Adafactor optimization." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Different model families and failure modes: broken VitPose models vs Gemma 4 multimodal token type defaults." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Phi4MultimodalConfig default-init bug is unrelated to Whisper word timestamp decode crashing on trailing replacement chars." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config default initialization and ModelOutput key assignment are distinct code paths and bugs." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are v5 tokenizer regressions, but they affect different behaviors and model families; no evidence of one underlying bug." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Importing sentence-transformers with missing nn is unrelated to a missing Qwen2.5 VL symbol export." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Generic tokenizer loading issue vs a specific add_special_tokens BOS/EOS regression; not the same bug." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Different tokenizer regressions for different models; similar theme but not the same underlying defect." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error is unrelated to a documentation parameter-name mismatch." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "KV-cache continuation crash in serve is unrelated to auto_docstring failing under future annotations." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Different multimodal model compatibility issues: missing image patch API vs remote-code checkpoint/meta-init incompatibility." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Reading a registered buffer and double optimizer stepping in deepspeed are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "Optimizer argument handling is unrelated to a documentation typo." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Both involve Qwen models, but one is a tied-weights warning bug and the other is missing tensor-parallel plan; not clearly the same change." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Tokenizer cleanup behavior change and Pegasus output drift are different regressions with different likely causes." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "A dependency version bump is unrelated to leftover dummy class cleanup." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 save_pretrained behavior and tokenizer save_pretrained metadata mismatch affect different save paths." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing image_utils import and leftover dummy classes are both export/namespace issues, but not the same concrete bug." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer config metadata and deepspeed optimizer stepping are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Pegasus v4/v5 output mismatch is unrelated to AttentionInterface.register behavior changes." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both touch save_pretrained, but one is FP8 model serialization and the other is processor file saving; different code paths." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "XLMRoberta Flash Attention 2 support is unrelated to MLflow reporting after training failures." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "Failed-job reporting in MLflow is unrelated to a Llama 3.2 vision model loading regression." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Defaulting multimodal configs to None and junk-filling non-persistent buffers are different initialization bugs." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FP8 accuracy concerns are unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 inference failure on B200 and missing chat template on a processor are different issues." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow reporting on failure is unrelated to a CustomTrainer signature change." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "Training-job reporting and output embedding reinitialization are unrelated bugs." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx module and missing PyTorch dependency errors are different import failures." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "PretrainedConfig type-checking regression and LlamaConfig head_dim validation are distinct config 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": 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-10000-451", + "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:45092", + "issue:45103", + "issue:45245", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45834", + "issue:46002", + "issue:46050", + "issue:46082", + "issue:46132", + "issue:46209", + "issue:46305", + "issue:46328", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:45245|issue:46132", + "issue:44297|issue:45310", + "issue:43122|issue:46209", + "issue:43475|issue:46305", + "issue:41628|issue:42757", + "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:44263|issue:44360", + "issue:42907|issue:46050", + "issue:30990|issue:38175", + "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": "2c18871f235082c312ffb436c7e5e5a34a49af8a15f0b6d5383327af19336602", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36130, + "estimated_input_tokens": 9033, + "estimated_eval_tokens": 18322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a grab bag of unrelated Transformers issues. Several are broadly about v5 loading/saving regressions, but the soft pairs are not the same concrete bug, so none should be merged as duplicates.", + "confidence": 0.79, + "canonical_issue_reason": "issue:43723 is the best cluster anchor: it is a clear, user-facing v5 tokenizer-loading regression in a generic API path, making it more representative than the model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43723 is the most broadly applicable and easiest-to-track issue among this set; it captures a common v5 loading failure without being tied to one model family.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Different bugs: category-count runtime error vs attention registration behavior change." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Both involve model/tokenizer loading, but one is tokenizer config serialization and the other is a Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Different failure modes: tokenizer output changes vs FP8 inference failure on B200." + }, + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "SAM3 video encoder attribute error is unrelated to MoonshineStreaming window masking off-by-one." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and different dependency paths." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Dataset loading in an example script is unrelated to LlamaConfig head_dim validation." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Both are tokenizer/model loading complaints, but one is a generic AutoTokenizer issue and the other is Qwen3.5 generation with flash-attention." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "One is a LasrFeatureExtractor argument regression; the other is an auto_docstring crash with future annotations." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Different sequence bugs: Qwen2.5-VL vision position ids vs Moonshine streaming mask off-by-one." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Generation kwargs update after sync_gpus is unrelated to cache max length honoring." + }, + { + "left": "issue:44734", + "right": "issue:46421", + "accept": false, + "reason": "KV-cache continuation tensor indexing crash is a different issue from lazy cache init causing recompilation." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 MoE save_pretrained bug is unrelated to the 50GB shard-size default question." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs handling and Kosmos2.5 OCR index errors are different code paths." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Position_ids buffer access bug is unrelated to processor.save_pretrained missing files." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "torch.split return handling in a DSA indexer is not the same as remote-code meta-init incompatibility." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring mismatch for position_ids is unrelated to GLM-5 RoPE implementation correctness." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer loading regression is different from eos_token_id config inconsistency in Gemma-3." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request rejection is unrelated to Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "Generic dependency trouble is not the same as a performance bug in SequenceFeatureExtractor.pad." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": false, + "reason": "Same broad area, but not the same bug: one is generic tokenizer loading, the other is a specific EMBEDDIA/est-roberta failure." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Both mention DSA indexer, but one is split return values and the other is a missing ReLU." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Saving dequantized Ministrals/Devstrals is unrelated to sinq quantization breaking after save/load." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang and zero probabilities in siglip2 are different model/runtime issues." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen/InternVL save/load regressions, but the concrete failures and code paths differ." + }, + { + "left": "issue:43761", + "right": "issue:46082", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression is unrelated to LlamaConfig head_dim validation." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer junk in v5 is unrelated to a streaming attention window off-by-one." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "FlashAttention-3 training degeneration is different from NaNs during checkout/init." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "MLA RMS norm epsilon precision issue is unrelated to remote-code checkpoint meta-init incompatibility." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FP8 accuracy issues on one side and Qwen3.5 MoE from_pretrained errors on the other 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": 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-10000-451", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28218", + "issue:30990", + "issue:33357", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43479", + "issue:43550", + "issue:43576", + "issue:43646", + "issue:43756", + "issue:43784", + "issue:43824", + "issue:44038", + "issue:44060", + "issue:44117", + "issue:44188", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44368", + "issue:44373", + "issue:44488", + "issue:44521", + "issue:44568", + "issue:44704", + "issue:44913", + "issue:44945", + "issue:45278", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46132", + "issue:46305", + "issue:46326", + "issue:46369", + "issue:46402" + ], + "soft_pairs": [ + "issue:43479|issue:44913", + "issue:41628|issue:43824", + "issue:42831|issue:44038", + "issue:43408|issue:46402", + "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:44368|issue:46305", + "issue:22355|issue:28218", + "issue:42886|issue:42994", + "issue:42831|issue:45753", + "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:42371|issue:44263", + "issue:30990|issue:33357", + "issue:44188|issue:44373", + "issue:42907|issue:44945", + "issue:43122|issue:44521", + "issue:43122|issue:46326", + "issue:41628|issue:45278", + "issue:39401|issue:44488" + ], + "prepared_review_unit_hash": "946934485050c01537b0267472f10ff20faadd5e51ed33ee8825c836e596f00b", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36187, + "estimated_input_tokens": 9047, + "estimated_eval_tokens": 18350 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is not a true duplicate cluster: it mixes unrelated tokenizer, import, quantization, generation, attention, and model-specific bugs. I rejected every soft edge; no single issue cleanly represents the rest.", + "confidence": 0.98, + "canonical_issue_reason": "No reliable canonical issue: the set is heterogeneous, so picking one issue as the duplicate anchor would be arbitrary.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue: none of these reports safely covers the rest of the cluster, so there is no strong representative anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Different models and failures: Phi4MultimodalConfig default-init bug vs GPTNeoX rotary_pct reload regression." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols and likely different breakpoints." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "Accuracy issue in FineGrainedFP8 is unrelated to the Qwen3-VL-Moe/transformers 5.0 bug." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "Same SAM3 family, but one is a warning mismatch and the other is a text-encoder weight-loading failure." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "A torch.split/indexing bug and a docstring bug are not the same underlying issue." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer regressions, but offset_mapping and add_special_tokens/BOS-EOS behavior are different code paths." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "Module import failure for transformers.onnx vs XLMRoberta Flash Attention 2 support; unrelated problems." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Specific huggingface_hub symbol import breakage vs a broad 'many import errors' report; not the same bug." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "Flash Attention 2 support and Sentence Transformers loading hangs are different issues." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure is different from wrong tokenizer decoder type in Transformers v5." + }, + { + "left": "issue:44368", + "right": "issue:46305", + "accept": false, + "reason": "A tied-embeddings warning and a sliding-window mask off-by-one are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "Importing transformers.onnx and tokenizer added-token spacing are different bugs." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Offline tokenizer cache loading and quantized model saving are separate code paths." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FP8 accuracy regression and a Qwen3_5 image infinite loop are not the same underlying problem." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "SAM3 warning/instantiation mismatch vs memory overflow in sam3-litetext; different failure modes." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs handling and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken transformers env command vs import-time NameError in sentence-transformers are different issues." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization breakage and NaNs in classification heads are distinct bugs." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch and Qwen3-Next tied-weights warning do not share the same code path." + }, + { + "left": "issue:43550", + "right": "issue:46132", + "accept": false, + "reason": "torch.compile/SDPA failure and AttentionInterface.register behavior change are different attention-related regressions." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Reading a registered position_ids buffer and tokenizer_class mismatch on save are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES None handling and AutoProcessor kwargs propagation are different loader bugs." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 settings guidance and torch.split return values are unrelated." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading hang and MacOS bus error are different runtime failures." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "torch.compile attention-kernel divergence and a wrong docstring are not the same issue." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Quantized model saving failure and pipeline-parallelism output corruption are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Tokenization differs across versions, but that is not the same as all-zero assistant_masks for multimodal chat templates." + }, + { + "left": "issue:43122", + "right": "issue:46326", + "accept": false, + "reason": "Both involve tokenizer behavior, but different_final_message handling and version-to-version tokenization drift are different bugs." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing export is not the same as a broad import-error regression report." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Wrong offset_mapping and failure to load a specific model are unrelated tokenizer/model-loading problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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-10000-451", + "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:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44222", + "issue:44261", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44938", + "issue:44945", + "issue:44964", + "issue:45072", + "issue:45200", + "issue:45216", + "issue:45290", + "issue:45478", + "issue:45600", + "issue:45753" + ], + "soft_pairs": [ + "issue:44261|issue:44373", + "issue:42907|issue:43937", + "issue:43976|issue:44945", + "issue:43784|issue:45600", + "issue:44704|issue:45290", + "issue:42907|issue:44964", + "issue:39401|issue:44521", + "issue:44038|issue:44297", + "issue:43116|issue:45200", + "issue:44625|issue:44829", + "issue:44360|issue:44485", + "issue:44038|issue:44295", + "issue:43122|issue:44514", + "issue:43408|issue:43475", + "issue:43122|issue:44779", + "issue:42907|issue:43452", + "issue:44038|issue:44222", + "issue:45216|issue:45478", + "issue:36296|issue:43116", + "issue:43824|issue:44351", + "issue:22355|issue:29159", + "issue:42757|issue:44351", + "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": "88c21cfb32505dcd25a46bf357bf45caeb8e22e0b28ece628b0c8de81a5539b5", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35462, + "estimated_input_tokens": 8866, + "estimated_eval_tokens": 17988 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous bug reports across unrelated models, tokenizers, import paths, and training/runtime paths. None of the soft-similarity pairs look like the same underlying issue, so I would not merge any of them as duplicates.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue fits the whole set because the items span unrelated subsystems and symptoms.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44038 is the strongest standalone bug report candidate: it is model-specific, regression-shaped, and concrete enough to anchor follow-up triage.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "One is a config/precision bug in MLA q_a_layernorm; the other is a docstring problem for position_ids. Different symptoms and fixes." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Dequantized save_pretrained failure and invalid GenerationConfig are unrelated code paths." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python-version support breakage is not the same as incorrect output under pipeline parallelism." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "Both are import/compatibility related, but they reference different missing pieces in different integration paths." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "cached_file kwargs propagation and chat-template tool-call crashes are distinct API bugs." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Both involve loading/saving, but they are different models and different failure modes." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Wrong offset_mapping and all-zero assistant masks are separate tokenizer/template behaviors." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "A Qwen3-VL-Moe bug and tokenizer_class metadata mismatch are not the same underlying issue." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Classification example output failure and Gemma 4 mm_token_type_ids defaulting are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "num_labels propagation and flash_attention_3 degenerate training affect different paths." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "A missing ReLU in one component is not the same as a GLM-5 RoPE implementation discussion." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe bug vs position_ids buffer access error are different code paths." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer-version output differences and batched chat-template crashes are different bugs." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Model-type warning and missing fpn_position_embeddings are distinct SAM 3 Video problems." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenization-related, but they describe different models and different regressions." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "save_pretrained for dequantized models and gguf_file loading are unrelated failure modes." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Same broad model family, but one is a general transformers/Qwen3-VL-Moe bug and the other is FP8 save_pretrained behavior." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Checkpoint save correctness and from_pretrained load errors are separate operations." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor parallel training bug and example script multi-label output bug are not the same issue." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Missing exports for Qwen2_5_VLForConditionalGeneration and HybridCache are different missing symbols." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "Missing transformers.onnx import and slow/fast tokenizer inconsistency are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different packages and different missing names." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "dtype drift between training and evaluate is not the same as incorrect LLM output in pipeline parallelism." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Packaging backend omission and Python 3.14 load failure are distinct environment/setup issues." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 HF-vs-JAX discrepancy and Qwen3_5 infinite loop are different model behaviors." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 and Deepseek tokenizer output regression are different tokenizer bugs." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM and dtype mismatches in CI inference are unrelated." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "Import failure for transformers.onnx and LayoutLMv3 error-message clarity are not the same bug." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder loading and cached_file kwarg forwarding 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": 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-10000-451", + "nodes": [ + "issue:33453", + "issue:38617", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:43064", + "issue:43116", + "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:44410", + "issue:44466", + "issue:44485", + "issue:44779", + "issue:44936", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45083", + "issue:45278", + "issue:45310", + "issue:45357", + "issue:45405", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:46002", + "issue:46077", + "issue:46231", + "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:42371|issue:43874", + "issue:42222|issue:43441", + "issue:44075|issue:44936", + "issue:44038|issue:45478", + "issue:45357|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:44466|issue:46002", + "issue:44291|issue:45071", + "issue:43937|issue:44977", + "issue:45405|issue:45446", + "issue:44117|issue:45081", + "issue:44410|issue:45310" + ], + "prepared_review_unit_hash": "c9a51276c1137ef49e0445648a55bae0c8fb3aeebc576a45286b565fd66e2bb4", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36585, + "estimated_input_tokens": 9147, + "estimated_eval_tokens": 18550 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a heterogeneous mix of Transformers v5 regressions and model-specific bugs. I do not see any soft pair that is the same underlying issue; the only clear duplicate-like item is the repeated Qwen3.5 MoE from_pretrained report (45310/45478), but it is not part of the provided soft-edge list.", + "confidence": 0.77, + "canonical_issue_reason": "Issue 45310 is the best canonical anchor only for the narrow Qwen3.5 MoE from_pretrained duplicate pair, since 45478 repeats the same title later. It is not a good canonical for the full cluster because the rest of the items are unrelated.", + "canonical_pr_reason": null, + "best_issue_reason": "45310 is the strongest issue anchor overall because it is an exact duplicate-style report with a clear, concrete failure mode. Even so, it should only represent that small subtopic, not the whole heterogeneous cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Both are import-related v5 regressions, but one is a specific missing symbol and the other is a broad upgrade-breakage report." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "Both touch GLM-5, but one is a generation config validation error and the other is a RoPE implementation discussion." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Base-model loading breakage and NaNs in classification heads are different code paths and symptoms." + }, + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Attention-kernel divergence and feature-extractor padding performance are unrelated bugs." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled-file runtime errors are not the same as a documentation parameter-name mismatch." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Multi-label classification output logic and SigLIP2 model discrepancy are different problems." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance and a missing image-processor method are unrelated issues." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "VitPose import breakage and Ministral FlashAttention failure are different model-specific regressions." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "SGD argument handling and trainer.evaluate() failing after train() are distinct trainer bugs." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5/Qwen3-VL-Moe, but one is a model-loading error and the other is a different from_pretrained regression." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is save_pretrained key serialization and the other is a load-time error." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python-version compatibility and a flex_attention import version check are not the same failure." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Classification example output and pipeline deprecation messaging are unrelated." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "A qwen3_omni_moe helper behavior bug is unrelated to a flex_attention dependency/version check." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Optimizer-state corruption under FSDP2/PEFT is not the same as missing GenerationMixin cache support." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both are tokenizer-related, but one is a generic loading regression and the other is an offset_mapping correctness bug." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig validation and a stale warning on num_return_sequences are different issues." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer correctness bugs, but they affect different models and different failure modes." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES handling and flash_attention s_aux=None crashes are unrelated subsystems." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Attention scaling being ignored and bidirectional masks being built are distinct attention bugs." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "A VRAM leak in dataloader threads and an isolated reproduction of another issue are not enough to call the same bug." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing pipeline() and a flex_attention import version check are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video shape errors and missing torchvision crashes are different video-processing problems." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Chunked-prefill cache handling and padding conversion overhead are unrelated." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Serialization of tied weights and NaNs in classification heads are different bugs." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Both are loading/type-checking regressions, but the error causes and code paths differ." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GenerationConfig validation and Qwen3.5 flash-attention generation failure are not the same issue." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "A PEFT minimum-version bump and a PyTorch version check for AuxRequest import are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping None handling and Mistral regex patch crashes are different tokenizer bugs." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Missing QKV/gate projections in qwen3next and a Qwen3.5 Moe from_pretrained error are different model regressions." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "These are both Qwen3-family reports, but they concern different models and different failure surfaces." + } + ] + }, + "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-10000-451", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42490", + "issue:42617", + "issue:42673", + "issue:42831", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43937", + "issue:44075", + "issue:44220", + "issue:44261", + "issue:44410", + "issue:44483", + "issue:44484", + "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:46298", + "issue:46328" + ], + "soft_pairs": [ + "issue:43122|issue:46298", + "issue:45341|issue:45561", + "issue:38617|issue:41628", + "issue:45341|issue:46328", + "issue:44829|issue:45216", + "issue:42617|issue:45341", + "issue:42222|issue:46178", + "issue:45362|issue:46328", + "issue:42673|issue:43582", + "issue:43408|issue:43742", + "issue:39401|issue:43421", + "issue:44877|issue:46231", + "issue:42490|issue:42673", + "issue:42371|issue:44261", + "issue:44484|issue:45310", + "issue:38617|issue:44220", + "issue:43937|issue:44496", + "issue:44933|issue:45278", + "issue:43066|issue:43122", + "issue:44410|issue:45478", + "issue:42222|issue:45542", + "issue:42371|issue:42831", + "issue:43012|issue:43577", + "issue:44075|issue:45656", + "issue:44483|issue:45464", + "issue:42175|issue:42222", + "issue:43097|issue:44493", + "issue:43673|issue:45341", + "issue:41628|issue:45071" + ], + "prepared_review_unit_hash": "7a27e7f980d90c2a7ee29a70a81dae3329f77dbfdfbf8ab84308d6feddfbab75", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "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": "The soft pairs are mostly broad topical similarities, but they do not describe the same underlying bug or change. This looks like a heterogeneous issue set rather than one duplicate cluster.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Unrelated regressions: tokenizer output change vs a noisy logging message." + }, + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both mention testing utilities, but one is a vague small bug and the other is a specific xdist file race; not the same bug." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Different import failures for different symbols and code paths." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "Testing-utils bug vs SequenceFeatureExtractor.pad performance issue; no shared underlying defect." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "Both relate to Qwen3.5/training behavior, but one is flash-attention training degeneration and the other is save_pretrained checkpoint corruption." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "Completely different areas: a script execution problem vs a testing-utils bug." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "VitPose model breakage and tensorflow_text backend registration are unrelated." + }, + { + "left": "issue:45362", + "right": "issue:46328", + "accept": false, + "reason": "Chat runtime failure vs audio feature padding performance; different subsystems." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM vs Apple Silicon TypeError in caching_allocator_warmup; not the same bug." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "Different model-loading problems for different models and causes." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Both are tokenizer-related, but offset_mapping is a separate bug from runtime post-processor updates." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Config-loading strictness issue vs documentation parameter mismatch." + }, + { + "left": "issue:42490", + "right": "issue:42673", + "accept": false, + "reason": "Local-path import warning and a Qwen3 VRAM leak are unrelated." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API guidance and MLA rms_norm_eps precision error are different concerns." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "save_pretrained shard-size question vs Qwen3.5 MoE from_pretrained regression; no shared code-path defect." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for a config symbol vs a feature-extraction bug." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GenerationConfig validation for GLM-5 vs model_type missing for a different checkpoint." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import and a broad upgrade regression are not the same bug." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Both are tokenizer regressions, but decoder-type mismatch and changed tokenization output are distinct issues." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen model families and different loading failures." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "Model breakage vs backend-installation problem; unrelated." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 settings and FineGrainedFP8 accuracy are separate performance/precision topics." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "A compile-time warning and a model dtype-loading bug are different code paths." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "Both involve optimizers, but one is unused SGD args and the other is a double-step deepspeed behavior." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Different chat-completions failures with different symptoms and likely different causes." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow backend packaging issue vs VitPose model breakage." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed config attribute and unexpected position-id keys are not the same regression." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache regression vs testing-utils bug; unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "Import error for AutoImageProcessor vs PreTrainedConfig type-checking regression." + } + ] + }, + "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": 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-10000-451", + "nodes": [ + "issue:27957", + "issue:38617", + "issue:42757", + "issue:42890", + "issue:43064", + "issue:43097", + "issue:43335", + "issue:43352", + "issue:43441", + "issue:43452", + "issue:43502", + "issue:43531", + "issue:43673", + "issue:43873", + "issue:43874", + "issue:43931", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44410", + "issue:44484", + "issue:44485", + "issue:44829", + "issue:44933", + "issue:44987", + "issue:45003", + "issue:45071", + "issue:45341", + "issue:45372", + "issue:45375", + "issue:45440", + "issue:45478", + "issue:45542", + "issue:45588", + "issue:45753", + "issue:46050", + "issue:46123", + "issue:46172", + "issue:46231", + "issue:46298", + "issue:46369" + ], + "soft_pairs": [ + "issue:44263|issue:44485", + "issue:45071|issue:45542", + "issue:27957|issue:43352", + "issue:43874|issue:44261", + "issue:44263|issue:46172", + "issue:44484|issue:45478", + "issue:44360|issue:46172", + "issue:43931|issue:44410", + "issue:43441|issue:44829", + "issue:43931|issue:44038", + "issue:43335|issue:45071", + "issue:43097|issue:43452", + "issue:45003|issue:45341", + "issue:43986|issue:45341", + "issue:43673|issue:46123", + "issue:44933|issue:46231", + "issue:45753|issue:46369", + "issue:44117|issue:44987", + "issue:43986|issue:44484", + "issue:43874|issue:44263", + "issue:43531|issue:45341", + "issue:45440|issue:46172", + "issue:43986|issue:45372", + "issue:44261|issue:44485", + "issue:43097|issue:45588", + "issue:43097|issue:46050", + "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": "1ae41e1e0f286754d97b9403bd0aaa7841e4e87e1c4991ec82c328a38b0eea75", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 37956, + "estimated_input_tokens": 9489, + "estimated_eval_tokens": 19234 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous; none of the soft pairs looks like the same underlying bug/change. If a single anchor is needed, issue 43097 is the broadest core regression report among the set.", + "confidence": 0.77, + "canonical_issue_reason": "issue:43097 is the most central, high-impact report here: it targets a core Trainer v5 regression, has discussion and inbound references, and is broader than the model-specific or one-off bugs around it.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the best representative issue because it describes a concrete, cross-cutting failure in core training behavior rather than a narrow model-specific or documentation/test issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "Both mention GLM, but one is about DSA indexer torch.split behavior and the other about RoPE implementation; different code paths and bugs." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "A PretrainedConfig type-checking regression is unrelated to a tensorboard-without-TF backend error." + }, + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Both involve Flash Attention 2 support, but they are model-specific requests/errors for different architectures and likely different fixes." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Missing image-processor method and an MLA layernorm epsilon precision issue are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "These are different model families and different failures; no shared concrete bug." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "A save_pretrained sharding question is not the same as a Qwen3.5 MoE from_pretrained loading error." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "A missing ReLU in the DSA indexer is unrelated to the deepseek-v4 transpose issue." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Both are Qwen3-related, but one is a weight-shape mismatch in Qwen3-VL-30B and the other is missing projections in qwen3next; not the same bug." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both mention FlashAttention, but the model, failure mode, and code path differ." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "These are different Qwen3-VL/MoE loading problems, not a single concrete defect." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction is unrelated to PretrainedConfig type-checking." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Tie-embeddings removal and gguf_file tokenizer/model loading are different regressions." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils is unrelated to a small testing_utils bug." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "A torchvision dependency crash in AutoProcessor is unrelated to a testing utility issue." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Generation cache handling and mask pipeline partial-batch handling are different subsystems and bugs." + }, + { + "left": "issue:44933", + "right": "issue:46231", + "accept": false, + "reason": "A missing import from image_utils is unrelated to a documentation parameter-name typo." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "An infinite loop on a specific Qwen3_5 image and a SAM memory overflow are different model-specific failures." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping returning None and a specific repo loading failure are not the same underlying issue." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "A processor/torchvision crash is unrelated to a save_pretrained shard-size discussion." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "These are unrelated GLM46V image-processing and DSA-indexer bugs." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "A sliding_window bug in Qwen3-MoE is unrelated to a small testing utility bug." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "DeepseekV3MoE divergence from remote implementation is not the same as the deepseek-v4 comb transpose issue." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Torchvision-missing processor loading and a mistral_common ReasoningEffort import error are different failures." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "A layernorm epsilon precision discussion is unrelated to the max_shard_size question." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Removal of tie_embeddings_and_encoder_decoder and a flash_attention.py sinkless-model crash are different code paths." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "A config/API removal issue is unrelated to a quantized save/load regression." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "Wrong optimizer states on FSDP2/PEFT ranks is unrelated to Liger Kernel application with model_init." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols and modules; no shared root cause is evident." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "A duplicate special kwarg in AddedToken is unrelated to TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Missing qwen3next projections and a missing deepstack_visual_indexes config field are different model/config bugs." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Both are Qwen3-family issues, but they affect different model variants and different loading paths." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "local_files_only leaking API calls is unrelated to a missing embed_tokens.weight in UMT5Encoder." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "A flaky integration test lacking set_seed and a noisy logging message are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports from different modules; not the same bug." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Tie-embeddings removal and quantized offloading behavior are different regression classes." + } + ] + }, + "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-10000-451", + "nodes": [ + "issue:36683", + "issue:38617", + "issue:42175", + "issue:42490", + "issue:42757", + "issue:42907", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43208", + "issue:43352", + "issue:43502", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43824", + "issue:43825", + "issue:43906", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44077", + "issue:44117", + "issue:44162", + "issue:44222", + "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:45584", + "issue:45712", + "issue:45753", + "issue:46132", + "issue:46172", + "issue:46178", + "issue:46298", + "issue:46365", + "issue:46421" + ], + "soft_pairs": [ + "issue:44936|issue:46365", + "issue:44704|issue:45103", + "issue: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:45584|issue:46365", + "issue:44077|issue:46365", + "issue:44380|issue:45479", + "issue:44380|issue:46132", + "issue:43064|issue:44380", + "issue:44222|issue:45478", + "issue:44483|issue:44514", + "issue:36683|issue:43582", + "issue:43208|issue:43526", + "issue:42175|issue:46365", + "issue:42490|issue:43012", + "issue:43012|issue:44683", + "issue:42757|issue:43824", + "issue:38617|issue:43502" + ], + "prepared_review_unit_hash": "7807488d83a7d61986eab1433db7d6db7d5cf2c1a914f9dd58e866cea4e05d77", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37445, + "estimated_input_tokens": 9362, + "estimated_eval_tokens": 18980 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is heterogeneous: most soft pairs are superficially similar regression/import/runtime reports but not the same underlying bug. No PRs are present. I would not merge any soft edge pair.", + "confidence": 0.16, + "canonical_issue_reason": "issue:43824 is the clearest, most self-contained bug report in the set, with a specific import failure and strong impact signal; it is the best representative of the overall import/regression theme, even though the cluster is not truly duplicate-heavy.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43824 is the most actionable and representative issue here: concrete failure mode, clear reproduction context, and broader relevance than the other candidates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "Different failures: trainer lifecycle/eval breakage vs an import dependency conflict affecting Trainer import." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Different code paths: cached_file kwargs propagation vs auto-docstring annotation handling." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "Attention backend config bug vs cache lazy initialization/recompilation behavior; not the same defect." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Video-model AutoProcessor crash without torchvision is unrelated to TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Both are import-related, but one is a missing class export and the other is an optional-dependency crash." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python-version loading failure and invalid backend mapping are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "The second item is only a reproduction issue reference, not evidence of the same tokenizer bug." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training blockers and Bamba torch.compile/SDPA failure are distinct model-specific issues." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing encoder weights on load vs an unrelated save_pretrained shard-size question." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "Precision warning during bfloat16 compilation is not the same as a Bamba SDPA compile failure." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "Different warnings and different triggers: bfloat16 precision warning vs tie_word_embeddings warning." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline translation messaging bug vs inability to save Pipeline objects." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Dummy class cleanup/check_repo issue is unrelated to a noisy logging message." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Saving dequantized models and saving Pipeline objects are different persistence paths." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "A removed config field is not the same as a noisy logging message." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Unsupported Flash Attention on Nemotron is a model capability issue, not GPT2 attention scaling being ignored." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Unsupported Flash Attention and a Python-version compatibility issue are unrelated." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage and a deepseek tensor transpose bug do not appear to share a code path." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper empty-transcription generation failure is unrelated to a Trainer import conflict." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "PatchTSMixer post_init validation and Trainer import failure are different problems." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention scaling being ignored is not the same as degenerate zero loss in sequence classification." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "AttentionInterface.register behavior change is a different API regression than GPT2 backend scaling." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "FP8 save_pretrained behavior and Qwen3.5 MOE from_pretrained errors are distinct failures." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat completions API rejection and batched processor crash are different surfaces." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config vocab_size AttributeError and Apple Silicon allocator TypeError are unrelated." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training bugs and BeitImageProcessorFast reduce_labels behavior are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Tensorflow backend packaging and Trainer import conflict are not the same issue." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local-model trailing-slash warning and bfloat16 precision warning are different behaviors." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "Precision warning during compilation and compiled flex_attention failure are different bugs." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Missing is_offline_mode import and missing Qwen2_5_VLForConditionalGeneration export are distinct import errors." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "ImportError for layer_type_validation and local_files_only still making API requests 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": 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-10000-451", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:42757", + "issue:42994", + "issue:43010", + "issue:43064", + "issue:43066", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43525", + "issue:43611", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44075", + "issue:44077", + "issue:44230", + "issue: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:45593", + "issue:45600", + "issue:45698", + "issue:46077", + "issue:46082", + "issue:46172", + "issue:46209", + "issue:46298", + "issue:46365" + ], + "soft_pairs": [ + "issue:45600|issue:46298", + "issue:44038|issue:44230", + "issue:45698|issue:46172", + "issue:43618|issue:44380", + "issue:43618|issue:43761", + "issue:44483|issue:45290", + "issue:44625|issue:46365", + "issue:43299|issue:43867", + "issue:36683|issue:43475", + "issue:42994|issue:44655", + "issue:43066|issue:46209", + "issue:43756|issue:44485", + "issue:45276|issue:46077", + "issue:43901|issue:44655", + "issue:36010|issue:42757", + "issue:43611|issue:44655", + "issue:36683|issue:44077", + "issue:43352|issue:43937", + "issue:43867|issue:45216", + "issue:44625|issue:46077", + "issue:43976|issue:44938", + "issue:46077|issue:46082", + "issue:43352|issue:45593", + "issue:36683|issue:43525", + "issue:43618|issue:45216", + "issue:36683|issue:44242", + "issue:43992|issue:45310", + "issue:43064|issue:44492", + "issue:44075|issue:46365", + "issue:43010|issue:45278", + "issue:44355|issue:44855" + ], + "prepared_review_unit_hash": "4606d0334b26b9e327a3138ba8b6780f60505dd217df21bba20db6ad0fff43ce", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36465, + "estimated_input_tokens": 9117, + "estimated_eval_tokens": 18490 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is only loosely related to Transformers v5 regressions; most soft pairs point to different model-, API-, or serialization bugs rather than true duplicates. No soft edge looks safe to merge.", + "confidence": 0.23, + "canonical_issue_reason": "issue:45278 is the broadest umbrella in the set: it reports multiple import failures after upgrading Transformers, which best matches the cluster\u2019s overall regression-heavy theme.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the best representative issue because it is broad enough to summarize the cluster, while the others are narrower model-specific or feature-specific bugs.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45600", + "right": "issue:46298", + "accept": false, + "reason": "Unrelated: removed Sam3LiteText auto-mapping CI break vs a noisy logging message." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Both mention Qwen3-VL/MoE, but one is a general loading bug and the other is fp8 inference support; different underlying problems." + }, + { + "left": "issue:45698", + "right": "issue:46172", + "accept": false, + "reason": "Different areas: save/load of custom modules vs a DeepSeek-V4 tensor transpose bug." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "Both are regressions, but CLIP attentions missing is not the same bug as GPT-2 attention scaling being ignored in SDPA/FlashAttention." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Same model family, different symptoms and code paths: missing attentions vs missing hidden_states." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Chat API request validation is a different problem from apply_chat_template crashing on tool-call assistant messages." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Qwen3.5 config propagation bug vs an import-order issue involving datasets/Trainer." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Both are model-loading failures, but they involve different models and failure modes." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "AttributeError on Gemma3Config vs missing field on Sam3VisionEncoderOutput; unrelated models and attributes." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model saving failure is not the same as saving Pipeline objects with save_pretrained." + }, + { + "left": "issue:43066", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer decoder type bug vs FP8 inference failure on B200; unrelated." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Different RoPE-related model issues; same general area but not the same concrete bug." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Both Gemma4-related, but resize_token_embeddings behavior differs from bidirectional attention mask construction." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Docs mismatch for TextClassificationPipeline vs serialization of Pipeline objects." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols and different dependency packages." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "Model loading with base_model_prefix is not the same as saving Pipeline objects." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3Config missing vocab_size and patchtsmixer post_init validation are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Flash Attention support error for NemotronH vs invalid GenerationConfig for GLM-5 are different bugs." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both involve loading/saving checkpoints, but one is a state_dict ordering error and the other is a Qwen3.5 save_pretrained regression." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Qwen3.5 num_labels propagation is unrelated to Gemma4 bidirectional attention masks." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Different Python-version compatibility failures with no clear shared code path." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Same version window, but Gemma4 attention masks and LlamaConfig head_dim validation are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "FlashAttention support limitation vs D-FINE auxiliary loss logic; unrelated." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Different AttributeErrors on different model configs." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP output regression and Qwen3.5 checkpoint saving regression are unrelated." + }, + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Gemma3Config vocab_size missing is unrelated to MoE load-balancing loss omission." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Both concern Qwen3.5, but one misses embed_tokens.weight and the other has a from_pretrained error; not the same concrete bug." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Optimizer state corruption under FSDP2/PEFT is unrelated to a cache strategy typo." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "SGD args ignored is an optimizer bug; importing Trainer being blocked by datasets is a different import/dependency issue." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "No_grad decoration for update methods is unrelated to import errors after upgrading Transformers." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Compiled Python file handling is unrelated to the Python 3.13 DebertaV2Model parsing error." + } + ] + }, + "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-10000-451", + "nodes": [ + "issue:28282", + "issue:30333", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42548", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43065", + "issue:43232", + "issue:43502", + "issue:43575", + "issue:43576", + "issue:43618", + "issue:43723", + "issue:43867", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44246", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44387", + "issue:44485", + "issue:44556", + "issue:44749", + "issue: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:46202", + "issue:46285" + ], + "soft_pairs": [ + "issue:43867|issue:46202", + "issue:43867|issue:45375", + "issue:41669|issue:44246", + "issue:44938|issue:45071", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:43010|issue:45850", + "issue: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:45125|issue:45375", + "issue:43010|issue:43232", + "issue:43502|issue:44222" + ], + "prepared_review_unit_hash": "fba921e4e0f07433ba7038ca95ce222eba0e81f99df9c1a2904b2d903102dea8", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36540, + "estimated_input_tokens": 9135, + "estimated_eval_tokens": 18526 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are largely a grab-bag of unrelated regressions and bugs. None of the proposed soft pairs looks like the same underlying problem closely enough to merge as duplicates, so I reject all soft edges. If a single issue must stand in for the set, the broad v5 reload regression report is the best representative, but only weakly so.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 44556 is the broadest, most user-facing regression in the set: a version-upgrade load/reload failure spanning v4.57 to v5.2/v5.3. It is still not a true duplicate of the others, but it is the least niche representative.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44556 best captures the general class of v5 loading regressions and is the most plausible umbrella representative among these unrelated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Both mention loading-related failures, but one is a generic state_dict loading error while the other is a Gemma pipeline roundtrip/tokenization issue. Different code paths and root causes." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "State_dict load sorting error and a missing Qwen3.5 vision config field are different bugs in different components." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Import-time slowdown from import * cleanup and intermittent import slowness are both performance-related, but not the same change or failure mode." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 import/load failure is an environment compatibility issue; the other is PretrainedConfig type checking in v5.4.0. Not the same bug." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config missing vocab_size and FSDP2/PEFT optimizer state corruption are unrelated model/config vs training-distribution bugs." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "A flaky SamHQ integration test due to missing set_seed differs from a pytest-xdist race in captured_info.txt." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "A cache/layer @torch.no_grad issue is a specific implementation bug; the other is a broad complaint that v5 breaks too many models." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "UMT5Encoder missing embed_tokens.weight and Qwen3.5 Moe from_pretrained errors are different model-loading failures." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Generic load error with sorted state_dict and incorrect visual encoder save_pretrained keys are different save/load bugs." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Cache update decoration and Python version compatibility are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor parallel training bug and OOM when loading Qwen2-57B-A14B-Instruct with tp are both TP-related, but the reported failures and likely fixes differ." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIPOutput attentions missing assignment and patchtsmixer optional post_init restriction are separate model API issues." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch-not-found ImportError and tokenizer loading regression are both import/load themed, but not the same underlying defect." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "Missing CLIPOutput attentions and inconsistent CLIP text model structure are distinct CLIP API regressions." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "Missing attentions assignment and LwDetrImageLoss AMP CUDA crash are unrelated model-output vs training/numerics bugs." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "A DSA indexer ReLU discussion and DeepseekV3MoE divergence are different implementation concerns." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "local_files_only still making API requests is a download/offline bug; AddedToken special kwarg duplication is a tokenizer construction bug." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Seed-related test flakiness and Phi-4 multimodal model loading failure are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "A missing set_seed in a test and a stale GraniteSpeech device override test failure are different CI/test problems." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "Seed flakiness and auto_mappings referencing removed Sam3LiteText configs are unrelated test vs mapping issues." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Both are loading regressions after v5, but one is tokenizer loading and the other is checkpoint reload across versions; different concrete failures." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization OOM from higher reserved CUDA memory is a memory-management regression, not a checkpoint reload failure." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained for moe and missing _tp_plan for tensor parallelism are different model export vs parallelism metadata bugs." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Torch.no_grad decoration for cache/layer updates and a dummy nn.Conv2d in Sam3PixelDecoder are unrelated implementation changes." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Broken transformers env command and AttentionInterface.register behavior change are separate CLI vs attention API issues." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Offline download leakage and GLM-5 RoPE implementation concern different subsystems and failure modes." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "cannot import PreTrainedModel and cannot import HybridCache are both import errors, but they affect different symbols and likely different export paths." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Missing tensor-parallel plan for Qwen3_5MoeForConditionalGeneration and Qwen3.5 Moe from_pretrained error are related only by model family, not the same bug." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file loading errors and slow filtering after upgrade are unrelated." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLFlow not reporting failed training jobs and tokenizer loading regression are entirely different problems." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Missing _tp_plan and missing deepstack_visual_indexes are different Qwen3.5 configuration/model export issues." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache update no_grad decoration and generation kwargs sync handling are different generation internals." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "local_files_only offline leakage and FP8 save_pretrained moE are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 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-10000-451", + "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:43299|issue:44556", + "issue:43502|issue:44961", + "issue:42831|issue:43874", + "issue:30333|issue:36296", + "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:43576|issue:44938", + "issue:44222|issue:44230", + "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": "3af8d5b294438fee42a60e4f8195947bf5d8b2795d3a6e7ea2583c24e1c1320b", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36411, + "estimated_input_tokens": 9103, + "estimated_eval_tokens": 18462 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a grab bag of Transformers v5-era regressions, import failures, model-loading issues, and unrelated docs/feature bugs. Only the RoFormer pair looks like the same underlying code-path problem; the rest are too different to merge as duplicates.", + "confidence": 0.41, + "canonical_issue_reason": "issue:45850 is the broadest umbrella in the set, capturing the shared theme of widespread v5 regressions across models and loading paths, even though it is very generic.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the best single representative because it summarizes the overarching cluster theme better than the more specific bugs, but it is still too vague to serve as a precise canonical fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Loading-stuck bug vs model.forward documentation issue; different problem and code path." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are reload regressions, but one is Qwen3VL-Moe-specific and the other is a generic v4.57\u2192v5.2/5.3 checkpoint reload failure." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "Local-files-only network leak vs unrelated placeholder issue; no shared bug." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "FP8 accuracy regression vs missing image-processor method causing multimodal token counting failure; different issues." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLflow reporting failure after job crash vs tensor-parallel training bug; unrelated." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-file execution error vs PretrainedConfig type-checking breakage; distinct regressions." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs still referencing pipeline() vs unrelated placeholder issue." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled Python file errors and save_pretrained pipeline serialization are separate bugs." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel application bug vs DSA indexer/ReLU discussion; unrelated." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer-loading regression vs Llama 3.2 90B Vision model-loading regression; similar symptom, different failure." + }, + { + "left": "issue:27957", + "right": "issue:33453", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta vs tokenizer-loading regression; different code paths." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing huggingface_hub import vs broken transformers env command; both are import/CLI symptoms but not the same bug." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "Broken env command and PretrainedConfig type-checking are separate regressions." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "env CLI breakage vs missing Qwen2_5_VL export; different export surfaces." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Documentation confusion vs ImportError for layer_type_validation; unrelated." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "Placeholder issue vs Gemma-4 audio positional encoding bug; no overlap." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "CLI env issue vs Python 3.14 load failure; too broad to be the same bug." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained bug vs FP8 inference support request; related theme, but not the same concrete defect." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Different missing top-level imports; likely separate export regressions." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": true, + "reason": "Both point to the same RoFormer forwarding/interface problem: kwargs are not propagated because the attention implementation bypasses the expected interface." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs bug vs unrelated placeholder issue." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion vs unrelated placeholder issue." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug vs pipeline docs mismatch; unrelated." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Model-init Liger Kernel application bug vs FP8 inference failure on B200; different failures." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor-parallel training bug vs BeitImageProcessorFast reduce_labels bug; unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Missing GenerationMixin cache in chunked_prefill vs documentation typo; not the same issue." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "Specific Qwen3.5-35B chat crash vs broad complaint about many v5 breakages; not a duplicate." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Processor save_pretrained missing files vs generic bug report placeholder; no evidence of sameness." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading failure vs undefined tf backend error when only tensorboard is installed; unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Qwen3-VL-Moe v5 bug vs missing _tp_plan for Qwen3.5Moe tensor parallelism; same family, but not the same concrete code-path problem." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support request vs DeepSeekV4 weight-conversion question; unrelated." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output problem vs PretrainedConfig type-checking breakage; different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet: issue:45850 is a reasonable broad canonical issue, and the cluster is accurately described as mostly heterogeneous v5 regressions rather than a tight duplicate set. The soft-edge calls are generally conservative, and the only accepted pair (46143/46144) is plausibly the same RoFormer interface/kwargs defect from the titles alone." + }, + "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": 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-10000-451", + "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:44295|issue:45230", + "issue:43352|issue:44393", + "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": "11951b5bfa07b38257532f90d8dd76794deb3d82c8978347926045f27e831bc6", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 42, + "serialized_chars": 39545, + "estimated_input_tokens": 9887, + "estimated_eval_tokens": 20030 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Cluster is mostly a grab bag of Transformers v5 regressions, import/loading issues, and model-specific bugs; there are a few thematic similarities, but no clear duplicate pair.", + "confidence": 0.66, + "canonical_issue_reason": "issue:45850 is the broadest umbrella for the cluster\u2019s recurring \u201cv5 breaks models\u201d theme, even though it is very generic.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45850 is the best representative of the cluster because it captures the common version-regression theme shared by several nearby reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Related import/lazy-loading area, but different bugs: intermittent import slowness vs unsafe `sys.modules` access." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Different failures: broken `transformers env` command vs checkpoint reload compatibility after upgrade." + }, + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Same model family, but different symptoms and code paths: bbox output bug vs image-specific infinite loop." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Specific `Llama4Config` attribute error vs a broad complaint about many models breaking." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-star cleanup/perf issue is unrelated to missing cache in chunked prefill." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both are v5 loading regressions, but for different models and different root causes." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "Model-specific missing `pad_token_idx` issue vs a vague global breakage report." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "Cache-strategy typo is unrelated to the broad v5 regression complaint." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Both describe v5 breakages, but not the same concrete failure or fix." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API migration issue has nothing to do with a cache-strategy typo." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Both touch imports, but one is import-time bloat and the other is a missing symbol." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Missing `set_seed()` in tests is unrelated to scheduler kwargs handling." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Model-type mismatch warning vs Liger Kernel not being applied under `model_init`." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer save metadata mismatch vs a too-generic bug report with no concrete match." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Same import/lazy-loading area, but one is intermittent slowness and the other is a broader lazy-loading failure." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Buffer access error vs generic bug report; no specific overlap." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Different model bugs and symptoms: Flash Attention 2 support error vs Qwen3-VL bbox issue." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env CLI and scheduler kwargs are unrelated." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwargs forwarding bug vs scheduler kwargs bug." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python-version compatibility regression vs Qwen3-VL output bug." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs bug is unrelated to `PretrainedConfig` type-checking regression." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import performance cleanup vs missing `HybridCache` symbol; different underlying problems." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Cache-strategy typo vs feature-extractor padding inefficiency." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import cleanup/perf issue vs a separate testing_utils bug." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Python support regression vs a broad claim that v5 breaks many models." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different Qwen3.5-related problems: vision bbox hallucination vs from_pretrained error." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "PatchTSMixer post-init policy issue is unrelated to `_tied_weights_keys` list handling." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python compatibility issue vs missing Qwen3_5 import; unrelated." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still mention pipelines, but the other issue is about scheduler kwargs." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion in `forward` vs tokenizer loading regression." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application issue vs flash_attention_3 training degeneration." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Env command failure vs lazy-loading issue; same broad area, but not the same bug." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Pipeline save_pretrained issue vs tied-weights crash; no concrete match." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "`45230` is too generic to establish the same bug as the Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA/ReLU discussion vs Gemma-4 audio positional-encoding bug; unrelated." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy-loading/import issue vs scheduler kwargs handling; different code paths." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Broken env CLI vs missing model import; unrelated." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp calculation bug vs GLM-5 RoPE implementation issue." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeek-V4 shared-expert gating bug vs a separate tensor transpose bug." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp calculation bug vs scheduler kwargs handling; unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-star cleanup/perf issue vs VRAM leak in multi-threaded dataloading." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Confusing docs issue vs Llama 3.2 Vision model-loading regression." + } + ] + }, + "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-10000-451", + "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": "569e822d64c6453389c7d4c69dceb57fb57c246e504fe0c434082cd440e02159", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862653, + "estimated_input_tokens": 215664, + "estimated_eval_tokens": 431584 + }, + "budget": { + "node_count": 44, + "item_count": 44, + "soft_pair_count": 42, + "serialized_chars": 37114, + "estimated_input_tokens": 9279, + "estimated_eval_tokens": 18814 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most items are unrelated regression reports, doc issues, or model-specific bugs, so none of the soft edges look like true duplicates. No PRs are present; the closest thing to a thematic umbrella is a vague v5-regression complaint, but it is not a safe duplicate hub.", + "confidence": 0.96, + "canonical_issue_reason": "No strong canonical issue exists because the items span unrelated subsystems and failure modes. If one must be chosen, issue 45850 is the broadest umbrella for the v5-regression theme, but it is too vague to be a true duplicate target.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45850 is the only broad, umbrella-style report in the set, so it is the closest fit for a cluster representative. That said, it is still too generic to serve as a high-quality canonical duplicate.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Checkpoint loading slowdown and documentation confusion are different problem types." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in Qwen3 causal LM is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "GLM-5 RoPE math and Gemma-4 audio positional encoding are different model-specific bugs." + }, + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Continuous batching for audio models is unrelated to DeepSeek-V4 weight conversion support." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "A broad v5 regression complaint is not the same as a specific import error for Qwen3.5." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "Local-files-only network calls and Qwen3VL timestamp calculation are different code paths." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 configuration guidance is unrelated to a VRAM leak bug." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Timestamp calculation and missing processor save files are separate processor issues." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "Qwen3VL timestamp logic and Qwen3 causal LM VRAM leakage are not the same bug." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch and Gemma-4 audio positional encoding do not share a concrete failure mode." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "torch.split return handling in GlmMoeDsaIndexer is unrelated to Gemma-4 audio positional encoding." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "Missing rms_norm_eps in MLA q_a_layernorm is a different bug from Gemma-4 audio encoding." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "HybridCache import failure and Qwen3.5 module import failure are separate missing-symbol issues." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention support in NemotronH is unrelated to Qwen3.5 importability." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both mention tokenizer v5 behavior, but one is a broad behavior change and the other is a specific decoder-type bug." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler kwargs handling is unrelated to the unrelated 'Unemployment' issue." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A cache-strategy typo and an unrelated issue titled 'Unemployment' do not match." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "The broad v5 regression complaint does not describe the same bug as the unrelated 'Unemployment' issue." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to the 'Unemployment' issue." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weights key handling and chat template tool-call crashes are different failures." + }, + { + "left": "issue:44998", + "right": "issue:45341", + "accept": false, + "reason": "The 'Unemployment' issue and a testing_utils bug are unrelated." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config is unrelated to tied-weights key crashes." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig is a config issue, not tied-weight key handling." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow checkpoint loading and a missing import symbol are unrelated." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support and processor save_pretrained file output are different topics." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLflow job-status reporting and checkpoint loading speed are unrelated." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "Checkpoint loading performance and tensor-parallel training bugs are different problems." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 support request and Qwen3.5 from_pretrained failure are not the same underlying issue." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Checkpoint loading slowdown and tokenizer-loading regression are separate issues." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models is unrelated to FP8 inference support." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init_weights behavior and RoFormer attention-interface usage are different model implementation bugs." + }, + { + "left": "issue:42936", + "right": "issue:46018", + "accept": false, + "reason": "Mask2Former ignore_value handling and DeepSeek-V4 shared-expert gating are unrelated." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "ModernBERT norm_eps typing and tokenizer regex warnings are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2Former ignore_value and loss-function kwargs passthrough are different code paths." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss kwargs passthrough and RoFormer attention-interface behavior are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "A model-specific ignore_value bug does not match a global tokenizer regex warning." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Mask2Former loss handling and RoFormer attention implementation are different bugs." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Ignore_value usage and a typo in quantization operation naming are unrelated." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss kwargs passthrough and config type hints are unrelated." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "ModernBERT norm_eps typing and a quantization-operation typo are different issues." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "A Llama4 norm question and a ModernBERT type-hint bug are unrelated." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value handling and ModernBERT norm_eps typing 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": 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": "b2abcc9379be3cc77abe5fd2ea0dfaebe4795400f1064d4a1f527edc69169032", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2876, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three issues are unrelated feature requests in different areas: tokenizer truncation behavior, learning-rate scheduling functionality, and Wav2Vec2 word boundary offsets. No duplicate cluster emerges from the titles or descriptions.", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical duplicate: the issues describe distinct feature requests with different code paths and no shared targets.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:10519 is the most representative only by slightly broader wording and discussion, but it is still unrelated to the other two; if one issue must anchor the cluster, this is the least specific.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Tokenizer truncation options and Wav2Vec2 word offset extraction are unrelated problems." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "Truncation-from-beginning behavior is unrelated to ReduceLROnPlateau-like scheduler functionality." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "Learning-rate scheduling features and Wav2Vec2 word offset timing are different subsystems and not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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": "6a61ee3b18fc0c0a7876602dc51e8fab3e76b2aceacce1923f3b8eba0b928801", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2478, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These three issues do not appear to be true duplicates: one is about Tapas tokenization deviating from TensorFlow, one is about a misleading auth/download error for private models, and one is about a generic inability to load bert-base-uncased. The soft links are too broad to merge confidently.", + "confidence": 0.89, + "canonical_issue_reason": "No exact duplicate stands out; if a single anchor is needed, issue:16618 is the broadest and most central model-loading report, with the most discussion activity.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:16618 is the best cluster anchor because it is the most general user-facing failure report and could subsume nearby model-loading problems better than the more specialized tokenization or auth-message issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Different underlying problems: Tapas tokenization parity with TensorFlow vs a generic bert-base-uncased loading failure. Same ML area, but not the same bug." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "This is about an error message when downloading a private model without login, not the same concrete failure as loading bert-base-uncased." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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": "78e8a0417f12fdb9243cb6b8276913359cac768ddbf3622671c47b42564fa50f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate feature-request issues: one for Flash Attention 2 architecture support and one for GGUF architecture support. They share a loose \u201ccommunity contribution\u201d framing but not the same underlying change or code path.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 26350 is the better representative issue only because it is older and has substantially more discussion activity; however, it is not a duplicate of 33260.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "Different requested capabilities (Flash Attention 2 vs GGUF) and different implementation areas; the shared wording is only a broad support request, not the same bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 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": "9f38b887dc888240898ea2a40c8b8a639f18aa85bac6ca53c242f896f4cff4b0", + "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": "The two issues are only loosely related at the level of training failing to start, but they do not clearly describe the same underlying bug. One is a specific `resume_from_checkpoint`/epoch iterator failure, the other is a broad FineWeb SLM training startup problem.", + "confidence": 0.87, + "canonical_issue_reason": "Issue 26413 is the better canonical issue because it names a concrete failure mode (`resume_from_checkpoint` causing an epoch_iterator error) and is more specific and actionable than the broader training-start symptom in 31501.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 26413 is the best representative because it targets a distinct code-path and failure message, making it easier to map to a single fix or regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "Both concern training problems, but the evidence is too weak to treat them as the same bug; 26413 is specifically about checkpoint resumption and an empty epoch iterator, while 31501 is a broad startup failure with no clear match to that 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": 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": "a577970bda1f6ae0f151100d2b91000a313d229e7b86a584d90338db695a8e34", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13530, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13530, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster looks like one soft-similarity group around the generic i18n template issues plus a very similar Thai-text pair, but the packet does not support treating any pair as confirmed duplicates because every edge is marked deterministic_accept:false. The Turkish, Spanish, and Romanian issues are best treated as separate language-specific requests rather than one merged underlying issue.", + "confidence": 0.66, + "canonical_issue_reason": "issue:43834 is the best representative of the generic-template i18n subgroup because it is one of the earliest near-identical placeholder issues, but this should be treated as a soft representative only, not a confirmed duplicate anchor.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43834 is the strongest global representative of the cluster\u2019s generic-template pattern, since it sits at the center of the near-identical placeholder issues and is the cleanest summary node for triage.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Same broad i18n topic, but Turkish vs generic placeholder request; not enough evidence for a true duplicate." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Turkish language request does not match the generic placeholder issue as the same underlying report." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Turkish request and generic template issue are only loosely similar, not the same issue." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Turkish and Romanian translation requests are distinct language-specific issues." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish and Spanish translation requests are different language targets, so not duplicates." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish request is only soft-similar to the generic template issue, not the same underlying report." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Spanish request and the generic placeholder issue are similar in structure but not clearly the same issue." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish and Romanian requests are separate language-specific issues." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Spanish request is not the same concrete issue as the generic placeholder report." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian request and generic template issue are only soft-similar, not confirmed duplicates." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian request does not appear to be the same underlying report as the generic template issue." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian request is only structurally similar to the generic placeholder issue." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Romanian request is unrelated to the Thai-text issue." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Romanian request is unrelated to the Thai-text issue." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": false, + "reason": "These are very similar generic-template issues, but the packet still only gives soft similarity, not a confirmed duplicate signal." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": false, + "reason": "Near-identical placeholder structure, but not enough evidence to merge as confirmed duplicates." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Generic template issue vs Thai-text issue; different content and not the same request." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Generic template issue vs Thai-text issue; different content and not the same request." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": false, + "reason": "Strong textual similarity, but still only soft-similarity evidence in the packet." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Generic template issue and Thai-text issue are different underlying reports." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Generic template issue and Thai-text issue are different underlying reports." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Generic template issue vs Thai-text issue; not the same concrete bug/request." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Generic template issue vs Thai-text issue; not the same concrete bug/request." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": false, + "reason": "These Thai-text issues are highly similar, but the packet does not justify upgrading that similarity to a confirmed duplicate." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The clustering summary is too strong relative to the packet. The packet only provides title/body similarity and shows `deterministic_accept: false` for every listed pair, so statements like \"true duplicate subgroup,\" \"same underlying request,\" or exact duplicate acceptance between the generic template issues and the Thai pair are not grounded. In particular, marking `issue:46222|issue:46226` as accepted and treating the generic template issues as confirmed duplicates overstates the evidence. The conservative part about the Turkish/Spanish/Romanian issues being distinct language-specific requests is reasonable, but the overall verdict should stay at the soft-similarity level." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "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": "dd024e2b05e7aa45e9a7fc264b29f34b5feb98ca483af56e584430f86394bae5", + "component_budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21658, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21658, + "estimated_input_tokens": 5415, + "estimated_eval_tokens": 11086 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This is a loose thematic cluster of model-support requests, not a duplicate set. Most items ask to add different models or modalities (text encoders, detectors, ASR/TTS, OCR, vision-language), so they should stay separate.", + "confidence": 0.56, + "canonical_issue_reason": "issue:43671 is the best cluster anchor: it is a concrete model-support request and sits closest to the other speech/audio-related items, but it is still not a duplicate of them.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43671 is the most representative anchor for this loose group of model-add requests, with the strongest overlap in theme and terminology among the speech/audio proposals.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different model requests: EuroBert config support vs DEIMv2. Same broad 'add model' pattern, but not the same change." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert is a text model/config request; Qwen3-TTS is a speech synthesis request. Different underlying code paths." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert and RF-DETR are unrelated model families (text encoder vs object detector)." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "Both are detection-related, but RF-DETR and DEIMv2 are distinct detector models with separate integrations." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR is an object detector; LFM2.5 Audio 1.5B is an audio model. Different functionality." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR and the weight-sparse transformer request are different architectures and different implementation targets." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "Object detection support vs ASR support are different model capabilities and code paths." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are unrelated tasks: detection vs text-to-speech." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are different model families and different modalities." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT support vs a sparse-transformer model request are distinct changes." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT is a text model; LFM2.5 Audio 1.5B is an audio model." + }, + { + "left": "issue:36879", + "right": "issue:37015", + "accept": false, + "reason": "RF-DETR and NeoBERT are unrelated model additions." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are different model requests, not the same underlying fix." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture support and Qwen3-ASR support are different model integrations." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and the OpenAI sparse-transformer request are unrelated models." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "Both are model-add requests, but MiniCPM3 and ModernVBERT are distinct families and likely separate implementation work." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 support and Voxtral TTS support are different modalities and code paths." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 detection support and Qwen3-TTS support are different model types." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT and Qwen3-ASR are unrelated model families and modalities." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT is a BERT-style text model; Qwen3-TTS is speech synthesis." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT and 'BERT-like models with RoPE' are only superficially related; one is a specific model request, the other is a general architectural pattern." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "General BERT-like RoPE discussion does not match Voxtral TTS support; different model classes." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Same model family, but TTS and ASR are different capabilities and would not be merged into one PR." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Both are speech models, but Qwen3-TTS and Voxtral-TTS are different model families with separate integrations." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "TTS support and DeepSeek-OCR2 support are different modalities and changes." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Speech model support vs sparse-transformer support are unrelated changes." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Both are audio-related, but Qwen3-TTS and LFM2.5 Audio 1.5B are distinct models/capabilities." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "ASR support and LFM2.5 Audio support are not the same underlying model change." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "ASR support and EUPE perception-encoder support are different model requests." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "ASR support and a weight-sparse transformer request are unrelated." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "Audio model support vs SAM3-LiteText support are different modalities and integrations." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Sparse-transformer support and EUPE support are different underlying changes." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "ASR and TTS are distinct speech tasks; these would not be one merged fix." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral-TTS are unrelated model families." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "ASR support and OCR2 support are different code paths and model types." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are separate vision-language model requests; related broad area, but not the same concrete change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "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": "b46e848e79ebfdefd15f150a66bd8098cc95365a74a19ae402de11e8a870f6cf", + "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 do not look like duplicates: one asks for generic block decoding changes, while the other requests support for a specific model version (DeepSeek v3.2).", + "confidence": 0.95, + "canonical_issue_reason": "No canonical issue: the titles/body indicate different topics rather than the same underlying bug or feature request.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43037 is the more concrete and user-facing request, but it is still not a duplicate of 41084.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "The requests are unrelated at the level needed for duplicate triage: block decoding vs. DeepSeek v3.2 support." + } + ] + }, + "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": "bb8ba802e3d3406b3870d9fea03e377faad2b70268cabc37cffe7697f6bc6ed2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1583, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1583, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related thematically around Trainer customization, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPs estimation for custom trainers. I would not merge them as duplicates.", + "confidence": 0.91, + "canonical_issue_reason": "Issue #43598 is the broader, more representative Trainer customizability request; it centers on a core Trainer behavior rather than a narrower metrics/estimation edge case.", + "canonical_pr_reason": null, + "best_issue_reason": "#43598 is the best issue to keep as the cluster representative because it addresses a foundational Trainer customization point and is phrased as a general design concern.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "They share the general theme of custom trainers in `Trainer`, but the underlying problems are different: one is about column removal/customizability, the other about FLOPs counting. They would not be fixed by the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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": "a3192c02b5be04c25afe036aa839dbd523d11d62078a4ad42ce8e5760883af27", + "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": "The two issues are about different Qwen model families and different API surfaces: one requests Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration, while the other requests AutoModelForSequenceClassification support for Qwen3.5. They are related only at a broad \"Qwen model support\" level, not the same underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43684 is the broader model-registration request, so it is the better cluster anchor, but it is still not a duplicate of 44405.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different model variants and different heads/APIs; this is not the same bug or change and would not plausibly be merged as one issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 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": "5623e52b10313574a62cb62f7b95cd6d9a207e33f2b4a76c59da6690b07dc1cd", + "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 issues are near-identical low-signal \"huggingface\" reports; one \"Fett\" issue appears unrelated. I treat 43845 and 43846 as the clearest duplicates, while 43859 is too under-specified to confidently merge with them from metadata alone.", + "confidence": 0.78, + "canonical_issue_reason": "issue:43845 is the best representative of the duplicate \"huggingface\" pair: it is the earliest of the two identical-title, identical body-length reports and has the clearest match to issue:43846.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43845 is the strongest global issue representative because it anchors the only convincingly duplicate pair (43845/43846) and is more complete than 43859.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no shared concrete evidence beyond generic similarity." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no shared concrete evidence beyond generic similarity." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles and no shared concrete evidence beyond generic similarity." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Same exact title and same body length, which strongly suggests the same underlying issue." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "Same broad title, but the shorter body and later activity make this too weak to treat as the same issue from metadata alone." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "Same broad title, but the shorter body and later activity make this too weak to treat as the same issue from metadata alone." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet: 43845/43846 share the same title and body length, while 43835 is clearly different and 43859 is weaker evidence from metadata alone. The soft-edge verdicts are conservative overall, and the single accepted duplicate pair is supported by the strongest available signals." + }, + "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": "99c954f4682a076783c6f8338c4953f5c90d346194c4ca447711b14e722b6731", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is dominated by low-information spam/generic placeholder issues with no substantive bug or feature signal. The only apparent overlaps are superficial title matches, not evidence of the same underlying issue.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "No issue in the set meaningfully represents a real problem statement; all are either spam placeholders or an unrelated generic documentation suggestion.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are titled \"\", but that is only a superficial match; there is no concrete bug, target, or code-path evidence to show they are the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Identical spam titles do not establish a shared defect or change request; no supporting body or references indicate the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam and the other is a generic documentation suggestion; these are not the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "A spam placeholder and a titled issue about \"Google LLC\" do not indicate the same bug or change." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "A documentation enhancement request and a spam placeholder are unrelated." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "A documentation enhancement request and a spam placeholder are unrelated." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam placeholders with no concrete problem description; identical titles alone are insufficient to treat them as the same underlying issue." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "A spam placeholder and an unrelated titled issue do not match on any concrete defect or change." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "A spam placeholder and an unrelated titled issue do not match on any concrete defect 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": 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": "f86c200bb5cc212e2035e96a5ed8f510eb9c0cb7648c5a3bce1f52b95ad34f1d", + "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": "These two closed issues are very likely duplicates of the same report: they have identical titles, very similar body lengths, and a soft-similarity signal. #45145 is the better canonical issue because it was created first and has the stronger reference activity.", + "confidence": 0.84, + "canonical_issue_reason": "#45145 is earlier and has the stronger reference activity, so it is the best canonical issue if these are treated as duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the strongest representative because it predates #45151 and is the better anchor for a probable duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": true, + "reason": "Identical titles plus similar body size and a soft-similarity signal make this a probable duplicate, though the body text is not visible here so the match is not fully provable." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The conclusion is plausible, but the writeup overstates the evidence. The packet only shows identical titles, similar body lengths, and a soft-similarity signal; it does not expose the body text needed to claim the reports are 'highly similar' or that they are definitively the same underlying issue. The canonical-choice reasoning is fine, but the soft-edge verdict and summary should be phrased more conservatively (e.g. 'likely duplicate' or 'probable duplicate') and the confidence should be lowered." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "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": "6eac56f30fd4015113e8620f5457ea5c4a07a42522a5f79094f263f96e1459c5", + "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": "These two issues are related only at a high level (4-bit inference / bitsandbytes), but they describe different failure modes and code paths: one is a Gemma4 audio-module uint8/`torch.finfo()` TypeError, the other is a `BitsAndBytesConfig.llm_int8_skip_modules` regression that removes the default `lm_head` exclusion and triggers an AssertionError. They should not be merged as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": "Issue #45674 is the better canonical issue for this cluster because it has the clearer, configuration-centric regression description and appears broader/more reusable as a report of a BitsAndBytesConfig bug.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #45674 is the strongest representative issue: it states the concrete failing behavior, the triggering option, and the 4-bit inference context more explicitly than #45672.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Both involve 4-bit inference, but the bugs are different: #45672 is a Gemma4/audio-module `torch.finfo()` TypeError on uint8 weights, while #45674 is an `llm_int8_skip_modules`/`lm_head` exclusion regression causing AssertionError. Different underlying code paths and fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 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": "4b6726ebcc38c080f374d0f065765c5eb44730f84d4173ecf246af7d1704d230", + "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 closed issues are superficially similar only in that they are both empty, name-like entries; they do not describe the same bug or change and should not be clustered as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": "Issue 45762 is the only reasonable placeholder canonical because it was created first, but it is not a substantive report and has no content to represent a real bug cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45762 is marginally preferable as the cluster representative due to earlier creation, but neither issue is suitable as a true canonical bug report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "Reject: the titles are different personal-name/date strings, with no body, targets, or discussion tying them to 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": 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": "c261127724d3050348ceb6d3ccbf1dc51c2f857c99c96845f5766dca2b326f1a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1481, + "estimated_input_tokens": 371, + "estimated_eval_tokens": 998 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues appear unrelated: both are personal-name/date strings with no body or linked context, so there is no evidence they describe the same bug or request.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "Different named/date-only issue titles with no shared context, body, or links; not enough evidence they are duplicates." + } + ] + }, + "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": "dc9c24502f0d75bef955755724e8489d63dcbb8ca3bd69b464e736e6e6a2d0c8", + "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": "These are unrelated open feature requests: one asks for Qwen3_5 token-classification support for value-model use, the other asks for FP8 training support.", + "confidence": 0.96, + "canonical_issue_reason": "issue:46295 is the broader, more general request, so it is the better cluster anchor if one must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:46295 is the more generally applicable issue, but there is no evidence these two are duplicates or the same underlying problem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "These requests concern different areas of the stack and different changes: model-head support for token classification versus FP8 training infrastructure. 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": "84c7f83c4b30ecd2a58ae76043daf86cc1eb963765f5fd019786411290962b9b", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5723, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5723, + "estimated_input_tokens": 1431, + "estimated_eval_tokens": 3118 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Best-effort split: the packet looks like two likely duplicate issue pairs (45870/45871 and 45872/45873) plus one separate dataset request (45869). The duplicate evidence is strong but still soft, so I would treat the pairs as likely/near-identical rather than confirmed exact duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "No single canonical issue for the whole cluster. Best-effort grouping suggests two near-identical issue pairs (45870/45871 and 45872/45873) and one distinct issue (45869), but the evidence is soft rather than definitive.", + "canonical_pr_reason": null, + "best_issue_reason": "45870 is the best representative if one issue must be chosen, because it is the earliest issue in the near-identical 45870/45871 pair. It should be treated as a representative for that pair only, not for the full cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "Different request type: dataset bundle vs. financial time-series example. Related theme only, not the same underlying issue." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "Different request type: dataset bundle vs. financial time-series example. Related theme only, not the same underlying issue." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Near-identical issues: same title, matching body length, and strong soft-similarity signals. Likely duplicates." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Same broad DataForge/financial theme, but the concrete ask is different: prediction example versus scraping service." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Same broad DataForge/financial theme, but the concrete ask is different: prediction example versus scraping service." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Same broad DataForge/financial theme, but the concrete ask is different: prediction example versus scraping service." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Same broad DataForge/financial theme, but the concrete ask is different: prediction example versus scraping service." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Near-identical issues: same title, matching body length, and strong soft-similarity signals. Likely duplicates." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster split is plausible, but the wording is slightly too strong for the packet. The evidence only shows soft similarity plus matching titles/body lengths, not confirmed exact duplicates. Please soften the verdicts to \"near-identical\" or \"likely duplicates\" rather than asserting exact duplication, and keep the canonical-issue statement framed as a best-effort split rather than a definitive duplicate grouping." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "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": "5fe63c347924732467957d5b352624b66714f24ecde2a50a7fc11460dd85690c", + "component_budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 22, + "serialized_chars": 43093, + "estimated_input_tokens": 10774, + "estimated_eval_tokens": 21804 + }, + "budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 22, + "serialized_chars": 43093, + "estimated_input_tokens": 10774, + "estimated_eval_tokens": 21804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the PRs cover unrelated model additions, refactors, docs, and bug fixes. The only clear duplicate-style overlap is the fixed_cross_entropy kwargs pair (43251/43254), which targets the same loss-path change and the same tracking issue.", + "confidence": 0.84, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best representative because it implements the concrete cross_entropy kwargs fix, explicitly references the tracking issue, and is effectively the same change as 43254.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest global representative: it fixes a specific runtime bug in fixed_cross_entropy and is the only PR in the set with a true duplicate partner.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Unrelated changes: new VideoPrism model addition vs. an executorch dynamic-shape test." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Tokenizer API additions are unrelated to executorch export testing." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Mistral tokenizer helper methods and adding MobileLLM are different model areas and different changes." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor robustness changes are not the same underlying bug as adding a new model." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss-function kwarg forwarding is unrelated to adding MobileLLM." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Different code paths: cross-entropy kwargs fix vs. Mistral4 test repairs." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch export coverage and Mistral4 test fixes are not the same change." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "No concrete overlap beyond both touching model-related code." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "A new model addition and a config-resolution fix are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "New model implementation vs. Falcon output-interface refactor are unrelated." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactoring and the model-creation skill are not the same code-path change." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon refactor and Mistral4 test fixes are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46344", + "accept": false, + "reason": "Mistral4 test fixes and VitPose graph-break fix target different bugs in different models." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "A typo rename in CLIP-like model conversion is unrelated to Mistral4 tests." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 tests and chat parsing/serve updates do not share the same underlying change." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46367", + "accept": false, + "reason": "Different fixes: Mistral4 tests vs. FP8 kernel integration changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "The model creation skill is a meta/documentation PR, not the same bug or code change as Mistral4 tests." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "Adding MobileLLM is unrelated to the model creation skill PR." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 test fixes and config resolution ordering are different issues." + }, + { + "left": "pull_request:46048", + "right": "pull_request:46367", + "accept": false, + "reason": "TDT loss kernel support and FP8 kernel integration changes affect different integrations and code paths." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 test fixes and TDT loss kernel support are unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs implement the same concrete fixed_cross_entropy kwargs support on the same function and same issue target, so they are mergeable duplicates." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is well grounded in the packet: the cluster is mostly unrelated, and the only accepted pair (43251/43254) clearly shares the same file, the same issue target, and the same fixed_cross_entropy kwargs fix. The rejection calls are conservative and do not overstate similarity." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 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": "34e86a7080c8f763ef7ba233fcabdc2f7096a6e924f96c67b5198b9408f08d92", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10380, + "estimated_input_tokens": 2595, + "estimated_eval_tokens": 5446 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10380, + "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 mixes one base VibeVoice implementation PR, one separate VibeVoice Realtime follow-up, and one unrelated Parakeet-RNNT PR. The shared auto-registry/docs files are not enough to make any pair a true duplicate or mergeable single PR.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:40546 is the original, broad VibeVoice implementation and is the clearest representative of the VibeVoice subset in this cluster.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:40546 is the most representative single PR here because it introduces the core VibeVoice model support without mixing in the separate realtime extension or an unrelated model family.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Both are VibeVoice-related, but 42781 is a separate realtime/acoustic-tokenizer extension rather than the same concrete implementation change; they are not the same underlying bug or mergeable into one PR." + }, + { + "left": "pull_request:42781", + "right": "pull_request:46331", + "accept": false, + "reason": "These target different model families (VibeVoice vs Parakeet-RNNT); overlap is only in generic auto-registration files, so they are unrelated." + } + ] + }, + "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": "68c5322b25f3c6cbc8be67595f188004fdb30fcf4aaa76d103fca27ef8de8ee1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4674, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4674, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated feature PRs in different model families: TimesFM covariate forecasting versus DINOv3 image classification support. They do not target the same bug or change set.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:41224 is the cleaner canonical representative because it is a focused model capability addition with clear code, auto-model, and test updates in one subsystem.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:41224 is the best standalone representative of the cluster because it is a concrete, self-contained feature PR with explicit model support and tests, whereas 40755 is a separate TimesFM enhancement.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Reject: these PRs touch different models and different code paths. TimesFM covariate forecasting and DINOv3 image classification support are unrelated feature additions, not a single mergeable change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 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": "2c5fe783972a3997226169de37cf28f7d2f074771c717c1f01f13aad98f93e40", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10436, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10436, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate new-model addition PRs (DeepSeek OCR, Isaac, Kimi2-6) that overlap mainly in Hugging Face boilerplate and auto-registration files, not in the same underlying change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No clear canonical duplicate PR: the PRs target different models and unrelated code paths, so any single PR would be only a loose representative.", + "best_issue_reason": null, + "best_pr_reason": "PR #45630 is the most mature and heavily reviewed of the set, so it is the strongest pragmatic representative if one must be chosen.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different model additions (DeepSeek OCR vs Isaac). Shared docs/auto-mapping scaffolding is boilerplate, not the same concrete bug or change." + }, + { + "left": "pull_request:45186", + "right": "pull_request:45630", + "accept": false, + "reason": "Different model additions (Isaac vs Kimi2-6) with only generic registration and docs overlap. They do not fix or implement the same underlying code path." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "e6b85d0c619464d82f549bff0f7f8b06dfbf7fea6705a1c0250cf2d1db649339", + "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 two PRs both update GitHub Actions in workflow files, but they target different action families and different upgrade goals: one is a Node 24 compatibility sweep for checkout/cache/upload-artifact, while the other bumps docker/build-push-action. They overlap in scope, not in the underlying change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "No single PR is a true canonical duplicate; if forced to pick a representative, 42975 is the broader workflow-action compatibility upgrade, but 42976 is a separate docker action bump.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Different dependency tracks and fix intents: Node 24-compatible GitHub Actions vs docker/build-push-action v6 upgrades. Shared workflow files alone is not enough to treat them as the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "dcccff41718c60ea2ca51e82b758c569e219cda15725dffe6ef50d54dfd233a3", + "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": "These are distinct model-addition PRs across different model families; the shared docs/auto-registry edits are broad integration boilerplate, not evidence of the same underlying change. No soft edges should be merged.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No true duplicate PR exists in this cluster; if one representative is needed, 43451 is the strongest candidate because it is the most extensive and has the highest review activity, but it is still a separate model addition.", + "best_issue_reason": null, + "best_pr_reason": "43451 is the best representative PR only in the sense of completeness and activity; it covers the broadest end-to-end model integration, but it is not equivalent to the other PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different models and codepaths: ViT NEPA vs Molmo. Shared docs/auto-registration files are generic model wiring, not the same change." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different model families: ViT NEPA vs Molmo2. The overlap is only in shared registry/docs updates." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Unrelated additions: ViT NEPA vision model vs CTSM time-series model. No shared bug or concrete implementation target." + }, + { + "left": "pull_request:42978", + "right": "pull_request:46411", + "accept": false, + "reason": "Completely different scope: ViT NEPA model integration vs a MiniMax demo/new model files. Not the same underlying change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are related model families, but they are distinct versions with different implementations and assets; they are not the same PR-worthy code-path fix/change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo vs CTSM are different models in different domains; shared auto/docs edits are not enough to treat them as duplicates." + }, + { + "left": "pull_request:43448", + "right": "pull_request:46411", + "accept": false, + "reason": "Molmo model integration is unrelated to the MiniMax demo/original-code PR." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are separate model integrations with different codepaths; overlap in registry/docs files is incidental." + }, + { + "left": "pull_request:43451", + "right": "pull_request:46411", + "accept": false, + "reason": "Molmo2 model integration is not the same as adding a MiniMax demo and model stub files." + }, + { + "left": "pull_request:45490", + "right": "pull_request:46411", + "accept": false, + "reason": "CTSM time-series model work is unrelated to the MiniMax demo/new model files; no shared underlying bug or mergeable change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "9e505281769031e3b41ae9103dfbe2b5479d9bd0ceea87c932717035255e3cc4", + "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 pull requests are not the same underlying change: they target different subsystems and different failure modes, with only superficial overlap in being small fixes/tests. No duplicate cluster emerges.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 46380 is the most substantial and concrete fix in the set, with a specific generation-config validation bug and corresponding code path update; however, it is not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Whisper tokenizer decoding/tests versus CLIP/ViT test edits; no shared bug or code path." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper tokenization change is unrelated to FP8 dequantization/core loading changes." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "Only commonality is Whisper appearing in filenames; one is tokenizer decode logic, the other is generation/watermarking plus broad test updates." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "Test-only CLIP/ViT changes do not match the FP8 dequant/model-loading bug fix." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "Both are generic-looking fixes, but they address different areas: image-model tests versus generation config and many model tests." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "Both touch model-loading-related code, but they fix different concrete problems 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": 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": "a9dc98cc914a95013da69fb2e61f84edc3e75b012aae4531fc04f04778339814", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4563, + "estimated_input_tokens": 1141, + "estimated_eval_tokens": 2538 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4563, + "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/versioning test change, while the other is a broader CI/workflow and testing-utilities change. The only overlap is a shared Vit test file, which appears incidental.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43532 is the better representative PR because it is the larger, more substantive change with broader discussion and multiple affected workflow/utility files; PR #43488 is a narrow bot-format/setup tweak.", + "best_issue_reason": null, + "best_pr_reason": "PR #43532 is the strongest standalone candidate in this cluster: it has the more meaningful code-path impact and broader change surface, whereas PR #43488 is a tiny setup/versioning test PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Reject: the diffs target unrelated concerns. #43488 changes setup/versioning and adds a pytest dependency/test tweak; #43532 rewires CI workflows and touches testing/notification utilities. Sharing one Vit test file is not enough to indicate the same underlying fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "edbfe99d2e579e1e25f5e5af3028aef50a7441d83e7a0c0f64389d9e0cb08299", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two open PRs add different pipelines: promptable concept segmentation and promptable visual segmentation. They share scaffolding and docs patterns, but they target different concrete features and model/test sets, so they should not be merged as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43613 is the better canonical representative because it has the broader, more integrated visual-segmentation implementation and test coverage; 43612 is a separate concept-segmentation feature, not a duplicate.", + "best_issue_reason": null, + "best_pr_reason": "PR 43613 is the strongest overall representative of the cluster due to its wider scope and more substantial code/test surface, while still being clearly distinct from 43612.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different features and code paths: concept segmentation vs visual segmentation. Shared documentation/pipeline scaffolding is not enough to treat them as the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "173fb11220d386ac58bff9d36b13ebb4f28debe13caef047983a77aa2d8c8c04", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4293, + "estimated_input_tokens": 1074, + "estimated_eval_tokens": 2404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated Trainer extension PRs in the same file: one changes gradient-accumulation loss scaling, the other refactors signature-column filtering for custom trainers/PEFT models. They are not duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs address different code paths in `trainer.py` and should not be collapsed into one duplicate group.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a clear duplicate of the other; if a representative is needed, `#43651` is slightly more self-contained and specific, but it does not subsume the other change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Different trainer behaviors: `#43651` adds `_loss_is_scaled_for_ga` for gradient-accumulation loss scaling, while `#43663` adds `_get_signature_columns` to control column filtering. Shared file alone is not enough to treat them as the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 12, + "component_count": 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": "0969a0749195deea83495164a4dc502b34fc8673fb2b9f4ff8d1c906cbbf7033", + "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": "The two PRs both touch `src/transformers/cli/serve.py`, but they fix different annotation/runtime-import problems: one is a broader postponed-annotation cleanup, the other specifically handles `TypeAdapter` when `pydantic` is unavailable. They are related but not the same change and should not be merged as duplicates.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "`pull_request:43842` is the better canonical PR because it has the clearest, most targeted fix and an explicit issue link; it addresses the concrete `TypeAdapter` import/annotation failure when `pydantic` is not installed.", + "best_issue_reason": null, + "best_pr_reason": "`pull_request:43842` is the strongest overall PR in this cluster: it is narrowly scoped, directly states the failure mode, and cleanly fixes the runtime path that would break CLI import without `pydantic`.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "They overlap in filename and general annotation theme, but they fix different bugs: `43656` is a broader future-annotations/type-hint cleanup, while `43842` specifically makes `TypeAdapter` safe when `pydantic` is absent. They are not the same underlying change and would not be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "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": "668598dbef95359f740f3c0aca511ea58cff14a546bd16a76688265692e5233b", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84898, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75094, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "A shared output-tracing migration cluster centered on issue #43979. Most PRs are model-specific refactors and should not be deduped across different models; only same-model retries look like true duplicates.", + "confidence": 0.86, + "canonical_issue_reason": "Issue #43979 is the shared tracker for the standardized output-tracing migration and is explicitly linked by many of these PRs.", + "canonical_pr_reason": "PR #44066 is a strong representative of the cluster: it is explicitly tied to #43979 and makes the concrete GPT-J/CodeGen output-tracing refactor on the relevant code path.", + "best_issue_reason": "Issue #43979 is the common upstream coordination point for this migration work.", + "best_pr_reason": "PR #44066 is the most representative, well-scoped implementation in the set and sits at the center of the duplicate GPT-J/CodeGen attempts.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models and code paths (CVT/FNet vs GPT-J); same theme only." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models/code paths; both reference #43979 but are not duplicates." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Same GPT-J output-tracing refactor on the same file; 44066 already covers the concrete GPT-J path." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "Different model families; only the refactor theme matches." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet vs GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "Different models and implementation paths; not the same bug/change." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 vs DeBERTa are unrelated model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models; shared migration style does not make them duplicates." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen vs MPT are separate code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both are GPT-Neo output-tracing refactors on the same file and same underlying code path." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo vs GPT-J/CodeGen are different model-specific refactors." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "Different models; same migration family only." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet vs GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR vs GPT-J/CodeGen are different model changes." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet vs GPT-J are different concrete code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are different models; not mergeable as one PR." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "Different models/code paths; only the output-tracing pattern is shared." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo vs Flaubert are separate model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "Different models and refactors; no shared concrete bug." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo vs GPT-J/CodeGen are different model-specific changes." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT vs GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo vs GPT-J are different model refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo vs MGP-STR are unrelated model changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "Different models; shared decorator migration theme is too broad." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT2 vs MGP-STR are unrelated code paths." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J vs MGP-STR are different model-specific changes." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT vs Flaubert are different model implementations." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT2 vs GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "Segformer vs GPT-J/CodeGen are different model refactors." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "Segformer vs LongT5 are unrelated concrete changes." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "Segformer vs VisualBert are different model paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo vs MPT are separate code paths." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44101", + "accept": false, + "reason": "TextNet vs Flaubert are unrelated model implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet vs LongT5 are different code paths." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet vs Segformer are different model refactors." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBERT vs MPT are unrelated models." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44074", + "accept": false, + "reason": "DeBERTa vs TextNet are different concrete changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J/CodeGen standardized output-tracing change on overlapping files; these are duplicate implementations." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Same GPT-J output-tracing refactor; 44722 is a direct duplicate-style follow-up on the same code path." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J vs SpeechT5 are different model-specific refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "Different model families; only the output-tracing migration theme overlaps." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet vs SpeechT5 are unrelated code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa vs SpeechT5 are different model changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa vs ViTS are different concrete refactors." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: issue #43979 is the shared tracker for this migration set, and the chosen canonical PR (#44066) is well supported by its explicit link and overlapping GPT-J/CodeGen scope. The soft-edge calls are conservative overall, especially by only accepting pairs that share the same concrete model/code path. " + }, + "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": "fc34f9b0c613b961f0848339e18ca2f704e9789f7c607c660cbdf75d5611cd5b", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84898, + "estimated_input_tokens": 21225, + "estimated_eval_tokens": 42706 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17125, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These PRs are all parallel model-specific refactors to standardized output tracing/capture_outputs, but they touch different architectures and code paths. They look thematically similar, not like duplicates or one mergeable change.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44073 is a clean representative of the cluster: it clearly expresses the shared standardized output-tracing refactor, and the diff directly shows the decorator/capture_outputs migration pattern.", + "best_issue_reason": null, + "best_pr_reason": "PR 44073 is the best global representative because its title and diff most clearly capture the common change pattern across the cluster without being unusually narrow or issue-specific.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are standardized-output refactors, but they target different models (FocalNet vs ImageGPT) and different implementation details. Same theme, not the same underlying bug/change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "Both migrate output tracing, but they are separate model implementations (VisualBert vs TextNet) with different forward signatures and test coverage. Not one concrete change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "These are model-specific output-tracing refactors for different architectures (MPNet vs VisualBert). They share a pattern, but not a mergeable single code-path fix." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "MPNet and TextNet are unrelated model code paths; the similarity is only the generic migration to capture_outputs. Not duplicates." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are different models with separate forward/output plumbing. The shared decorator migration is too broad to count as the same change." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 are different subsystems and the diffs indicate different output-tracing adaptations, not one concrete bug or feature merge." + } + ] + }, + "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": "8615dc41d7f316d63538a8d118b78531311e814ca41ecb0d0abc8af719b6a986", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39790, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39790, + "estimated_input_tokens": 9948, + "estimated_eval_tokens": 20152 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster contains a series of independent, model-specific output-tracing refactors. They share a common pattern (`can_return_tuple` / tracing helpers), but they target different model files and code paths, so they should not be treated as duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model families (`timm_backbone` vs `upernet`); same refactor theme only." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "`timm_backbone` and `vision_text_dual_encoder` are unrelated code paths despite similar tracing changes." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Separate models (`timm_backbone` vs `univnet`); not the same underlying bug or change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "`timm_backbone` and `mamba`/`falcon_mamba` are distinct implementations; no shared concrete fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "Different model code paths (`timm_backbone` vs `rwkv`); only the refactor style overlaps." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "`mobilenet_v1` and `univnet` are unrelated models; same output-tracing motif only." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Different model implementations (`mobilenet_v1` vs `upernet`), so not mergeable as one PR." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "`mobilenet_v1` and `depth_anything` touch different forward paths and files." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "Separate models (`mobilenet_v1` vs `vision_encoder_decoder`); no common concrete bug." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "`mobilenet_v1` and `speech_encoder_decoder` are unrelated model wrappers." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "`vision_text_dual_encoder` and `univnet` have different runtime behavior and code paths." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models (`vision_text_dual_encoder` vs `upernet`); only the tracing refactor is common." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "`vision_text_dual_encoder` and `mamba` are not the same underlying change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "`vision_text_dual_encoder` and `depth_anything` are separate model-specific refactors." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "`univnet` and `upernet` are unrelated models with independent forward implementations." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "`univnet` and `mamba`/`falcon_mamba` do not fix the same concrete issue." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Different models (`univnet` vs `depth_anything`); same refactor pattern only." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44026", + "accept": false, + "reason": "`univnet` and `vision_encoder_decoder` are distinct code paths." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "`upernet` and `rwkv` are unrelated model implementations." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "`upernet` and `mamba`/`falcon_mamba` do not share the same bug or patch." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "`mamba`/`falcon_mamba` vs `codegen`: different model subsystems and code paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "`timm_backbone` and `codegen` are unrelated models; similarity is only thematic." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "`univnet` and `codegen` target different forward logic and cannot be merged as one fix." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "`upernet` and `codegen` are separate model changes." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "`vision_text_dual_encoder` and `codegen` are different concrete code paths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "`depth_anything` and `vision_encoder_decoder` are different models; same refactor theme only." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "`depth_anything` and `speech_encoder_decoder` are independent wrappers with separate fixes." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "`depth_anything` and `superpoint` are unrelated model implementations." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "`depth_anything` and `rwkv` touch different model stacks and code paths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "`depth_anything` and `dpr` are distinct model-specific refactors." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "`vision_encoder_decoder` and `rwkv` are unrelated models." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are different wrappers, but not the same change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "`vision_encoder_decoder` and `superpoint` are separate code paths." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "`vision_encoder_decoder` and `dpr` fix different model implementations." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "`speech_encoder_decoder` and `superpoint` are unrelated model files." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "`speech_encoder_decoder` and `rwkv` do not share a concrete underlying bug." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "`speech_encoder_decoder` and `dpr` are distinct model-specific changes." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "`superpoint` and `rwkv` are different model paths; only the refactor pattern matches." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "`superpoint` and `dpr` are unrelated changes." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "`rwkv` and `dpr` are separate model implementations, 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": 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": "1318a490434d4a6d3478457ae12151b1dce8b8d077582f1e36871f513196779a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6551, + "estimated_input_tokens": 1638, + "estimated_eval_tokens": 3532 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs add different model integrations: PenguinVL vs Voxtral TTS. They share some boilerplate auto-registration/docs paths, but they do not fix the same underlying code-path or belong in one merged change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR as a duplicate representative: the PRs target different models and different feature sets, so neither subsumes the other.", + "best_issue_reason": null, + "best_pr_reason": "PR #45401 is the more self-contained candidate because it has an explicit issue target and a narrower, clearly scoped model-support addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Reject: different model implementations (PenguinVL vs Voxtral TTS). Shared files are generic registration/docs plumbing, not evidence of 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": 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": "6a9cdc14bb680094c56e09404250dc42fa71c9ca697a4aa20c821ce5320f6855", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 9634, + "estimated_input_tokens": 2409, + "estimated_eval_tokens": 5074 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 9634, + "estimated_input_tokens": 2409, + "estimated_eval_tokens": 5074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All four PRs are on the same kernel-fusion feature line, but they are sequential expansions of the API rather than the same concrete change. The latest PR is the broadest and most representative of the final direction.", + "confidence": 0.66, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46339 is the most complete and recent kernel-fusion implementation: it simplifies and extends the KernelConfig-based n-to-1 fusion path, touching the core loader, hub kernel integration, modeling utils, and tests.", + "best_issue_reason": null, + "best_pr_reason": "PR 46339 best represents the cluster because it subsumes the earlier KernelConfig fusion work and includes the widest implementation and test coverage.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Related feature area, but not the same change: 44979 introduces a lower-level Module Fusion API, while 45363 builds a separate KernelConfig-based n-to-1 fusion integration on top of it." + }, + { + "left": "pull_request:45363", + "right": "pull_request:46339", + "accept": false, + "reason": "Same general feature family, but 46339 is a later extension/simplification of the KernelConfig fusion path with additional loader and integration changes, so it is not the same concrete PR." + }, + { + "left": "pull_request:45666", + "right": "pull_request:46339", + "accept": false, + "reason": "Both work on n-to-1 kernel fusion, but 45666 is an intermediate extension and 46339 further refactors/simplifies the implementation; they are successive iterations, not a single duplicate change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "2a14e027e89c2fceb643be078cdceb96160df2bb98b769722811c0eae3c0c6bf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests target the same Mistral regex patch bug in the same file, with the same issue target and nearly identical code change; 45317 is the more complete version because it also adds tests.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the stronger canonical PR because it is the later, more complete fix for the same AttributeError path and includes test coverage in addition to the code change.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the best representative of the cluster because it fixes the concrete bug directly and adds tests, making it the more self-contained and merge-ready change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying AttributeError in `_patch_mistral_regex`, same affected file and issue target, and the diffs change the same code path in a mergeable way." + } + ] + }, + "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": "c314a3b8b7feac8e08f4ea514b2d1f77d5631ecfb8ec438ccfeb10df4794831b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6065, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6065, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two NemotronH-related pull requests, but they address different scopes: one is an inheritance/refactor toward GraniteMoeHybrid, the other is a broader model split introducing NemotronHDense/NemotronHSparse and docs. They are related, not duplicates.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45569 is the better canonical PR because it appears to deliver the fuller, user-facing model packaging change (new dense/sparse variants, auto-mapping, docs, and tests) rather than a narrow refactor.", + "best_issue_reason": null, + "best_pr_reason": "PR #45569 is the strongest representative of the cluster: it has the broadest and most complete implementation of the NemotronH workstream and is more likely to stand alone as the primary change. PR #45115 is a narrower preparatory refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "They share NemotronH files, but the changes are not the same underlying fix. #45115 is a targeted refactor/inheritance mapping toward GraniteMoeHybrid, while #45569 splits NemotronH into dense/sparse models with docs and tests. Different goals, not mergeable as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "4686420d3bcd26cfbc83bcdd77248f001ae3a8c2e34620aa0e20bc5bba3af36b", + "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": "The two PRs are related pieces of the same heterogeneous-model effort, but they are not duplicates: one adds per-layer config support, while the other builds out broader heterogeneous modeling/cache/masking support on top of that.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45332 is the broader umbrella change and is the better canonical representative of the heterogeneous-model feature work; PR 45333 looks like a narrower config-layer prerequisite rather than the same end-to-end change.", + "best_issue_reason": null, + "best_pr_reason": "PR 45332 is the best standalone PR because it covers the full code-path change set (config, modeling, cache, masking, tests) rather than only the configuration layer.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Related but not the same concrete change: 45333 is per-layer configuration support, while 45332 is a broader heterogeneous model implementation that depends on and extends that foundation. They would not be a single mergeable PR without collapsing distinct scopes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 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": "3e5a1c3392d12dabdb6584f5aec738ab240e9a0dd743ef0545bc4b5e5f74f111", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5514, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are typing-related PRs, but they are not the same change: one is a broad type-checking sweep across many Transformer files, while the other adds a smaller set of typing protocols/helpers around modeling_utils/PEFT integration. They overlap in theme and a few files, but not in concrete code-path or patch scope.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45415 is the broader umbrella change: it touches many Transformer modules plus pyproject typing settings, so it best represents the cluster\u2019s main typing-effort theme.", + "best_issue_reason": null, + "best_pr_reason": "PR 45415 is the strongest representative of the cluster because it is the most comprehensive and central typing cleanup among the two.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Same general typing topic, but different concrete changes and scopes: 45415 is a wide type-checking sweep, while 45425 adds typing helpers/protocols for specific integrations. They do not look mergeable as one PR for the same bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "f9f3d5987e143fd3a815c2f7adce6a32b1e0d83e9310d497500e32b09c388a65", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4858, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4858, + "estimated_input_tokens": 1215, + "estimated_eval_tokens": 2686 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are about different models and different cache bugs: PR #45513 fixes multi-token cached forward behavior in Qwen3.5/related linear-attention code, while PR #46084 fixes use_cache handling in Mamba2 when seq_len > 1. They are not the same underlying change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45513 is the broader, more representative cache-fix PR in this cluster, spanning multiple Qwen-family model files and a shared multi-token cached-forward bug pattern.", + "best_issue_reason": null, + "best_pr_reason": "PR #45513 is the best representative artifact because it addresses a concrete cache-forward bug across several related model implementations and is more central to the cluster than the unrelated Mamba2 fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45513", + "right": "pull_request:46084", + "accept": false, + "reason": "Different model families and different code paths: Qwen3.5 linear-attention cached forward vs. Mamba2 use_cache with seq_len > 1. They do not look 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": 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": "9955ab9939942577b3c46fb2364e3fbd50069f6a83de3792590986cf81e2e5f6", + "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 are related through expert-parallel loading, but they fix different bugs and touch different code paths. PR 45548 is about EP + DeepSpeed ZeRO-3 loading and config serialization; PR 45662 is about EP + FSDP2 experts being overwritten during rank-0 broadcast.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45662 is the more concrete and representative fix: it targets a specific expert-overwrite loading bug in the core parameter-loading path, with coordinated changes across loading and MoE integration. PR 45548 addresses a separate EP/ZeRO-3 accelerate-launch issue.", + "best_issue_reason": null, + "best_pr_reason": "PR 45662 is the best representative of the cluster because it fixes a clearly defined code-path bug with broader supporting changes, while PR 45548 is narrower and concerns a different loading/configuration problem.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "They share the EP theme, but the bugs are different: 45548 handles EP + DeepSpeed ZeRO-3 loading/config serialization, while 45662 fixes expert tensors being overwritten during FSDP2/rank-0 broadcast. 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": 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": "6d47cb1f0682f7a5a3a1267460846669b04a645cb109ef1ef7dd9a9089f431e3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4589, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same xdist-related captured_info handling in `testing_utils.py`, but #45645 is a broader follow-up that also changes CI log collection and notification handling. They are related, yet not close enough to treat as the same PR change.", + "confidence": 0.79, + "canonical_issue_reason": null, + "canonical_pr_reason": "#45645 is the more complete representative: it includes the core xdist-safe `captured_info` fix plus the CI-side log aggregation changes around that artifact.", + "best_issue_reason": null, + "best_pr_reason": "#45645 is the strongest overall fix in the cluster because it covers the main bug and the surrounding CI behavior needed to preserve and collect the split logs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": false, + "reason": "They overlap on the xdist-safe debug-log fix, but #45645 adds additional workflow and notification changes, so they are not the same underlying change and would not cleanly merge as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "f32ab67a40c856f8d2f1befbb646932e3acdda3502bedc6d1fe53d378ccd8173", + "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 are related only at a theme level (model support / attention dispatch), but they fix different models and different code paths, so they should not be merged as duplicates.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: the items are separate changes for different model families and different behaviors, not the same underlying fix.", + "best_issue_reason": null, + "best_pr_reason": "PR 46401 is the strongest standalone change: it adds SDPA dispatch support across BLIP-2/InstructBLIP/InstructBLIPVideo and includes tests, making it the most concrete and substantial PR in the set.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45670", + "right": "pull_request:46320", + "accept": false, + "reason": "These are unrelated changes: 45670 only adjusts AutoModel registration for glmasr, while 46320 adds attn_implementation support for RecurrentGemma." + }, + { + "left": "pull_request:46320", + "right": "pull_request:46401", + "accept": false, + "reason": "Both mention attn_implementation, but they target different model families (RecurrentGemma vs BLIP-2/InstructBLIP) and different code paths, so they are not the same underlying bug/change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "2a1f4dad7f082fa7bd6dfc2256eff4933a4699b3b583acc1f4780e77a2523c62", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7188, + "estimated_input_tokens": 1797, + "estimated_eval_tokens": 3850 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7188, + "estimated_input_tokens": 1797, + "estimated_eval_tokens": 3850 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs are not duplicates: #45720 implements TRF003-style forward/return_dict tuple handling across several model forwards, while #46259 adds a different linter rule (TRF018/TRF016) for image/video processor boolean flags and updates unrelated model processors. The overlap in some model filenames is incidental to broad repo-wide cleanup work.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR for this cluster: the PRs target different rules and different code paths, so they should not be merged as one change.", + "best_issue_reason": null, + "best_pr_reason": "#45720 is the better representative only in the sense that it is the more concrete code-path change; #46259 is mostly a separate linter-rule/docs rollout. Even so, they are not duplicates.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "Different underlying changes: #45720 patches model forward signatures/tuple-return behavior for TRF003, while #46259 introduces a separate processor-flag consumption rule and associated edits. Shared files are broad-model overlap, 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": 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": "6789b4734ef1f5eb556c9269da7ccbd818681d8ea14195de3bdc51dcd99bcf68", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4301, + "estimated_input_tokens": 1076, + "estimated_eval_tokens": 2408 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4301, + "estimated_input_tokens": 1076, + "estimated_eval_tokens": 2408 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two PRs are both non-production \u201cdo not merge\u201d changes, but they do not fix the same underlying problem: one is a mixed CI/model/config tweak PR, the other is a test-file migration that intentionally introduces failing assertions. I would not cluster them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45932 is the better representative of the cluster only in the loose sense that it contains the broader change set; PR #46100 is just a minimal test-only placeholder and does not share the same concrete code-path change.", + "best_issue_reason": null, + "best_pr_reason": "PR #45932 is the more substantive artifact and the closest thing to a cluster representative, but it still should not be treated as a duplicate of #46100.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Reject: the PRs touch different things and do not appear to fix the same bug or implement the same change. #45932 changes CI/workflow and several model/config files; #46100 only edits a ViT test file and adds obvious failing assertions. They are not plausibly mergeable into one PR as a single code-path 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": 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": "436d1e25452d526b165362ad851896c0b80e80561ac70dfbe36cfda7a78c8e64", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8081, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8081, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all SDPA-dispatch enablements, but they target different model families and code paths (CTRL, T5, wav2vec2/SeamlessM4T), so they are related feature work rather than duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:46073", + "right": "pull_request:46154", + "accept": false, + "reason": "Both add SDPA support, but one is for CTRL and the other for T5; they touch different model implementations and are not the same underlying fix." + }, + { + "left": "pull_request:46073", + "right": "pull_request:46196", + "accept": false, + "reason": "Shared SDPA theme only; CTRL is a separate architecture from wav2vec2/SeamlessM4T, so these are distinct model-specific changes." + }, + { + "left": "pull_request:46154", + "right": "pull_request:46196", + "accept": false, + "reason": "T5 SDPA dispatch and wav2vec2/SeamlessM4T SDPA dispatch affect different code paths and models, so they could not realistically 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": 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": "8f64c7ecbe2543b106c0906a48a6af0bf91c1a3a318653f5d406c0c6f80f3ac1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5572, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5572, + "estimated_input_tokens": 1393, + "estimated_eval_tokens": 3042 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two PRs are both in distributed-model territory, but they are not the same change: 46126 is a focused EP/router/trainer integration, while 46269 is a broad distributed branch-base PR spanning many subsystems. Shared filenames alone are not enough to treat them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "46126 is the more focused, self-contained change; 46269 is a broad integration/base branch PR with many unrelated distributed edits.", + "best_issue_reason": null, + "best_pr_reason": "46126 is the clearest representative artifact in this cluster because it describes a specific feature change rather than an umbrella branch-base sweep.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Different scope and likely different underlying changes: 46126 adds EP/router/trainer integration, while 46269 is a large distributed base branch touching many areas beyond that feature." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "ba1d5d1577ba511732fc1fa5c5b21dd354bed918c4051dfc7c8cfdd077b23b0f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related in theme (removing unsafe/source-file open() access), but they fix different code paths in different modules and should not be merged as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46152 is the broader, more central change: it adds a reusable resource-based source resolution path in import_utils, which addresses the underlying file-open pattern at the utility level.", + "best_issue_reason": null, + "best_pr_reason": "PR 46152 is the best representative for the cluster because it is the more general source-access refactor; PR 46207 is a separate, narrower caller-specific replacement in modeling_utils.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46152", + "right": "pull_request:46207", + "accept": false, + "reason": "Both avoid open() on source files, but they touch different modules and different behaviors: import_utils helper/source resolution vs modeling_utils attention-implementation detection. They do not appear to fix the same concrete bug or be mergeable into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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": "4d7adc4aad4138c855ae2e57f161c2ec1192d96093a529e17bfb311eca758327", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 8271, + "estimated_input_tokens": 2068, + "estimated_eval_tokens": 4392 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 4, + "serialized_chars": 8271, + "estimated_input_tokens": 2068, + "estimated_eval_tokens": 4392 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all model-test maintenance changes, but they address different underlying failures: one Cohere2MoE device-map test setup fix, two model-specific quant-cache skips, and one DeepSeek V4 XPU/distributed-worker backend adjustment. None of the soft pairs look like the same concrete bug or change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46362 is the clearest narrow fix: it adds a model-specific skip with an explicit incompatibility reason for DeepSeek V4 quant-cache testing.", + "best_issue_reason": null, + "best_pr_reason": "PR 46362 is the most representative single change in this set because it is a small, self-contained test adjustment with a concrete failure explanation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46337", + "right": "pull_request:46368", + "accept": false, + "reason": "Both are test-only model PRs, but they fix different models and different failures: Cohere2MoE device-map/offload setup vs NemotronH quant-cache incompatibility." + }, + { + "left": "pull_request:46337", + "right": "pull_request:46362", + "accept": false, + "reason": "Unrelated concrete issues: Cohere2MoE test configuration for hf_device_map vs DeepSeek V4 skipping an incompatible quant-cache test." + }, + { + "left": "pull_request:46362", + "right": "pull_request:46368", + "accept": false, + "reason": "Both skip generate_with_quant_cache, but they target different model implementations with separate cache incompatibility details; they are not the same underlying bug or a single mergeable change." + }, + { + "left": "pull_request:46362", + "right": "pull_request:46366", + "accept": false, + "reason": "DeepSeek V4 quant-cache test skip and DeepSeek V4 XPU/distributed-worker backend support affect different code paths and fix different failures, so they should not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 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-46376-3", + "nodes": [ + "pull_request:46376", + "pull_request:46417", + "pull_request:46425" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417", + "pull_request:46376|pull_request:46425" + ], + "prepared_review_unit_hash": "dd890f8eec3ef7fba556d07989f187b1690d7dc4d2839899646a1455034a7ceb", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 9319, + "estimated_input_tokens": 2330, + "estimated_eval_tokens": 4916 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 9319, + "estimated_input_tokens": 2330, + "estimated_eval_tokens": 4916 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The soft similarities are superficial: one PR adds multimodal continuous-batching runtime support, while the others add two different model families (HunYuan VL and RADIO). They are not the same underlying change or bug.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No clear canonical duplicate PR: the three PRs target distinct subsystems and model families, and the overlap is only broad 'add support' language.", + "best_issue_reason": null, + "best_pr_reason": "46376 is the broadest infrastructure PR, but it is still not a duplicate of the standalone HunYuan VL or RADIO model-addition PRs.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Continuous batching multimodal runtime support is unrelated to adding the HunYuan VL model package, docs, and tests; different code paths and deliverables." + }, + { + "left": "pull_request:46376", + "right": "pull_request:46425", + "accept": false, + "reason": "A multimodal continuous-batching runtime change is not the same as RADIO model integration; they touch different model families and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/manifest.json b/snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..35b927346ea7d20970a6509148f25f704996c076 --- /dev/null +++ b/snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260605t060128z", + "artifacts": { + "hybrid": "snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-06-05T06:10:47Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260605T060128Z", + "variant": "hybrid" +} diff --git a/snapshots/20260605T060128Z/manifest.json b/snapshots/20260605T060128Z/manifest.json index 7ebffd5b878a2797896255cddccf48dedbdd8fbf..059585b09738ff9009774fdc6422e934b1732092 100644 --- a/snapshots/20260605T060128Z/manifest.json +++ b/snapshots/20260605T060128Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 843 }, "extracted_at": "2026-06-05T06:01:28Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260605t060128z", + "runs": { + "hybrid-model-20260605t060128z": { + "analysis_id": "hybrid-model-20260605t060128z", + "artifacts": { + "hybrid": "snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260605T060128Z/analysis-runs/hybrid-model-20260605t060128z/manifest.json", + "model": null, + "published_at": "2026-06-05T06:10:47Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260605T060128Z", "watermark": {