diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 58e6fc48cf3bbb1e3c6fb433cca7d182a5b1e1e3..1781002beeae1eedc1c8058590f49a24c0386610 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": "20260605T120123Z", - "generated_at": "2026-06-05T12:11:20Z", + "snapshot_id": "20260605T180026Z", + "generated_at": "2026-06-05T18:11:20Z", "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.", @@ -1973,73 +1615,169 @@ ] }, { - "cluster_id": "cluster-43656-2", + "cluster_id": "cluster-45666-2", "summary": "Cluster of 2 related pull requests linked by soft_similarity.", "status": "open", "confidence": 0.5, "canonical_issue_number": null, - "canonical_pr_number": 43656, + "canonical_pr_number": 46339, "issue_numbers": [], "pr_numbers": [ - 43656, - 43842 + 45666, + 46339 ], "evidence_types": [ "soft_similarity" ], "pr_comparisons": [ { - "left_pr_number": 43656, - "right_pr_number": 43842, - "code_similarity": 0.405, - "size_similarity": 0.5, - "file_overlap": 1.0, - "area_overlap": 0.0, - "patch_similarity": 0.036, + "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/cli/serve.py" + "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/cli/serve.py", + "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, - 18 + 53 ], [ - 30, - 36 + 286, + 295 ], [ - 315, - 323 + 521, + 1077 + ] + ], + "right_ranges": [ + [ + 14, + 52 + ], + [ + 285, + 294 ], [ - 665, - 671 + 521, + 685 + ] + ] + }, + { + "filename": "src/transformers/modeling_utils.py", + "left_ranges": [ + [ + 1330, + 1338 ], [ - 931, - 937 + 3763, + 3789 ], [ - 1843, - 1849 + 4223, + 4234 ], [ - 1868, - 1874 + 4628, + 4641 ] ], "right_ranges": [ [ - 54, - 61 + 1346, + 1354 ], [ - 587, - 593 + 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 ] ] } @@ -2352,15 +2090,6 @@ "target_issue_number": 43240, "reason": "PRs in cluster-43240-3 are treated as duplicates because they converge on issue #43240 with closing_reference, shared_issue_target evidence." }, - { - "cluster_id": "cluster-43656-2", - "canonical_pr_number": 43656, - "duplicate_pr_numbers": [ - 43842 - ], - "target_issue_number": null, - "reason": "PRs in cluster-43656-2 are treated as duplicates because they share soft_similarity evidence." - }, { "cluster_id": "cluster-43698-3", "canonical_pr_number": 43779, @@ -2405,16 +2134,6 @@ "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, @@ -2432,6 +2151,15 @@ ], "target_issue_number": 45647, "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-45666-2", + "canonical_pr_number": 46339, + "duplicate_pr_numbers": [ + 45666 + ], + "target_issue_number": null, + "reason": "PRs in cluster-45666-2 are treated as duplicates because they share soft_similarity evidence." } ], "best_issue": { diff --git a/analysis/current/analysis-report-hybrid.llm-reviews.json b/analysis/current/analysis-report-hybrid.llm-reviews.json index 67defbffd398513ff36fca58a8cb9a23228bb3df..9f1bc3651e13c60841b4b990ecfe822e22026c8f 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": "20260605T120123Z", - "generated_at": "2026-06-05T12:11:20Z", + "snapshot_id": "20260605T180026Z", + "generated_at": "2026-06-05T18:11:20Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -133,340 +133,382 @@ "issue:41762|issue:43596", "issue:39290|issue:41720" ], - "prepared_review_unit_hash": "dc7f45d01bc684abea6c310e50525d68f7b321546428085ec445efc286cdf7a7", + "prepared_review_unit_hash": "e28d7c4fe4a737de77d8a55d006c8ddd05ec39aaf83dcdff5180fd4373b4b3af", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 64, - "serialized_chars": 46879, - "estimated_input_tokens": 11720, - "estimated_eval_tokens": 23696 + "serialized_chars": 46874, + "estimated_input_tokens": 11719, + "estimated_eval_tokens": 23694 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous; the only clear duplicate is the ZeRO-3 BERT index-out-of-bounds pair (43596/43638). The other soft links look like broad subsystem overlap or similar symptoms, not the same concrete bug.", - "confidence": 0.58, - "canonical_issue_reason": "43596 is the best issue representative of the small ZeRO-3 index-error family: it states the shared failure cleanly, and 43638 appears to be the same bug on a non-pretrained BERT variant.", + "summary": "This cluster is mostly a grab-bag of unrelated transformer/model/runtime issues. The only clear duplicate-like pair is the two DeepSpeed ZeRO-3 Bert IndexError reports; the rest should stay separate. No PRs are present.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43596 is the cleaner anchor for the shared ZeRO-3/Bert size-0 IndexError, and 43638 reads like a narrower variant of the same failure.", "canonical_pr_reason": null, - "best_issue_reason": "43596 is the strongest issue candidate overall because it is the clearest, most general statement of the duplicated ZeRO-3 'index 0 is out of bounds' problem.", + "best_issue_reason": "Issue 43596 is the best representative issue in this set because it captures the recurring ZeRO-3/Bert loading failure in a general form.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Different failures in different model paths: Qwen3.5 config attribute error vs streaming chat/completions failure." + "reason": "Different failures: config attribute missing vs streaming chat/completions error." }, { "left": "issue:45237", "right": "issue:45685", "accept": false, - "reason": "Both mention model/runtime problems, but one is AMD GPU support for GPT-OSS and the other is an MPS histogram op limitation; not the same bug." + "reason": "Different backends and symptoms: AMD GPU failure vs MPS int histogram limitation." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different model families and symptoms: dtype mismatch in Mistral-3 vs GPT-OSS failing on AMD GPUs." + "reason": "Different bug classes: image dtype mismatch vs AMD GPU runtime failure." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Autocast dtype mismatch in Phi-tiny-MoE is not the same issue as GPT-OSS running on AMD GPUs." + "reason": "Different code paths: autocast dtype mismatch vs AMD GPU model failure." }, { "left": "issue:45237", "right": "issue:45507", "accept": false, - "reason": "Separate GPT-OSS-related reports, but one is an AMD GPU failure and the other is an invalid method call in GraniteMoEHybrid." + "reason": "Unrelated model/backend issues; no shared underlying defect." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error is unrelated to GPT-OSS AMD GPU support." + "reason": "Different model families and errors; not the same bug." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "Both are DeepSpeed/model runtime issues, but collective mismatch under ZeRO-3 is not the same as ModernBERT ONNX export failure." + "reason": "Deepspeed collective mismatch is unrelated to ONNX export errors." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "Different models and code paths: ONNX export for ModernBERT vs Qwen3Moe with FineGrainedFP8." + "reason": "ONNX export issue vs Qwen3Moe FP8 failure are different problems." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "ONNX export error and Qwen3 auto device-map cuda assert are distinct failures." + "reason": "Export error and CUDA assert on auto device map are unrelated." }, { "left": "issue:45901", "right": "issue:46015", "accept": false, - "reason": "Both are documentation or task-related, but one is a table QA crash and the other is outdated RAG examples." + "reason": "One is a task crash, the other is outdated RAG docs; not the same issue." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Related v5 docs cleanup, but they target different removed content: pipeline tasks vs the transformers run command." + "reason": "Both are docs problems, but they concern different removed references and would be fixed separately." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ONNX export issue is unrelated to the GenerationMixin import regression." + "reason": "ONNX export error and GenerationMixin import error are different failures." }, { "left": "issue:44918", "right": "issue:46165", "accept": false, - "reason": "TRL SFT unpacking issue and resume-training/trust-remote-code issue are different training flows and failure modes." + "reason": "TRL SFT embedding handling is unrelated to resuming training with trust_remote_code." }, { "left": "issue:44805", "right": "issue:45823", "accept": false, - "reason": "Different models and likely different root causes: a mask/index shape error vs Gemma4 PLE device mismatch under auto mapping." + "reason": "Mask/index shape mismatch is not the same as Gemma4 PLE device placement." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both mention index/shape errors, but 43638 is a ZeRO-3 BERT load/train failure while 44805 is a mask-shape mismatch; not the same concrete bug." + "reason": "Both involve indexing errors, but the concrete failure modes and code paths differ." }, { "left": "issue:43638", "right": "issue:46165", "accept": false, - "reason": "ZeRO-3 BERT index error is unrelated to resume-training with locally downloaded models and remote code." + "reason": "ZeRO-3 Bert loading failure is unrelated to training resume/trust_remote_code." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "Different failure classes: ZeRO-3 index-out-of-bounds during BERT loading vs a mask/tensor shape mismatch." + "reason": "Similar IndexError wording, but the underlying tensor-shape issue is different." }, { "left": "issue:44805", "right": "issue:46294", "accept": false, - "reason": "Mask/index shape mismatch is unrelated to Qwen3.5 video feature/token mismatch." + "reason": "Different multimodal/video input bugs with distinct symptoms." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Generic shape mismatch is not the same as GPT-OSS MoE tensor-parallel malfunction." + "reason": "Different tensor shape/indexing bug vs GPT-OSS tensor-parallel issue." }, { "left": "issue:44560", "right": "issue:46165", "accept": false, - "reason": "Video StopIteration in Qwen3-vl-embedding is unrelated to resuming training with trust_remote_code." + "reason": "Video chat-template StopIteration is unrelated to resume-training behavior." }, { "left": "issue:43596", "right": "issue:46165", "accept": false, - "reason": "ZeRO-3 BERT index-out-of-bounds is unrelated to the resume-training/trust-remote-code issue." + "reason": "Different failure modes and code paths." }, { "left": "issue:43572", "right": "issue:46165", "accept": false, - "reason": "StableLmConfig missing pad_token_idx after v5 update is a different config regression." + "reason": "Config field regression is unrelated to training resume behavior." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice_preset bug and ModernBERT ONNX export error affect unrelated subsystems." + "reason": "BarkProcessor preset handling and ONNX export are unrelated." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Both involve DeepSpeed, but collective mismatch in Jamba-v01 is not the same bug as the ZeRO-3 BERT index error." + "reason": "Collective mismatch and ZeRO-3 Bert IndexError are different bugs." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Same underlying ZeRO-3 failure: both report the 'index 0 is out of bounds' error during BERT model load/train, with only minor wording/model-variant differences." + "reason": "Same ZeRO-3/Bert size-0 IndexError family; 43638 is a narrower variant of the same underlying bug." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger Kernel cross-entropy calling and GPT-OSS MoE tensor-parallel failure are different problems." + "reason": "Liger cross_entropy being called is unrelated to GPT-OSS tensor parallelism." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Incorrect detokenization and LoRA merge/save/load output mismatch are unrelated." + "reason": "Detokenization mismatch is unrelated to LoRA merge/save/load inconsistency." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error is unrelated to LoRA merge/save/reload behavior." + "reason": "Qwen-3-8B NVFP4 runtime failure is unrelated to LoRA save/reload output divergence." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Both mention tensor-shape issues, but FSDP activation checkpointing recompute mismatch is not the same as a mask/index shape error." + "reason": "Activation checkpointing/FSDP recomputation mismatch is a different bug from mask indexing." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "DeepSpeed training with a 2-D weight error is a different failure mode from collective mismatch on ranks." + "reason": "Different DeepSpeed failures: weight dimensionality vs collective mismatch." }, { "left": "issue:45859", "right": "issue:45902", "accept": false, - "reason": "Both involve Qwen3_5MoeTextRotaryEmbedding, but one is CPU-offload incompatibility and the other is uninitialized-memory access after materialization." + "reason": "Same module name, but CPU offload incompatibility and uninitialized-memory-after-to_empty are different bugs." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "ZeRO-3 BERT index error and GPT-OSS tensor-parallel failure do not share the same code path." + "reason": "Zero-3 Bert loading failure is unrelated to GPT-OSS tensor-parallel behavior." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Activation checkpointing/FSDP recompute mismatch is unrelated to the custom trainer compute_loss signature regression." + "reason": "FSDP checkpointing mismatch and custom trainer signature regression are unrelated." }, { "left": "issue:43596", "right": "issue:45161", "accept": false, - "reason": "ZeRO-3 BERT index error is not the same as GPT-OSS MoE tensor-parallel failure." + "reason": "Different model systems and failure modes." }, { "left": "issue:43716", "right": "issue:45507", "accept": false, - "reason": "Different model families and symptoms: Mistral-3 dtype mismatch vs GraniteMoEHybrid invalid method call." + "reason": "Image dtype mismatch and GraniteMoE invalid method call are unrelated." }, { "left": "issue:45874", "right": "issue:46294", "accept": false, - "reason": "Gemma4 reverse-embedding lookup is unrelated to Qwen3.5 video token/feature mismatch." + "reason": "Gemma4 inputs_embeds reverse lookup is unrelated to Qwen3.5 video token mismatch." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3TextConfig sliding_window_pattern regression is unrelated to the Mistral tokenizer regex patch crash." + "reason": "Gemma3/vLLM attribute regression is unrelated to Mistral regex patch failure." }, { "left": "issue:45823", "right": "issue:45874", "accept": false, - "reason": "Gemma4 device mismatch under auto mapping and Gemma4 reverse-embedding lookup are different forward-path issues." + "reason": "Different Gemma4 issues: device placement vs expensive reverse embedding lookup." }, { "left": "issue:43425", "right": "issue:44292", "accept": false, - "reason": "Torch 2.10 incompatibility and Qwen-3-8B-NVFP4 runtime error are distinct version/runtime problems." + "reason": "Torch 2.10 incompatibility and Qwen-3-8B NVFP4 runtime failure are different problems." }, { "left": "issue:34928", "right": "issue:41720", "accept": false, - "reason": "FSDP checkpointing recompute mismatch is not the same as Qwen3 auto device mapping cuda assert." + "reason": "Activation checkpointing/FSDP mismatch is unrelated to Qwen3 auto-device-map CUDA assert." }, { "left": "issue:43828", "right": "issue:45161", "accept": false, - "reason": "Phi-tiny-MoE autocast dtype mismatch and GPT-OSS only-TP failure are unrelated." + "reason": "Autocast dtype mismatch is unrelated to GPT-OSS tensor parallel failure." }, { "left": "issue:30277", "right": "issue:36331", "accept": false, - "reason": "DeepSpeed collective mismatch and CustomTrainer.compute_loss signature breakage are different bugs." + "reason": "DeepSpeed collective mismatch and trainer signature mismatch are different bugs." }, { "left": "issue:34928", "right": "issue:36010", "accept": false, - "reason": "Activation checkpointing/FSDP issue is unrelated to the GenerationMixin import error." + "reason": "FSDP recomputation mismatch is unrelated to GenerationMixin import failure." }, { "left": "issue:43039", "right": "issue:44805", "accept": false, - "reason": "Liger Kernel cross_entropy being called and a tensor mask/index shape error are not the same bug." + "reason": "Liger cross_entropy dispatch and tensor mask shape mismatch are different issues." }, { "left": "issue:45237", "right": "issue:46015", "accept": false, - "reason": "GPT-OSS AMD GPU failure is unrelated to outdated RAG documentation." + "reason": "AMD GPU runtime failure is unrelated to outdated RAG examples." }, { "left": "issue:43866", "right": "issue:46402", "accept": false, - "reason": "Corrupted Ovis2 checkpoint and Sam3Model text encoder weights not loading are different model/data problems." + "reason": "Checkpoint corruption and missing Sam3Model weight loading are not the same bug." }, { "left": "issue:45800", "right": "issue:46304", "accept": false, - "reason": "Torch/transformers version incompatibility and USE_DISTRIBUTED=0 breakage are separate compatibility regressions." + "reason": "Torch/version incompatibility and USE_DISTRIBUTED=0 regression are separate compatibility issues." }, { "left": "issue:43296", "right": "issue:46304", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and USE_DISTRIBUTED=0 support are unrelated breakages." + "reason": "PaddleOCR-VL/vLLM loading failure is unrelated to the distributed-build regression." }, { "left": "issue:41762", "right": "issue:43596", "accept": false, - "reason": "Both are ZeRO-3 index errors, but they occur on different models and appear to be separate manifestations rather than one concrete duplicate." + "reason": "Both are ZeRO-3 IndexErrors, but on different models and code paths; too weak to merge." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues, but they concern different removed guidance and are separate edits." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Different multimodal preprocessing failures." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Different dtype/autocast issue vs MPS histogram integer limitation." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "TRl SFT embedding unpacking and MPS histogram failure are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different tensor kernel/autocast failures with no shared code path." + }, + { + "left": "issue:30277", + "right": "issue:36010", + "accept": false, + "reason": "Collective mismatch and import failure are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP bug is unrelated to import regression." }, { "left": "issue:39290", "right": "issue:41720", "accept": false, - "reason": "Gemma3TextConfig sliding-window regression is unrelated to Qwen3 auto device mapping cuda asserts." + "reason": "Gemma3/vLLM attribute issue is unrelated to Qwen3 CUDA assert on device mapping." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The ZeRO-3/BERT index-out-of-bounds pair (43596/43638) is a plausible duplicate, while the other soft-edge rejections are cautious and supported by the issue titles. I don\u2019t see clear evidence of overstatement." + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft edge (43596/43638) is well supported by the shared ZeRO-3/Bert size-0 IndexError wording, and the rest of the rejections do not overstate similarity. No PRs are present, so there is no mergeability concern to flag." }, "evaluator_used": true, "retried": false, @@ -539,8 +581,8 @@ "soft_pairs": [ "issue:44589|issue:46304", "issue:45823|issue:46294", - "issue:43901|issue:44512", "issue:45237|issue:45901", + "issue:43901|issue:44512", "issue:44841|issue:45823", "issue:43039|issue:44560", "issue:43975|issue:46015", @@ -556,15 +598,15 @@ "issue:43854|issue:43866", "issue:43425|issue:45070", "issue:43596|issue:45800", - "issue:43975|issue:45901", "issue:43792|issue:45070", + "issue:43975|issue:45901", "issue:45362|issue:45464", "issue:44292|issue:45901", "issue:39290|issue:45823", "issue:45464|issue:45685", "issue:43541|issue:45685", - "issue:43404|issue:43866", "issue:43541|issue:43828", + "issue:43404|issue:43866", "issue:44841|issue:45084", "issue:36010|issue:39290", "issue:45800|issue:45823", @@ -579,20 +621,20 @@ "issue:43366|issue:45464", "issue:43531|issue:45070" ], - "prepared_review_unit_hash": "5df8b8afa4dadcf766d94d8d54f38f89e8db7d02d186cbf679a1a17f70432e77", + "prepared_review_unit_hash": "93a1fc7acbcd7a2bd6e38f494def14a99b72bc19c39b969fbf0d9b5b588c439c", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 41, - "serialized_chars": 39794, + "serialized_chars": 39795, "estimated_input_tokens": 9949, "estimated_eval_tokens": 20154 }, @@ -601,264 +643,264 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous; nearly all soft pairs are different bugs or docs regressions. The only clear duplicate-like pair is the v5 AutoTokenizer wrong-token-IDs regression across model families (45812 and 45920).", - "confidence": 0.89, - "canonical_issue_reason": "If one issue must serve as the cluster anchor, 45920 is the broadest representative: it describes the v5 AutoTokenizer token-ID regression across multiple model families and has inbound references.", + "summary": "This cluster is mostly a mix of unrelated transformer regressions, but one duplicate family stands out: the v5 AutoTokenizer wrong-token-id regression across multiple model families. I only accept that soft duplicate; the rest are too model- or symptom-specific to merge safely.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45920 is the broadest, clearest formulation of the shared AutoTokenizer token-ID regression and covers multiple impacted model families, making it the best representative of the only strong duplicate family here.", "canonical_pr_reason": null, - "best_issue_reason": "45920 is the strongest global representative because it captures the broadest, most cross-cutting user-visible regression in the set and is more general than the Granite-only report.", + "best_issue_reason": "issue:45920 is the best global issue representative because it generalizes the tokenizer regression beyond a single model and has the strongest evidence of being the shared underlying bug.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44589", "right": "issue:46304", "accept": false, - "reason": "Different failures: a missing Float8 storage type versus a broader torch/distributed compatibility break." + "reason": "Different failures: a Float8 storage lookup error vs a distributed-training compatibility issue." }, { "left": "issue:45823", "right": "issue:46294", "accept": false, - "reason": "Both involve model loading/inference, but they are different models and different code paths (Gemma4 device mismatch vs Qwen3.5 video chat template handling)." + "reason": "Both are model-loading/runtime bugs, but they involve different models and different code paths (device mismatch vs chat-template/video-token mismatch)." }, { - "left": "issue:43901", - "right": "issue:44512", + "left": "issue:45237", + "right": "issue:45901", "accept": false, - "reason": "Both are docs-related, but they cover different APIs and different stale references; not the same underlying change." + "reason": "Completely different symptoms and subsystems: AMD GPU support vs table-question-answering crash." }, { - "left": "issue:45237", - "right": "issue:45901", + "left": "issue:43901", + "right": "issue:44512", "accept": false, - "reason": "Unrelated problems: AMD GPU support for GPT-OSS versus a table-question-answering task crash." + "reason": "Both are docs issues, but they reference different API changes and do not describe the same underlying change." }, { "left": "issue:44841", "right": "issue:45823", "accept": false, - "reason": "Different model families and symptoms; processor failure is not the same as a PLE device-mismatch bug." + "reason": "Different models and different runtime failures; no evidence they share the same bug." }, { "left": "issue:43039", "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross-entropy routing and Qwen3-vl video StopIteration are unrelated." + "reason": "Unrelated features: Liger Kernel cross-entropy dispatch vs Qwen3-VL video input StopIteration." }, { "left": "issue:43975", "right": "issue:46015", "accept": false, - "reason": "Tokenizer detokenization bug and outdated RAG examples do not share the same code path." + "reason": "A decoding regression is unrelated to outdated RAG documentation." }, { "left": "issue:45812", "right": "issue:45920", "accept": true, - "reason": "Both report the same v5 AutoTokenizer regression producing wrong token IDs across multiple model families; they are plausibly one fixable tokenizer-mapping bug." + "reason": "Same underlying AutoTokenizer v5 regression: wrong token IDs across model families, with 45920 just describing a broader set of affected models." }, { "left": "issue:44863", "right": "issue:46015", "accept": false, - "reason": "NemotronH checkpoint loading and outdated RAG docs are unrelated." + "reason": "Checkpoint loading failure vs documentation staleness; no shared code-path." }, { "left": "issue:45020", "right": "issue:46304", "accept": false, - "reason": "Both mention recent-version breakage, but one is remote_code model loading and the other is USE_DISTRIBUTED=0 PyTorch support; different issues." + "reason": "Broad remote_code breakage is not the same as the USE_DISTRIBUTED=0 PyTorch compatibility issue." }, { "left": "issue:43866", "right": "issue:44863", "accept": false, - "reason": "Corrupted Ovis2 checkpoint versus NemotronH implementation loading failure are different underlying problems." + "reason": "Both mention loading problems, but the model families and failure modes differ." }, { "left": "issue:43872", "right": "issue:46304", "accept": false, - "reason": "Bitsandbytes _is_hf_initialized incompatibility is unrelated to the PyTorch USE_DISTRIBUTED=0 breakage." + "reason": "bitsandbytes constructor incompatibility is unrelated to distributed PyTorch support." }, { "left": "issue:44292", "right": "issue:46015", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error and outdated RAG examples are unrelated." + "reason": "Model execution/runtime error vs docs problem; unrelated." }, { "left": "issue:39290", "right": "issue:43531", "accept": false, - "reason": "Both mention sliding-window concepts, but one is a missing config attribute with vLLM and the other is a Qwen3-MoE sliding_window issue; not the same bug." + "reason": "Both mention sliding_window, but they are different models and different missing/incorrect config behavior." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Both are IndexErrors with DeepSpeed/BERT context, but the concrete failures differ (mask shape mismatch vs out-of-bounds on size 0)." + "reason": "Both are IndexErrors under ZeRO-3, but the symptoms and model contexts are different." }, { "left": "issue:43792", "right": "issue:45901", "accept": false, - "reason": "Whisper model loading failure and table-question-answering task crash are unrelated." + "reason": "Whisper model loading/running issue is unrelated to table-question-answering crashes." }, { "left": "issue:43854", "right": "issue:43866", "accept": false, - "reason": "GLM-4.7 unit-test loading issue and Ovis2 checkpoint corruption are different model-specific failures." + "reason": "Different models and distinct load failures; not the same bug." }, { "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Torch version incompatibility and a pydantic PretrainedConfig regression are not the same bug." + "reason": "Torch version incompatibility and a pydantic PretrainedConfig regression are unrelated." }, { "left": "issue:43596", "right": "issue:45800", "accept": false, - "reason": "Both involve incompatibility reports, but one is DeepSpeed zero3/BERT and the other is torch 2.4.1 with transformers 5.8.0; different code paths." + "reason": "Zero-3/BERT IndexError is a different failure from the torch 2.4.1 / transformers 5.8.0 incompatibility report." }, { - "left": "issue:43975", - "right": "issue:45901", + "left": "issue:43792", + "right": "issue:45070", "accept": false, - "reason": "Detokenization regression and table-question-answering crash are unrelated." + "reason": "Whisper runtime failure and pydantic model breakage are unrelated." }, { - "left": "issue:43792", - "right": "issue:45070", + "left": "issue:43975", + "right": "issue:45901", "accept": false, - "reason": "Whisper run failure and pydantic field breakage are unrelated." + "reason": "Detokenization regression is unrelated to table-question-answering crashes." }, { "left": "issue:45362", "right": "issue:45464", "accept": false, - "reason": "Both are Qwen3.5-related, but one is a chat crash and the other is a streaming chat/completions API failure; too different to merge." + "reason": "Both involve Qwen3.5, but one is chat crashes and the other is streaming API failure; too different to merge." }, { "left": "issue:44292", "right": "issue:45901", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error and table QA crash are unrelated." + "reason": "Unrelated model load/run bug vs table QA task crash." }, { "left": "issue:39290", "right": "issue:45823", "accept": false, - "reason": "Different models and symptoms; missing sliding_window_pattern is not a device-map forward mismatch." + "reason": "Different models and different symptom classes; no shared bug evidenced." }, { "left": "issue:45464", "right": "issue:45685", "accept": false, - "reason": "Qwen3.5 streaming API failure and MPS histogram Int error are unrelated." + "reason": "Streaming inference failure and MPS histogram kernel limitation are unrelated." }, { "left": "issue:43541", "right": "issue:45685", "accept": false, - "reason": "Grouped_mm tracing failure and MPS histogram Int error are different issues." + "reason": "Different backend-specific MoE failures; not the same concrete bug." }, { - "left": "issue:43404", - "right": "issue:43866", + "left": "issue:43541", + "right": "issue:43828", "accept": false, - "reason": "Tied lm_head weights in Mistral3 and corrupted Ovis2 checkpoint are not the same problem." + "reason": "Both are MoE-related, but one is torch-dynamo/grouped_mm tracing and the other is autocast dtype mismatch." }, { - "left": "issue:43541", - "right": "issue:43828", + "left": "issue:43404", + "right": "issue:43866", "accept": false, - "reason": "Both involve runtime/tracing/dtype problems, but the concrete failures and code paths differ." + "reason": "Tied-weight bug in Mistral3 is unrelated to checkpoint corruption/loading." }, { "left": "issue:44841", "right": "issue:45084", "accept": false, - "reason": "Processor failure for Voxtral and non-template-node compilation are unrelated." + "reason": "Processor failure and template compilation error are different components and failures." }, { "left": "issue:36010", "right": "issue:39290", "accept": false, - "reason": "ImportError for GenerationMixin and Gemma3 sliding_window_pattern are unrelated API breakages." + "reason": "ImportError for GenerationMixin is unrelated to Gemma3 sliding_window config errors." }, { "left": "issue:45800", "right": "issue:45823", "accept": false, - "reason": "Torch/transformers compatibility and Gemma4 device mismatch are unrelated." + "reason": "Torch/transformers version incompatibility is not the same as the Gemma4 device-map forward mismatch." }, { "left": "issue:41093", "right": "issue:41762", "accept": false, - "reason": "Both are IndexErrors, but one is a mask/tensor length mismatch and the other is zero-sized tensor access during Gemma3 ZeRO-3 loading." + "reason": "Both are IndexErrors, but the model context and exact failure differ; not enough to treat as the same bug." }, { "left": "issue:43296", "right": "issue:43366", "accept": false, - "reason": "Different model-support requests: PaddleOCR-VL loading versus GGUF gpt-oss architecture support." + "reason": "PaddleOCR-VL loading failure and GGUF gpt-oss support are different model-specific issues." }, { "left": "issue:43572", "right": "issue:45070", "accept": false, - "reason": "StableLm pad_token_idx regression and PretrainedConfig pydantic field breakage are unrelated." + "reason": "Missing pad_token_idx after v5 update is unrelated to the pydantic PretrainedConfig regression." }, { "left": "issue:44805", "right": "issue:44918", "accept": false, - "reason": "Both are shape-related training errors, but they involve different models and different failure modes." + "reason": "Similar shape-related wording, but the underlying failures are different (mask indexing vs embedding unpacking with TRL)." }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Int4 OOM regression and tied-weights translation-model issues are different bugs." + "reason": "Int4 OOM / CUDA memory growth is unrelated to tied-weights issues in translation models." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Torch dynamo grouped_mm tracing error and GPT-OSS only-TP issue are unrelated." + "reason": "MoE grouped_mm tracing failure is not the same as GPT-OSS tensor-parallel support problems." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch and Qwen3.5 streaming API failure are unrelated." + "reason": "Autocast dtype mismatch and streaming chat/completions failure are different bugs." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "_is_hf_initialized TypeError during init_empty_weights and the pydantic PretrainedConfig regression are different incompatibilities." + "reason": "_is_hf_initialized constructor error is unrelated to the PretrainedConfig pydantic regression." }, { "left": "issue:43366", "right": "issue:45464", "accept": false, - "reason": "GGUF gpt-oss architecture support and Qwen3.5 streaming inference failure are unrelated." + "reason": "GGUF architecture support and Qwen3.5 streaming failure are different issues." }, { "left": "issue:43531", "right": "issue:45070", "accept": false, - "reason": "Qwen3-MoE sliding_window handling and pydantic PretrainedConfig field breakage are different issues." + "reason": "Sliding-window config behavior and pydantic PretrainedConfig breakage are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The evaluation is conservative and well grounded in the packet. The only accepted soft edge, 45812\u219445920, is supported by the titles and the broader v5 AutoTokenizer regression theme. The other pairs are reasonably rejected as different bugs or different code paths, and the cluster-anchor choice of 45920 is defensible because it is the broader cross-model representative." + "feedback": "The summary is grounded in the packet and stays conservative. The accepted duplicate pair, issue:45812 vs issue:45920, is well supported by the shared AutoTokenizer v5 wrong-token-ID regression. The remaining soft-edge rejections are appropriately cautious and do not overstate shared causality." }, "evaluator_used": true, "retried": false, @@ -962,241 +1004,241 @@ "issue:43866|issue:44403", "issue:38175|issue:43054" ], - "prepared_review_unit_hash": "101ad5c93328545a1bd1aa39d033d4d3c69c8b8ed7affdde3e34cf41f07b4a12", + "prepared_review_unit_hash": "a67f443e02065387555e888c91616393f8cae27e56125b798a5bf440c02c3d22", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 33, - "serialized_chars": 36897, - "estimated_input_tokens": 9225, - "estimated_eval_tokens": 18706 + "serialized_chars": 36891, + "estimated_input_tokens": 9223, + "estimated_eval_tokens": 18702 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose grab-bag of unrelated Transformers issues. The only clear duplicate chain is the non-persistent-buffer regression in v5; most other soft pairs share a broad subsystem or model family but not the same concrete bug.", + "summary": "The cluster is best anchored on issue 43950, with 43644 and 44534 as the strongest near-duplicates around the same Transformers 5.x non-persistent-buffer corruption problem. The rest of the cluster is mostly adjacent but unrelated reports, so the grouping should stay fairly tentative overall.", "confidence": 0.74, - "canonical_issue_reason": "No single canonical issue fits the whole set well; as the best local anchor, issue 43950 is the clearest representative of the repeated non-persistent-buffer corruption regression and directly absorbs the neighboring duplicates.", + "canonical_issue_reason": "Issue 43950 is the cleanest anchor for the shared non-persistent-buffer corruption topic: its title directly names the regression and it is the most specific representative of that bug family among the related reports.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43950 is the strongest issue to keep as the representative duplicate target because it is specific, regression-focused, and has explicit near-duplicates in this cluster (43644 and 44534).", + "best_issue_reason": "Issue 43950 is the best single issue to represent the cluster because it most directly captures the common bug theme shared by the strongest nearby duplicates.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43716", "right": "issue:46294", "accept": false, - "reason": "Both are multimodal input problems, but one is a dtype mismatch in Mistral-3 preprocessing and the other is a Qwen3.5 video token/feature mismatch; different model families and failure modes." + "reason": "Both are multimodal/video-related, but they describe different failures: image-preprocessor dtype mismatch versus chat-template video token/feature mismatch." }, { "left": "issue:45020", "right": "issue:45800", "accept": false, - "reason": "One is a broad remote_code breakage report, the other is a specific torch/transformers version incompatibility; not the same concrete bug." + "reason": "These are different bug reports: a broad remote_code regression versus a specific torch/transformers version incompatibility." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Different runtime errors and code paths: grouped_mm tracing failure vs missing Float8 storage object." + "reason": "Different code paths and failure modes: grouped_mm tracing errors versus missing Float8 storage." }, { "left": "issue:44479", "right": "issue:46294", "accept": false, - "reason": "Both involve video input regressions, but the titles point to different bugs: a v5.3.0 model-family regression versus a chat-template/video-token mismatch." + "reason": "Both involve video, but the concrete regressions differ and the symptoms do not point to the same root cause." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are IndexError reports under ZeRO-3, but the shapes and loading paths differ; not clearly the same underlying defect." + "reason": "Both are IndexErrors, but one is a mask/index shape mismatch and the other is a zero-sized tensor during ZeRO-3 init." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Same model family, but the symptoms are not specific enough to prove one bug: degraded embeddings versus nonsensical pipeline outputs." + "reason": "Same model family, but different observed problems: degraded text embeddings versus nonsensical pipeline/AutoModel outputs." }, { "left": "issue:45759", "right": "issue:45902", "accept": false, - "reason": "Different bugs: config unwrapping for composite Qwen multimodal models versus uninitialized rotary-embedding memory after to_empty." + "reason": "These are different issues: composite-config handling versus uninitialized-memory access after meta to_empty(cuda) materialization." }, { "left": "issue:45542", "right": "issue:45800", "accept": false, - "reason": "One is a backend-selection problem with tensorboard/TensorFlow, the other is a torch/transformers compatibility issue." + "reason": "TensorBoard/TensorFlow backend detection is unrelated to the torch/transformers compatibility issue." }, { "left": "issue:39186", "right": "issue:44918", "accept": false, - "reason": "Different model families and failures: FSDP weight-dimension error versus Qwen3.5 input-embedding unpacking with TRL SFT." + "reason": "Both are model-loading/training errors, but the underlying failures are different: FSDP weight-shape handling versus TRL embedding unpacking." }, { "left": "issue:43716", "right": "issue:44918", "accept": false, - "reason": "Unrelated multimodal dtype mismatch versus Qwen3.5 embedding unpacking issue." + "reason": "Different model families and different failure modes; these are not the same bug." }, { "left": "issue:40990", "right": "issue:44960", "accept": false, - "reason": "A perplexity regression on gpt-oss-20b is unrelated to the generic GLM5 issue." + "reason": "Both are model-behavior complaints, but they concern unrelated models and problems." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Notebook custom-model initialization and flash-attn fallback blocking are different code paths." + "reason": "Notebook custom-model initialization and flash-attn2 fallback gating are different bugs in different paths." }, { "left": "issue:45632", "right": "issue:45698", "accept": false, - "reason": "Both involve custom-code loading, but one is a leaf-dir cache collision and the other is wrong module selection after save_pretrained; not the same bug." + "reason": "Both involve custom code loading, but one is a cache-path collision and the other is loading the wrong module after save_pretrained." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "lm_head weight tying in Mistral3 is distinct from the broader tied-weights issue for translation models." + "reason": "Both mention tied weights, but the concrete model families and failure modes differ too much to treat as one bug." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "FineGrainedFP8Config failures are unrelated to flash-attn fallback gating." + "reason": "FineGrainedFP8Config failure and flash-attn fallback gating are unrelated." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache group-type crash on Qwen3.5 is not the same as tied-weight handling." + "reason": "PagedAttentionCache group-type handling is unrelated to tied-weight loading regressions." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior is a different problem from tied weights." + "reason": "Quantization offloading behavior is a different subsystem and symptom than tied-weight issues." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading failure is unrelated to translation-model tied weights." + "reason": "NemotronH checkpoint loading failure is not the same bug as the translation-model tied-weights regression." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "A ReLU discussion in DSA indexing is unrelated to the removed transformers run command docs issue." + "reason": "A DSA indexer discussion and a docs/runtime command issue are unrelated." }, { "left": "issue:44918", "right": "issue:45507", "accept": false, - "reason": "Qwen3.5 SFT embedding unpacking and GraniteMoEHybrid invalid method calls are different bugs." + "reason": "TRl SFT embedding unpacking and GraniteMoEHybrid invalid method calls are different code paths and failures." }, { "left": "issue:44492", "right": "issue:44512", "accept": false, - "reason": "A typo fix in cache strategies is not the same as stale docs mentioning a removed command." + "reason": "A typo in cache strategies is unrelated to the removed run-command docs issue." }, { "left": "issue:43975", "right": "issue:44292", "accept": false, - "reason": "Both are model/runtime issues, but one is a detokenization bug and the other is a Qwen-3-8B-NVFP4 execution failure." + "reason": "Detokenization behavior for one model is unrelated to running a different quantized model." }, { "left": "issue:44292", "right": "issue:45507", "accept": false, - "reason": "Different models and failures: NVFP4 runtime error versus GraniteMoEHybrid invalid-method loading." + "reason": "Different model/runtime issues with no shared concrete failure path." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Both describe the same v5 regression where non-persistent buffers are corrupted/fill with junk during loading or initialization." + "reason": "Both describe the same Transformers v5 non-persistent-buffer corruption regression." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "The first is a docs example with model/processor mismatch and quantization failure; the second is a SigLIP2 output-quality complaint." + "reason": "Related by model family only; the documentation/example issues are not the same bug as the text-embedding quality complaint." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Same underlying non-persistent-buffer corruption regression in Transformers 5.x, described with equivalent symptoms." + "reason": "These reports point to the same non-persistent-buffer junk/regression; 43950 is the more specific instance." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different problems: autocast dtype mismatch in Phi-tiny-MoE versus RLHF NaN explosion from Qwen3.5 position_ids/SDPA fallback." + "reason": "Both mention dtype-related failures, but autocast mismatch and RLHF NaN/position-id collapse are different problems." }, { "left": "issue:36331", "right": "issue:41762", "accept": false, - "reason": "Trainer compute_loss signature mismatch is unrelated to DeepSpeed ZeRO-3 zero-size index loading errors." + "reason": "CustomTrainer unexpected kwarg and ZeRO-3 zero-size index errors are unrelated." }, { "left": "issue:43872", "right": "issue:44589", "accept": false, - "reason": "Bitsandbytes constructor incompatibility and Float8 storage lookup failure are different quantization-related bugs." + "reason": "bitsandbytes Int8Params incompatibility and missing Float8 storage are different dependency/runtime issues." }, { "left": "issue:43950", "right": "issue:45902", "accept": false, - "reason": "Non-persistent buffer corruption is not the same as Qwen3_5MoeTextRotaryEmbedding uninitialized-memory after to_empty." + "reason": "The titles are adjacent around buffer/materialization behavior, but 45902 is not clearly the same non-persistent-buffer corruption bug as 43950." }, { "left": "issue:43054", "right": "issue:45701", "accept": false, - "reason": "SigLIP2 embedding quality regression and tokenization-version-change reports are different issues." + "reason": "Different symptom sets: SigLIP2 embedding quality versus version-dependent tokenization changes." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Ovis2 checkpoint corruption is not the same as loading noise during transformer initialization." + "reason": "A corrupted checkpoint and load-time noise are not the same underlying bug." }, { "left": "issue:38175", "right": "issue:43054", "accept": false, - "reason": "Both involve SigLIP2, but one is zero probabilities and the other is worse embeddings; not enough evidence they are the same concrete bug." + "reason": "Same model family, but zero-probability outputs and weaker text embeddings are not enough to establish one root cause." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The clustering summary is well grounded in the packet and stays conservative. The only clearly supported duplicate chain is the non-persistent-buffer regression around 43644/43950/44534, and the other soft-edge rejections do not overclaim equivalence based on the available titles/bodies." + "accept": false, + "feedback": "The summary is a bit too strong for the packet. You can anchor the cluster on 43950, but the claims about an 'explicit repro' and 'clear regression framing' are not supported by the metadata shown here. More importantly, the accept on issue:45902 is not conservative enough from the title alone; it is adjacent but not clearly the same bug as the non-persistent-buffer corruption reports. The 43644/43950/44534 linkage is well grounded, but the rest should stay more tentative." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -1214,13 +1256,14 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ + "issue:36331", + "issue:41093", "issue:41553", "issue:41762", "issue:42491", "issue:42915", "issue:42947", "issue:43054", - "issue:43257", "issue:43404", "issue:43493", "issue:43643", @@ -1250,7 +1293,6 @@ "issue:44991", "issue:45020", "issue:45081", - "issue:45305", "issue:45356", "issue:45399", "issue:45520", @@ -1270,8 +1312,8 @@ "issue:43054|issue:43493", "issue:41553|issue:43881", "issue:44403|issue:44451", - "issue:43975|issue:45520", "issue:44403|issue:44488", + "issue:43975|issue:45520", "issue:43643|issue:45020", "issue:42947|issue:43856", "issue:41762|issue:42915", @@ -1293,228 +1335,228 @@ "issue:43872|issue:44291", "issue:42915|issue:43872", "issue:43792|issue:44220", - "issue:43257|issue:45305" + "issue:36331|issue:41093" ], - "prepared_review_unit_hash": "b90635dfa6438e8554e6753adad754ce8c8b1935900e6528be293525a2f4300a", + "prepared_review_unit_hash": "cfdd37fb995c472a1e762c6943bcaf8105a3be3e659de3f9c7932a921b99064d", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36796, - "estimated_input_tokens": 9199, - "estimated_eval_tokens": 18654 + "serialized_chars": 36830, + "estimated_input_tokens": 9208, + "estimated_eval_tokens": 18672 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a mixed bag of unrelated Transformers issues, with a few clear duplicate subclusters around docs pipeline removals, SigLIP2 implementation drift, the `_is_hf_initialized` loading regression, and Whisper feature extraction.", - "confidence": 0.64, - "canonical_issue_reason": "issue:45020 is the broadest, most representative v5 regression report in this set, covering the common theme of recent loading/remote_code breakage across models.", + "summary": "This cluster is mostly over-broad: it mixes many unrelated model-loading, tokenizer, training, and docs regressions. Only two soft pairs look like true duplicates; the rest are similar at a high level but not the same concrete bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43872 is the best anchor for the only clear duplicate pair: it names the concrete `_is_hf_initialized` / `init_empty_weights` TypeError, and issue:44291 reads like the same bug with a more specific wrapper title.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 has the widest scope and highest cross-model impact, making it the best anchor for the overall cluster theme of v5 loading regressions.", + "best_issue_reason": "issue:43872 is the most suitable global issue anchor because it is a crisp, reproducible library regression with a direct duplicate, whereas the other items are mostly model-specific or unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41553", "right": "issue:43883", "accept": false, - "reason": "Different failures: a bad AutoTokenizer error for Voxtral vs a missing `all_tied_weights_keys` attribute in Molmo." + "reason": "Different problems: Voxtral loading error message vs Molmo missing `all_tied_weights_keys`." }, { "left": "issue:45663", "right": "issue:45902", "accept": false, - "reason": "Both involve Gemma/Qwen MoE-adjacent models, but the concrete bugs differ: FSDP2 `shared_kv_states` KeyError vs uninitialized `inv_freq` memory." + "reason": "Both are Gemma/Qwen MoE-related, but the bugs are different: FSDP2 `KeyError` vs uninitialized `inv_freq` after `to_empty(cuda)`." }, { "left": "issue:44910", "right": "issue:45859", "accept": false, - "reason": "Different code paths: Flash Attention position-id handling on Qwen3.5 vs CPU-offload incompatibility in rotary embedding." + "reason": "Different code paths: FlashAttention packed-sequence crash vs CPU offload incompatibility in a MoE rotary embedding." }, { "left": "issue:42947", "right": "issue:44387", "accept": false, - "reason": "Gradient checkpointing/LoRA inefficacy is not the same bug as int4 quantization increasing reserved CUDA memory." + "reason": "One is gradient-checkpointing not being effective with PEFT LoRA; the other is increased reserved memory under int4 quantization." }, { "left": "issue:43054", "right": "issue:43493", - "accept": true, - "reason": "Both point to the same SigLIP2 implementation mismatch with the intended/reference behavior, surfaced as degraded text embeddings." + "accept": false, + "reason": "Same model family, but not enough to call it the same bug; one reports poor text embeddings, the other a HF-vs-JAX implementation discrepancy." }, { "left": "issue:41553", "right": "issue:43881", "accept": false, - "reason": "Unrelated models and errors: Voxtral tokenizer messaging vs glm-4v-9b load failure." + "reason": "Both are loading-related, but they concern different models and different failure modes." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "One is about noisy loading logs; the other is a specific model load failure for ScandiBERT." + "reason": "One is noisy logging on load; the other is an actual model-loading failure for ScandiBERT." }, { - "left": "issue:43975", - "right": "issue:45520", + "left": "issue:44403", + "right": "issue:44488", "accept": false, - "reason": "Tokenizer detokenization regression is unrelated to the Python 3.13 `flash_attn` import/key error." + "reason": "Same as above: docs/log noise vs a specific model load failure, not the same bug." }, { - "left": "issue:44403", - "right": "issue:44488", + "left": "issue:43975", + "right": "issue:45520", "accept": false, - "reason": "Loading noise is not the same as a concrete tokenizer/model load failure for `cjvt/sleng-bert`." + "reason": "Detokenization regression and a Python 3.13 import-utils KeyError are unrelated." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Both mention remote code, but one is a specific missing-fields bug in `AutoConfig` and the other is a broader model-loading regression." + "reason": "Both mention remote code, but one is missing fields in `AutoConfig`, the other is a broad report of recent remote-code breakage." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Distinct memory/training issues: ineffective gradient checkpointing vs inefficient MoE training memory usage." + "reason": "Different regressions: ineffective checkpointing vs MoE training memory usage." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Different model-loading failures: Gemma3 ZeRO-3 `IndexError` vs Qwen3 MoE fine-grained FP8 conversion/loading." + "reason": "Different models and different failure classes: ZeRO-3 loading index error vs FineGrainedFP8Config failure." }, { "left": "issue:45685", "right": "issue:45823", "accept": false, - "reason": "MPS histogram/int support in MoE is unrelated to Gemma4 PLE device placement with `device_map='auto'`." + "reason": "MPS integer support error and Gemma4 device mismatch are unrelated." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Qwen3VL weight-only loading and GLM5 are unrelated model-specific issues." + "reason": "Qwen3VL weight-only loading error vs a GLM5 issue; no same underlying bug is evident." }, { "left": "issue:43824", "right": "issue:43872", "accept": false, - "reason": "Model import error for Qwen2.5-VL is unrelated to the bitsandbytes `_is_hf_initialized` constructor-arg regression." + "reason": "ImportError for a missing class and a bitsandbytes `Int8Params` constructor regression are different bugs." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer loading for `est-roberta` and Flash Attention fallback gating are different problems." + "reason": "Tokenizer load failure and FlashAttention fallback gating are unrelated." }, { "left": "issue:43881", "right": "issue:43883", "accept": false, - "reason": "glm-4v-9b load failure is unrelated to Molmo missing tied-weights metadata." + "reason": "Different model families and different symptoms: load failure vs missing tied-weights metadata." }, { "left": "issue:45020", "right": "issue:45823", "accept": false, - "reason": "Remote-code loading regressions and Gemma4 PLE device mismatch are different concrete bugs." + "reason": "Remote-code breakage and Gemma4 device-map forward mismatch are distinct problems." }, { "left": "issue:44661", "right": "issue:45020", "accept": false, - "reason": "`add-new-model-like` failing with `TOKENIZER_MAPPING_NAMES` is not the same underlying issue as the broader remote-code regression." + "reason": "Tokenizer mapping registration failure and broad remote-code regressions are not the same bug." }, { "left": "issue:43828", "right": "issue:45823", "accept": false, - "reason": "Autocast dtype mismatch on Phi-tiny-MoE is distinct from Gemma4 PLE device mismatch." + "reason": "Autocast dtype mismatch and device-placement mismatch are different runtime issues." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "A version incompatibility for Qwen3 MoE LoRA is not the same as a tie-word-embeddings warning in Qwen3.5 fine-tuning." + "reason": "Version incompatibility for a trained LoRA model and a warning about `tie_word_embeddings` are not the same bug." }, { "left": "issue:45081", "right": "issue:45399", "accept": false, - "reason": "Mistral regex patch crash is unrelated to Flash Attention fallback being blocked by other checks." + "reason": "Tokenizer regex patch crash vs FlashAttention fallback checks; different code paths." }, { "left": "issue:45715", "right": "issue:46315", "accept": false, - "reason": "Both are tokenizer performance regressions, but they affect different conversion APIs and appear to have different root causes." + "reason": "Both are tokenizer performance regressions, but they affect opposite conversion directions and different loops." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Remote-code loading breakage and Kimi-K2.5 tokenizer codec/warning regression are separate issues." + "reason": "Broad remote-code regression vs Kimi-K2.5 tokenizer codec regression are not the same issue." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "MobileLLM key error and Flash Attention fallback gating are unrelated." + "reason": "Model load KeyError and FlashAttention availability gating are unrelated." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "`lm_head` tying in Mistral3 is a model-weight issue, not the same as non-persistent buffer corruption in `from_pretrained`." + "reason": "lm_head tying bug and silent corruption of non-persistent buffers are different regressions." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "These are the same docs regression: lingering references to pipeline tasks removed in v5." + "reason": "These describe the same docs bug: pipeline tasks removed in v5 are still referenced in documentation." }, { "left": "issue:43872", "right": "issue:44291", "accept": true, - "reason": "Both describe the same `_is_hf_initialized` incompatibility during meta/init-based loading, just from different user-facing entry points." + "reason": "Same underlying `_is_hf_initialized` / `init_empty_weights` TypeError during loading." }, { "left": "issue:42915", "right": "issue:43872", "accept": false, - "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is a different bug from the bitsandbytes constructor-arg incompatibility." + "reason": "Qwen3Moe FineGrainedFP8Config failure and bitsandbytes `_is_hf_initialized` regression are different bugs." }, { "left": "issue:43792", "right": "issue:44220", - "accept": true, - "reason": "Both point to the Whisper `_torch_extract_fbank_features()` path failing, with the broad model failure and the specific function-level bug matching." + "accept": false, + "reason": "Whisper loading failure and `_torch_extract_fbank_features()` issue are too weakly connected to treat as duplicates." }, { - "left": "issue:43257", - "right": "issue:45305", + "left": "issue:36331", + "right": "issue:41093", "accept": false, - "reason": "Qwen3 MoE weight conversion/loading and gradient averaging under DeepSpeed are different training-time bugs." + "reason": "Completely unrelated: trainer `compute_loss` API mismatch vs tensor mask shape mismatch." } ] }, "evaluator_result": { "accept": true, - "feedback": "The cluster summary is grounded in the packet and stays mostly conservative. The canonical/best issue choice of issue:45020 is reasonable as a broad remote_code regression anchor. The accepted soft-edge pairs are also defensible from the titles and issue descriptions, especially the docs pipeline-removal pair, the `_is_hf_initialized` loading regression pair, and the Whisper feature-extraction pair. I did not see any clear overstatement of evidence in the rejected pairs." + "feedback": "Grounded and conservative overall. The two accepted duplicate pairs are well supported by the packet titles: issue:43827 \u2194 issue:44509 (same docs/pipeline removal bug) and issue:43872 \u2194 issue:44291 (same `_is_hf_initialized` / `init_empty_weights` TypeError). The other soft-edge rejections are phrased cautiously and do not overclaim duplication." }, "evaluator_used": true, "retried": false, @@ -1538,18 +1580,17 @@ "issue:24643", "issue:30064", "issue:36010", - "issue:36331", "issue:38175", "issue:41093", "issue:41553", "issue:42617", "issue:42915", "issue:42947", + "issue:43257", "issue:43366", "issue:43531", "issue:43550", "issue:43632", - "issue:43638", "issue:43645", "issue:43646", "issue:43824", @@ -1572,6 +1613,7 @@ "issue:44918", "issue:44928", "issue:44991", + "issue:45305", "issue:45313", "issue:45325", "issue:45584", @@ -1585,7 +1627,7 @@ "issue:46402" ], "soft_pairs": [ - "issue:36331|issue:41093", + "issue:43257|issue:45305", "issue:42947|issue:45923", "issue:41553|issue:44991", "issue:43940|issue:45313", @@ -1615,235 +1657,228 @@ "issue:44661|issue:45823", "issue:44560|issue:44805", "issue:38175|issue:43994", - "issue:43632|issue:45834", - "issue:43638|issue:44661" + "issue:43632|issue:45834" ], - "prepared_review_unit_hash": "2d4c1dec1661a247c9c48811676498e23467b4404c9a41e660f4889f46a852ac", + "prepared_review_unit_hash": "4018b61d8a5c1c487d14b14caa0b57ac0af5f99bc2abbe2f0fd3bafba1ead3aa", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37483, - "estimated_input_tokens": 9371, - "estimated_eval_tokens": 18998 + "soft_pair_count": 31, + "serialized_chars": 37144, + "estimated_input_tokens": 9286, + "estimated_eval_tokens": 18828 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous set of bug reports rather than one tight duplicate cluster. The only strong duplicate-like pair is the Transformers 5.0 custom model initialization regression (43645/43646); the rest are mostly model- or error-specific issues that do not look mergeable.", - "confidence": 0.63, - "canonical_issue_reason": "issue:43646 is the cleaner, broader representative of the custom model initialization regression and best subsumes issue:43645.", + "summary": "This set is mostly heterogeneous: the items span unrelated model-loading, attention, training, and tokenizer regressions. I only see two true duplicate-like pairs: the transformers v5 custom model initialization regression (43645/43646) and the siglip2-base-patch16-224 bad-output reports (38175/43994).", + "confidence": 0.83, + "canonical_issue_reason": "issue:43646 is the broader, cleaner statement of the v5 custom-model initialization regression; issue:43645 is the notebook-specific reproduction.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best issue anchor here because it captures the clearest true duplicate pair and is phrased more generally than issue:43645.", + "best_issue_reason": "issue:43646 is the best representative issue because it states the underlying regression in a general, actionable way without narrowing it to Jupyter.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:36331", - "right": "issue:41093", + "left": "issue:43257", + "right": "issue:45305", "accept": false, - "reason": "Different failures: trainer API keyword mismatch vs. tensor/mask shape mismatch; no sign of the same code-path bug." + "reason": "Both involve Qwen/DeepSpeed training, but one is weight conversion during load and the other is gradient averaging with GAS; different bugs and fixes." }, { "left": "issue:42947", "right": "issue:45923", "accept": false, - "reason": "Both mention gradient checkpointing, but one is PEFT/LoRA effectiveness and the other is a missing flag on a trust_remote_code model." + "reason": "Gradient checkpointing appears in both, but one is LoRA checkpointing being ineffective and the other is a missing class flag on a trust_remote_code model." }, { "left": "issue:41553", "right": "issue:44991", "accept": false, - "reason": "One is a bad AutoTokenizer error message, the other is an actual tokenizer load failure for a different model." + "reason": "Both are tokenizer-loading complaints, but they hit different models and failure modes." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both involve ZeRO-3 weight loading, but they are different Qwen model families and likely different loader issues." + "reason": "Both are ZeRO-3 weight-loading failures, but one is Qwen3-Next with all params missing and the other is Qwen3.5 language_model weights; too model-specific to assume one fix." }, { "left": "issue:44928", "right": "issue:45588", "accept": false, - "reason": "Unrelated problems: Qwen3.5 RLHF/position_ids vs. flash-attention sinkless-model crash." + "reason": "Different code paths: RLHF NaNs from 3D position_ids vs flash_attention crash on s_aux=None." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Different models and different load failures; too broad to treat as the same bug." + "reason": "Both are model-loading problems, but they concern different models and errors." }, { "left": "issue:41093", "right": "issue:44805", "accept": false, - "reason": "The error strings are similar, but the reports are generic symptom matches without enough evidence of the same root cause." + "reason": "Same error class, but no evidence of the same root cause; likely separate model/path reproductions." }, { "left": "issue:44479", "right": "issue:44560", "accept": false, - "reason": "Both are video regressions, but they affect different model stacks and do not clearly share one concrete fix." + "reason": "Both mention video, but one is a broad v5.3.0 regression across Qwen VL variants and the other is a StopIteration in Qwen3-vl-embedding." }, { "left": "issue:41553", "right": "issue:44488", "accept": false, - "reason": "Both are tokenizer-related, but one is Voxtral error handling and the other is a specific model load regression." + "reason": "Both are loading failures, but Voxtral error messaging and sleng-bert tokenizer loading are different issues." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Completely different domains: DeepSpeed training runtime error vs. image processor segmentation-map handling." + "reason": "Unrelated: DeepSpeed training weight shape error vs image processor void segmentation maps." }, { "left": "issue:44863", "right": "issue:46402", "accept": false, - "reason": "Both are model-loading issues, but for different architectures and different failure modes." + "reason": "Different models and load paths; no shared underlying bug." }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Qwen3.5 PagedAttentionCache crash vs. NemotronH checkpoint loading; no shared concrete bug." + "reason": "PagedAttention/linear_attention crash vs checkpoint loading failure; unrelated." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Both mention model support/precision, but one is GGUF gpt-oss support and the other is an autocast dtype mismatch." + "reason": "GGUF architecture support vs autocast dtype mismatch; different problems." }, { "left": "issue:45325", "right": "issue:45902", "accept": false, - "reason": "Both concern rotary embeddings, but the specific bug is different: temporal scaling vs. uninitialized inv_freq after materialization." + "reason": "Different regressions: Qwen2.5-VL rope scaling vs Qwen3_5Moe inv_freq materialization." }, { "left": "issue:45663", "right": "issue:45823", "accept": false, - "reason": "Gemma4 FSDP2 per-layer state rebuild bug is unrelated to the device-mismatch forward failure." + "reason": "Same model family, but one is an FSDP2 shared_kv_states KeyError and the other is a device-map mismatch." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "One is GGUF architecture support, the other is a sliding_window issue in Qwen3-MoE; different changes are needed." + "reason": "GGUF gpt-oss support and Qwen3-MoE sliding_window are unrelated." }, { "left": "issue:44155", "right": "issue:45584", "accept": false, - "reason": "AudioFlamingo3 batch-leak issue and Whisper empty-transcription failure are unrelated." + "reason": "AudioFlamingo3 track leakage vs Whisper empty transcription handling; unrelated." }, { "left": "issue:43950", "right": "issue:45834", "accept": false, - "reason": "Non-persistent buffer corruption in from_pretrained is not the same as the Kosmos2.5 long-input index error." + "reason": "Non-persistent buffer corruption vs OCR index error; different bugs." }, { "left": "issue:44560", "right": "issue:45685", "accept": false, - "reason": "Video StopIteration regression and MPS histogram int-support error are unrelated." + "reason": "Video StopIteration vs MPS histogram limitation; unrelated." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "GGUF support and missing Float8 storage are different loading/runtime problems." + "reason": "GGUF support vs Float8 storage lookup failure; unrelated." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "ImportError for a missing class and meta-device model loading breakage are different failures." + "reason": "Missing exported class vs meta-device loading regression; different failure modes." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Qwen3Moe FP8Config failure is not the same as generic meta-device loading regressions." + "reason": "FineGrainedFP8Config failure vs meta-device loading regression; unrelated." }, { "left": "issue:43645", "right": "issue:43646", "accept": true, - "reason": "These describe the same Transformers 5.0 custom model initialization regression; the notebook wording in 43645 is just a usage context." + "reason": "Same v5 custom model initialization regression; the notebook report is just a narrower reproduction of the broader custom-model init bug." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Different model families and different symptoms: torch.compile+SDPA vs. Flash Attention position_ids misinterpretation." + "reason": "torch.compile SDPA crash on Bamba vs flash attention illegal memory access on Qwen3.5; different code paths." }, { "left": "issue:44560", "right": "issue:44918", "accept": false, - "reason": "Video StopIteration in Qwen3-vl-embedding is not the same as unpacking Qwen3.5 input embeddings in TRL SFT." + "reason": "Video inference StopIteration vs TRL SFT input-embedding unpacking; unrelated." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "3d_parallel.py runtime failure and GGUF gpt-oss support are unrelated." + "reason": "3d_parallel.py runtime issue vs GGUF gpt-oss support; unrelated." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "Importing GenerationMixin and Qwen3Moe FP8 config issues are unrelated." + "reason": "GenerationMixin import break vs Qwen3Moe FP8 config failure; unrelated." }, { "left": "issue:44661", "right": "issue:45823", "accept": false, - "reason": "Tokenizer mapping bookkeeping bug vs. Gemma4 device mismatch; different subsystem and fix." + "reason": "`add-new-model-like` tokenizer mapping bug vs Gemma4 device mismatch; different subsystems." }, { "left": "issue:44560", "right": "issue:44805", "accept": false, - "reason": "Same broad video/modeling area is not enough; the reported errors and models differ materially." + "reason": "Different models and symptoms; both are generic shape/runtime errors but no shared root cause." }, { "left": "issue:38175", "right": "issue:43994", - "accept": false, - "reason": "Both involve siglip2-base-patch16-224, but one reports zero probabilities and the other nonsensical AutoModel/pipeline outputs; not enough to prove one underlying bug." + "accept": true, + "reason": "Same siglip2-base-patch16-224 output regression; one reports zero probabilities, the other nonsensical AutoModel/pipeline outputs." }, { "left": "issue:43632", "right": "issue:45834", "accept": false, - "reason": "_is_hf_initialized flag regression and Kosmos2.5 OCR index error are unrelated." - }, - { - "left": "issue:43638", - "right": "issue:44661", - "accept": false, - "reason": "DeepSpeed ZeRO-3 zero-sized tensor error and tokenizer-mapping helper bug are different problems." + "reason": "Transformers v5 `_is_hf_initialized` regression and Kosmos2.5 OCR index error are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only clearly supported duplicate-like pair is 43645/43646, and the other soft edges are reasonably marked false without overclaiming shared root causes. No mergeability overstatement is evident." + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted near-duplicates are the clearly related Transformers v5 custom-model initialization reports (43645/43646) and the siglip2-base-patch16-224 output regression reports (38175/43994); the other pair rejections are cautious and supported by the issue titles." }, "evaluator_used": true, "retried": false, @@ -1871,10 +1906,9 @@ "issue:42915", "issue:43257", "issue:43317", - "issue:43404", "issue:43425", - "issue:43454", "issue:43596", + "issue:43638", "issue:43644", "issue:43645", "issue:43708", @@ -1914,6 +1948,7 @@ "issue:46050" ], "soft_pairs": [ + "issue:43638|issue:44661", "issue:43596|issue:44661", "issue:44877|issue:45030", "issue:44661|issue:44805", @@ -1927,9 +1962,9 @@ "issue:44451|issue:44488", "issue:42491|issue:45276", "issue:45313|issue:45357", - "issue:43708|issue:44075", "issue:44488|issue:44991", "issue:39692|issue:43994", + "issue:43708|issue:44075", "issue:43644|issue:43645", "issue:45488|issue:45812", "issue:43856|issue:43873", @@ -1943,229 +1978,228 @@ "issue:44164|issue:44291", "issue:44910|issue:45588", "issue:45276|issue:45335", - "issue:43425|issue:43957", - "issue:43404|issue:43454" + "issue:43425|issue:43957" ], - "prepared_review_unit_hash": "9ff6eee356d2ac1722a76a6b966fe3d8404cd161853a21cc2970dd0a16a8bc30", + "prepared_review_unit_hash": "3d8af0154bbb4f7e9c879d43829d2ced4d9b9c23e24bf26cb2cafc6a1eef5b61", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 31, - "serialized_chars": 36975, - "estimated_input_tokens": 9244, - "estimated_eval_tokens": 18744 + "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": "Most items are distinct model- or feature-specific bugs, not duplicates. The only clear duplicate pair is the Transformers v5 non-persistent buffer corruption regression; the rest share broad subsystems at most, but not the same concrete code-path failure.", - "confidence": 0.86, - "canonical_issue_reason": "issue:44534 is the cleanest representative of the only confirmed duplicate subcluster: it states the v5 non-persistent-buffer junk regression in a general form, while issue:43644 is the same bug described more specifically.", + "summary": "This cluster is mostly a loose grab bag of distinct Transformers regressions, with only a few true duplicate pairs around the same concrete bug. The strongest duplicate groups are the non-persistent buffer initialization regression, the resize_token_embeddings propagation bug, and the repeated BERT loading failure reports.", + "confidence": 0.76, + "canonical_issue_reason": "issue:44488 is the best representative issue here: it is a concrete, user-facing loading regression, has a direct follow-up report (issue:44451), and is broader/more established than the earlier report.", "canonical_pr_reason": null, - "best_issue_reason": "There is no single issue that cleanly represents the whole set because the cluster spans many unrelated bugs; issue:44534 is the best representative of the one true duplicate theme and has the broadest, most canonical wording.", + "best_issue_reason": "issue:44488 is the strongest global issue candidate because it most clearly captures a real duplicate pair and has the best combination of breadth and discussion activity.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Different failures in different code paths: ZeRO-3/BertModel indexing versus add-new-model-like tokenizer mapping behavior." + }, { "left": "issue:43596", "right": "issue:44661", "accept": false, - "reason": "Different failures: Deepspeed ZeRO-3/BertModel index error vs tokenizer-mapping failure when adding a model-like entry." + "reason": "Both mention BertModel-related failures, but the underlying bug reports are unrelated." }, { "left": "issue:44877", "right": "issue:45030", "accept": false, - "reason": "Both involve config validation, but they hit different models and different validation paths; not enough evidence of one underlying bug." + "reason": "Both are config-validation complaints, but they target different models and different validation surfaces." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Tokenizer-mapping insertion bug and mask-shape IndexError are unrelated." + "reason": "Tokenizer mapping registration and mask shape IndexError are unrelated bugs." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Both touch loading/offload/memory, but one is quantized gpu+cpu offload loading and the other is mmap OOM on Strix Halo." + "reason": "Both involve loading/offload behavior, but one is dequantized model device mapping 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 at inference and the other is a documentation example with model/processor mismatch and quantization errors." + "reason": "Same model family, but one is a zero-probability runtime issue and the other is a documentation/example error with additional quantization problems." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Same v5 regression: non-persistent buffers are filled with junk; the titles describe the same concrete bug." + "reason": "These are the same non-persistent-buffer regression in Transformers 5, just reported under two titles." }, { "left": "issue:42491", "right": "issue:43720", "accept": false, - "reason": "Different code paths: LoRA model incompatibility on hf4.x/hf5.x vs packed-weight unpacking during accelerate loading." + "reason": "Different model stacks and failures: Qwen3 MoE LoRA loading versus BitNet packed-weight unpacking." }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Both are Qwen3-family issues, but one is auto device mapping cuda assert and the other is FineGrainedFP8Config failure." + "reason": "Both are Qwen3/Qwen3-MoE related, but the concrete failures and triggers differ." }, { "left": "issue:44960", "right": "issue:44977", "accept": false, - "reason": "Different model families and different flash-attention symptoms." + "reason": "GLM5 and Qwen3.5 flash-attention generation failures are unrelated." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Qwen3 MoE FP8 failure is unrelated to Strix Halo mmap OOM." + "reason": "Qwen3-MoE FP8 loading and Strix Halo mmap/OOM are separate issues." }, { "left": "issue:44451", "right": "issue:44488", - "accept": false, - "reason": "Both are model-loading failures, but they are separate model-specific reports with no shared concrete bug shown." + "accept": true, + "reason": "Both report the same BERT loading regression on different models, with 44488 as the broader follow-up report." }, { "left": "issue:42491", "right": "issue:45276", "accept": false, - "reason": "LoRA/accelerate compatibility issue vs resize_token_embeddings not propagating to Gemma4 outputs." + "reason": "Qwen3 MoE LoRA loading and resize_token_embeddings propagation are different problems." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 problems: ZeRO-3 language_model loading vs incorrect visual encoder keys on save_pretrained." - }, - { - "left": "issue:43708", - "right": "issue:44075", - "accept": false, - "reason": "Trainer checkpoint math bug and SGD argument handling are unrelated." + "reason": "Both are Qwen3.5-related, but one is ZeRO-3 weight loading and the other is save_pretrained visual-encoder key regression." }, { "left": "issue:44488", "right": "issue:44991", "accept": false, - "reason": "Separate tokenizer/model-loading failures for different model families." + "reason": "Both are tokenizer/load regressions, but they affect different model families and likely different root causes." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both mention SigLIP2, but one is a broken docs example and the other is incorrect inference behavior with AutoModel/pipeline." + "reason": "Both involve SigLIP2, but the docs/example issue also includes quantization failures and is not clearly the same underlying bug as the nonsensical-output report." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer checkpoint resume math and SGD optimizer-arg handling are unrelated." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "One is a buffer-initialization regression; the other is notebook behavior for custom models. Different bugs." + "reason": "Shared Transformers 5 context is too broad; one is buffer initialization junk, the other is Jupyter custom-model initialization." }, { "left": "issue:45488", "right": "issue:45812", "accept": false, - "reason": "Tokenizer corruption bugs, but different mechanisms and different model families (LlamaTokenizer pre-tokenizer override vs Granite token IDs regression)." + "reason": "Both are tokenizer regressions, but they affect different model families and different tokenizer code paths." }, { "left": "issue:43856", "right": "issue:43873", "accept": false, - "reason": "Qwen3 MoE training memory inefficiency vs quantization offloading not working are different problems." + "reason": "Training memory usage and quantized offloading behavior are different issues." }, { "left": "issue:43856", "right": "issue:44387", "accept": false, - "reason": "Both concern memory, but one is training inefficiency and the other is int4 reserved-memory OOM regression." + "reason": "Both mention memory pressure, but one is Qwen3 MoE training efficiency and the other is int4 quantization OOM in Transformers 5.x." }, { "left": "issue:43854", "right": "issue:43931", "accept": false, - "reason": "Different model-load failures: unit-test load issue for GLM-4.7-Flash vs weight-shape mismatch for Qwen3-VL." + "reason": "GLM-4.7-Flash test loading and Qwen3-VL weight-shape mismatch are unrelated model-loading bugs." }, { "left": "issue:43257", "right": "issue:43854", "accept": false, - "reason": "Accelerate+deepspeed MoE weight conversion bug is unrelated to GLM-4.7-Flash loading." + "reason": "Qwen3 MoE accelerate+deepspeed weight conversion and GLM-4.7-Flash unit-test loading do not share the same root cause." }, { "left": "issue:45362", "right": "issue:45406", "accept": false, - "reason": "Qwen3.5 chat crash and Gemma4 processor missing _tokenizer are unrelated failures." + "reason": "Qwen3.5 chat crash and Gemma4Processor serve crash are distinct failures." }, { "left": "issue:44960", "right": "issue:46050", "accept": false, - "reason": "Unrelated: GLM5 vs quantized model save/load regression." + "reason": "GLM5 and sinq save/load quantization breakage are unrelated." }, { "left": "issue:42491", "right": "issue:45292", "accept": false, - "reason": "LoRA/qwen3_moe compatibility issue vs a separate resize_token_embeddings/output_embeddings bug." + "reason": "LoRA model compatibility on Qwen3 MoE is not the same bug as resize_token_embeddings not updating output embeddings." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Different model-loading failures with no shared concrete root cause." + "reason": "Different model families and failure modes." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Save/from_pretrained extra_state handling and init_empty_weights _is_hf_initialized are different serialization/loading bugs." + "reason": "Both touch loading/saving, but extra_state handling and unexpected _is_hf_initialized arguments are different mechanics." }, { "left": "issue:44910", "right": "issue:45588", "accept": false, - "reason": "Both are flash-attention crashes, but one is a packed-sequence position_ids assertion and the other is an s_aux=None AttributeError." + "reason": "Both are flash-attention crashes, but the triggering conditions and code-path failures are different." }, { "left": "issue:45276", "right": "issue:45335", - "accept": false, - "reason": "Both are resize_token_embeddings regressions, but they affect different architectures and different embedding targets; not clearly the same code-path bug." + "accept": true, + "reason": "Same resize_token_embeddings propagation bug, reported on two model variants." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Torch 2.10 compatibility and meta-device model loading are unrelated." - }, - { - "left": "issue:43404", - "right": "issue:43454", - "accept": false, - "reason": "Both involve untied lm_head weights, but they are reported on different model classes and do not clearly describe the same concrete defect." + "reason": "Torch version incompatibility and meta-device model loading are separate problems." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and the soft-edge calls are appropriately conservative. The only clearly supported duplicate pair is issue:43644 / issue:44534, which share the same v5 non-persistent-buffer regression wording. The other rejected pairs are plausibly distinct based on the titles and available metadata." + "feedback": "The summary is grounded in the packet and the soft-edge decisions are conservative. The accepted pairs are well supported by title similarity and cluster metadata, and the rejected pairs avoid overclaiming when only broad thematic overlap is present. The canonical issue choice for issue:44488 is also reasonable given its discussion activity and the close 44451/44488 pair." }, "evaluator_used": true, "retried": false, @@ -2193,12 +2227,12 @@ "issue:42947", "issue:43335", "issue:43381", + "issue:43404", "issue:43425", - "issue:43611", + "issue:43454", "issue:43645", "issue:43646", "issue:43708", - "issue:43716", "issue:43828", "issue:43854", "issue:43873", @@ -2236,6 +2270,7 @@ "issue:46315" ], "soft_pairs": [ + "issue:43404|issue:43454", "issue:41720|issue:44155", "issue:39692|issue:44960", "issue:44402|issue:44488", @@ -2260,262 +2295,254 @@ "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:45292|issue:45335", "issue:43425|issue:43828", "issue:44402|issue:44451", "issue:44589|issue:44756", "issue:39404|issue:45701", "issue:44683|issue:45823", - "issue:43425|issue:45823", - "issue:43611|issue:43645", - "issue:43425|issue:43716" + "issue:43425|issue:45823" ], - "prepared_review_unit_hash": "bebe1b04a878b46d3e597927c4dfaa096d6f77b89c3183667c57854af1a73b6b", + "prepared_review_unit_hash": "d297deeff14f51190a8914a58e4a854ddad4b743dbced83337cb8f5446355f79", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 35, - "serialized_chars": 38217, - "estimated_input_tokens": 9555, - "estimated_eval_tokens": 19366 + "soft_pair_count": 34, + "serialized_chars": 37960, + "estimated_input_tokens": 9490, + "estimated_eval_tokens": 19236 }, "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 regression issues. The only clear duplicate family I see is the `resize_token_embeddings` bug affecting model-specific embeddings (`output_embeddings`, `embed_tokens_per_layer`, and `decoder.embed_tokens`).", - "confidence": 0.72, - "canonical_issue_reason": "Issue 45292 is the most generic statement of the duplicate family (`resize_token_embeddings` not updating `output_embeddings`), so it best represents the shared bug across the model-specific reports.", + "summary": "The cluster is mostly heterogeneous; nearly all soft pairs are related only by broad subsystem similarity, not the same concrete bug. The only clear duplicate-like subcluster is around `resize_token_embeddings` propagation to tied/output embeddings.", + "confidence": 0.36, + "canonical_issue_reason": "`issue:45292` is the clearest, most general representative of the resize-token-embedding regression; `issue:45276` and `issue:45335` look like narrower manifestations of the same fix area.", "canonical_pr_reason": null, - "best_issue_reason": "45292 is the broadest and most reusable issue for the resize-token-embedding regression; the others are narrower model-specific variants.", + "best_issue_reason": "`issue:45292` is the most succinct and broadly framed report of the only near-duplicate issue family in this set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both are weight-tying complaints, but they concern different models and likely different code paths." + }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Different models and failures: CUDA assert in Qwen3 auto device map vs embedding/token leakage in AudioFlamingo3 batched inference." + "reason": "Different models and failures: CUDA assert during device mapping vs cross-track token/embedding leakage." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "Unrelated topics: SigLIP2 doc example errors vs a generic GLM5 issue title with no matching failure mode." + "reason": "Docs/example errors for SigLIP2 vs an unrelated GLM5 issue; no shared underlying bug." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Both involve model loading, but the titles point to different models and no clear shared code-path failure is shown." + "reason": "Both mention model loading/tokenizer issues, but the affected models and failure modes differ." }, { "left": "issue:42947", "right": "issue:45704", "accept": false, - "reason": "Both mention memory/optimization concerns, but one is PEFT gradient checkpointing and the other is an apex RMSNorm memory leak." + "reason": "Gradient checkpointing/LoRA vs an Apex RMSNorm memory leak are unrelated bugs." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "ImportError for `GenerationMixin` is unrelated to Whisper pipeline `return_language` regression." + "reason": "Import path breakage for `GenerationMixin` is unrelated to Whisper pipeline `return_language` behavior." }, { "left": "issue:43873", "right": "issue:44387", "accept": false, - "reason": "Both mention quantization/OOM, but the symptoms and likely code paths differ enough that they do not look like the same bug." + "reason": "Both touch quantization/offloading, but one is offloading failure and the other is increased reserved memory/OOM." }, { "left": "issue:45663", "right": "issue:45759", "accept": false, - "reason": "Gemma4 FSDP2 KeyError and Qwen multimodal `from_config` unwrapping are different bugs." + "reason": "Different models and problems: Gemma4 FSDP attention state vs Qwen config unwrapping." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "Unexpected `_is_hf_initialized` argument during `init_empty_weights` is unrelated to Strix Halo mmap OOM." + "reason": "`init_empty_weights` argument handling and Strix Halo mmap/OOM are unrelated." }, { "left": "issue:45663", "right": "issue:45859", "accept": false, - "reason": "Gemma4 shared_kv_states FSDP bug and Qwen3_5Moe CPU-offload incompatibility are different failures." + "reason": "Unrelated model-specific issues: Gemma4 FSDP key error vs Qwen3.5-MoE CPU offload incompatibility." }, { "left": "issue:44155", "right": "issue:45823", "accept": false, - "reason": "AudioFlamingo3 embedding leakage is unrelated to Gemma4 device-map forward mismatch." + "reason": "AudioFlamingo batched inference leakage and Gemma4 device mismatch are different bugs." }, { "left": "issue:44387", "right": "issue:44756", "accept": false, - "reason": "Int4 quantization OOM due reserved memory and Strix Halo mmap OOM are different underlying issues." + "reason": "Both mention memory issues, but the causes and subsystems are different." }, { "left": "issue:44292", "right": "issue:44912", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error and git-oss MXFP4 fallback to bf16 are different quantization-loading problems." + "reason": "Different quantization/model-loading regressions with no concrete shared code-path." }, { "left": "issue:43957", "right": "issue:44756", "accept": false, - "reason": "Meta-device model loading regressions and Strix Halo mmap OOM do not look like the same bug." + "reason": "Meta-device loading breakage is unrelated to the mmap/OOM report." }, { "left": "issue:43335", "right": "issue:45663", "accept": false, - "reason": "SwitchTransformers sparse-layer config bug is unrelated to Gemma4 FSDP2 attention state reconstruction." + "reason": "SwitchTransformers sparse-layer construction and Gemma4 FSDP attention state are unrelated." }, { "left": "issue:42947", "right": "issue:43381", "accept": false, - "reason": "Both mention gradient checkpointing, but one is ineffective LoRA training behavior and the other is eval-mode incompatibility." + "reason": "Both are about gradient checkpointing, but one is ignored with PEFT LoRA and the other is an eval-mode restriction." }, { "left": "issue:45276", "right": "issue:45292", "accept": true, - "reason": "Same underlying `resize_token_embeddings` regression; 45276 is the Gemma4-specific manifestation and 45292 is the generic `output_embeddings` report." + "reason": "Same `resize_token_embeddings` propagation bug; `45276` is a model-specific instance of the same missing update to output embeddings." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "`save/from_pretrained` extra_state handling is unrelated to Strix Halo mmap OOM." + "reason": "Checkpoint save/load `extra_state` handling is unrelated to Strix Halo mmap OOM." }, { "left": "issue:44479", "right": "issue:44877", "accept": false, - "reason": "Video-input regression for Qwen VL models and strict-config loading for granite_speech are different issues." + "reason": "Video-input regression for Qwen VL models and strict Granite config loading are unrelated." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "GLM-4.7 unit-test loading failure and Qwen3.5 checkpoint save regression are separate save/load problems with different symptoms." + "reason": "Different model-loading regressions with no shared underlying failure." }, { "left": "issue:43381", "right": "issue:44928", "accept": false, - "reason": "Eval-mode gradient checkpointing failure is unrelated to RLHF NaN explosion from 3D position_ids and SDPA fallback." + "reason": "Gradient checkpointing in eval mode is unrelated to RLHF gradient explosion from 3D `position_ids`." }, { "left": "issue:45684", "right": "issue:45698", "accept": false, - "reason": "Both involve custom-model save/load, but one is about read-only file permissions and the other about loading the wrong custom module." + "reason": "Both involve save/load of custom models, but one is a permissions propagation issue and the other is wrong-module loading." }, { "left": "issue:45588", "right": "issue:45663", "accept": false, - "reason": "Flash-attention crash on `s_aux=None` and Gemma4 FSDP2 KeyError are distinct code-path failures." + "reason": "Flash-attention `s_aux=None` crash and Gemma4 FSDP attention-state rebuilding are unrelated." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Custom model initialization regressions and Trainer `resume_from_checkpoint` max_steps miscalculation are unrelated." + "reason": "Custom model initialization breakage and checkpoint resume batch-size math are different problems." }, { "left": "issue:43645", "right": "issue:44164", "accept": false, - "reason": "Notebook custom-model initialization and `save/from_pretrained` extra_state handling are different bugs." - }, - { - "left": "issue:45292", - "right": "issue:45335", - "accept": true, - "reason": "Same `resize_token_embeddings` bug family; both report model-specific embeddings not being updated, here `output_embeddings` versus `decoder.embed_tokens`." + "reason": "Notebook custom-model initialization and `extra_state` save/load handling are not the same bug." }, { "left": "issue:45812", "right": "issue:46315", "accept": false, - "reason": "Tokenizer ID regression and slow `convert_tokens_to_ids` performance are different tokenizer issues." + "reason": "Wrong token IDs and tokenizer performance regression are different failure modes." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Autocast dtype mismatch for Phi-tiny-MoE and meta-device loading breaks are not the same failure." + "reason": "Autocast dtype mismatch and meta-device loading failures are unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same resizing propagation bug across architectures: both report `resize_token_embeddings` failing to update tied embeddings." }, { "left": "issue:43425", "right": "issue:43828", "accept": false, - "reason": "Torch 2.10 compatibility issue and autocast dtype mismatch are too different to merge as one bug." + "reason": "Torch version compatibility and Phi MoE autocast dtype mismatch are unrelated." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Tokenizer vocab-size mismatch for one model and inability to load another model are not clearly the same underlying bug." + "reason": "Tokenizer vocab mismatch for one model and load failure for another do not look like the same defect." }, { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Float8 storage lookup failure and Strix Halo mmap OOM are unrelated." + "reason": "Float8 storage lookup error and mmap/OOM are unrelated." }, { "left": "issue:39404", "right": "issue:45701", "accept": false, - "reason": "Whisper pipeline language return regression is unrelated to tokenizer behavior changes across versions." + "reason": "Whisper pipeline `return_language` regression is unrelated to tokenizer output changing across versions." }, { "left": "issue:44683", "right": "issue:45823", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and Gemma4 device-map forward mismatch are different issues." + "reason": "Compiled flex attention on new torch versions and Gemma4 device mismatch are unrelated." }, { "left": "issue:43425", "right": "issue:45823", "accept": false, - "reason": "Torch 2.10 incompatibility and Gemma4 device mismatch are separate problems." - }, - { - "left": "issue:43611", - "right": "issue:43645", - "accept": false, - "reason": "Both are custom-model loading regressions in Transformers 5.0, but the titles indicate different failure modes and not one concrete bug." - }, - { - "left": "issue:43425", - "right": "issue:43716", - "accept": false, - "reason": "Torch compatibility issue and Mistral-3 image preprocessor/model dtype mismatch are unrelated." + "reason": "Torch 2.10 compatibility and Gemma4 `device_map=\"auto\"` forward mismatch are different issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analyst output is grounded in the packet and stays conservative on the soft-edge pairs. The only clearly supported duplicate family is the resize_token_embeddings regression centered on issue:45292, and the accepted links to issue:45276 and issue:45335 are well-justified by the titles. No overstatement of evidence stands out." + "feedback": "The summary is grounded in the packet and stays conservative. The resize_token_embeddings cluster (45276/45292/45335) is the only clearly supported near-duplicate family, and the other pairwise rejections are generally cautious and aligned with the issue titles. I don\u2019t see an overstatement that would require rejection." }, "evaluator_used": true, "retried": false, @@ -2549,8 +2576,8 @@ "issue:43596", "issue:43611", "issue:43632", - "issue:43643", "issue:43644", + "issue:43645", "issue:43716", "issue:43856", "issue:43901", @@ -2586,6 +2613,8 @@ "issue:46402" ], "soft_pairs": [ + "issue:43611|issue:43645", + "issue:43425|issue:43716", "issue:44530|issue:45084", "issue:43906|issue:44530", "issue:39404|issue:44912", @@ -2593,8 +2622,8 @@ "issue:44479|issue:44805", "issue:44589|issue:45084", "issue:43906|issue:46326", - "issue:43901|issue:44509", "issue:43425|issue:44918", + "issue:43901|issue:44509", "issue:46050|issue:46326", "issue:43994|issue:44079", "issue:36032|issue:43452", @@ -2616,229 +2645,210 @@ "issue:43716|issue:45084", "issue:46285|issue:46402", "issue:40990|issue:43421", - "issue:36032|issue:42915", - "issue:43643|issue:44530" + "issue:36032|issue:42915" ], - "prepared_review_unit_hash": "7db6ba6ab36aad951037da6239c779da2983faea122bd37d3ff2f5b6f8ffcb75", + "prepared_review_unit_hash": "a47c9c1f9bf57fe409a20d3e11f6d4e5040e04dc9188671147834f4f230144d6", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37406, - "estimated_input_tokens": 9352, - "estimated_eval_tokens": 18960 + "soft_pair_count": 33, + "serialized_chars": 37700, + "estimated_input_tokens": 9425, + "estimated_eval_tokens": 19106 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: most items are independent model-/tokenizer-/platform-specific regressions or docs issues, with no true duplicate pair among the provided soft edges. The closest umbrella-style issue is a Transformers v5 model-loading regression, but it is not a duplicate anchor for the rest.", - "confidence": 0.28, - "canonical_issue_reason": "Issue 43611 is the broadest representative of the cluster because it describes a general Transformers v5 model-loading regression; however, the rest of the items are too varied to treat as one duplicate family.", + "summary": "These issues are mostly distinct regressions or feature requests across different models and subsystems; the soft-linked pairs share broad themes but not the same concrete bug or fix. No duplicate cluster emerges here.", + "confidence": 0.9, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "43611 is the best single issue to represent the cluster since it is a clear, broad loading-regression report and more umbrella-like than the other narrow, model-specific bugs.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both relate to Transformers v5 model-loading regressions, but one is about `base_model_prefix` handling and the other about custom model init in Jupyter; different failure modes and fixes." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Unrelated problems: torch version incompatibility versus an image-preprocessor/model dtype mismatch for Mistral-3." + }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "Different bugs: a Qwen3.5 PagedAttentionCache crash vs a template-node compilation TypeError." + "reason": "Both are crashes, but one is Qwen3.5 cache/group-type handling and the other is a template compilation error; not the same code path." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "Unrelated failures; one is a reproduction of an earlier issue, the other is a Qwen3.5 cache crash." + "reason": "One is a reproduction of another issue, but the topics differ: an isolated prior bug reproduction versus a Qwen3.5 PagedAttentionCache crash." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to gpt-oss MXFP4 quantization fallback." + "reason": "Whisper `return_language` pipeline regression and MXFP4 quantization fallback are unrelated model/pipeline bugs." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "Different code paths and symptoms: torch.export + torch_compilable_check vs Perceiver interpolation failure." + "reason": "Different areas: `torch.export`/`torch_compilable_check` versus Perceiver image classification with non-default resolution." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "Video-input regression for Qwen VL models is not the same as a masking shape IndexError." + "reason": "Both involve runtime shape/path failures, but one is a video-input regression for Qwen VL models and the other is a generic mask/tensor index error." }, { "left": "issue:44589", "right": "issue:45084", "accept": false, - "reason": "Float8 storage lookup failure is unrelated to the template-node compilation error." + "reason": "Distinct failures: missing Float8 storage type versus template compilation on non-template nodes." }, { "left": "issue:43906", "right": "issue:46326", "accept": false, - "reason": "Different tokenization/pipeline issues with no shared underlying bug." + "reason": "An isolated reproduction of an earlier issue is not the same bug as the mistral_common `continue_final_message` tokenizer bug." }, { - "left": "issue:43901", - "right": "issue:44509", + "left": "issue:43425", + "right": "issue:44918", "accept": false, - "reason": "Both are docs-related, but they cover different pipeline API removals and behaviors." + "reason": "Torch incompatibility is not the same as Qwen3.5 embedding unpacking with TRL SFT trainer." }, { - "left": "issue:43425", - "right": "issue:44918", + "left": "issue:43901", + "right": "issue:44509", "accept": false, - "reason": "Torch version incompatibility is unrelated to the TRL SFT embedding-unpacking failure." + "reason": "One is a docs/behavior mismatch for text classification; the other is documentation still referencing removed pipeline tasks." }, { "left": "issue:46050", "right": "issue:46326", "accept": false, - "reason": "Quantized model save/load breakage is unrelated to continue_final_message behavior." + "reason": "Quantized-model save/load corruption and mistral_common tokenizer state handling are unrelated." }, { "left": "issue:43994", "right": "issue:44079", "accept": false, - "reason": "Nonsensical model outputs are not the same bug as ModelOutput key assignment with None." + "reason": "SigLIP2 model/pipeline outputs and `ModelOutput` key assignment with `None` are different bugs in different layers." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "Tokenizer method-name conflict is unrelated to gguf_file loading failures." + "reason": "Tokenizer special-method conflict and `gguf_file` breaking model/tokenizer loading are different loading regressions." }, { "left": "issue:43299", "right": "issue:43931", "accept": false, - "reason": "Both involve Qwen3-VL, but one is a v5 loading break and the other is a weight-shape mismatch; not the same fix." + "reason": "Both affect Qwen3-VL/Qwen3VL loading, but one is a Moe-model version regression and the other is a weight-shape mismatch for a specific checkpoint." }, { "left": "issue:43596", "right": "issue:44683", "accept": false, - "reason": "DeepSpeed zero3 init IndexError is unrelated to compiled flex_attention on newer torch." + "reason": "DeepSpeed ZeRO-3 init indexing error and compiled flex_attention failure on newer torch are unrelated." }, { "left": "issue:44164", "right": "issue:45759", "accept": false, - "reason": "extra_state save/load handling and composite config unwrapping are separate code paths." + "reason": "`extra_state` save/load handling is not the same as `from_config` not unwrapping `text_config` for composite multimodal configs." }, { "left": "issue:44671", "right": "issue:44991", "accept": false, - "reason": "CamemBERT prediction regression and tokenizer loading failure are different model-specific bugs." + "reason": "CamemBERT masked LM prediction regression and tokenizer loading failure for a specific model are different issues." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Different tokenizer failures: est-roberta loading vs Mistral regex patching." + "reason": "Both involve tokenizer loading, but one is an EMBEDDIA tokenizer failure and the other is a Mistral regex patch crash; different causes." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Both are memory-related, but one is Qwen3 MoE training inefficiency and the other is mmap OOM on Strix Halo." + "reason": "Qwen3 MoE training memory inefficiency is not the same as disabling mmap on Strix Halo to avoid OOM." }, { "left": "issue:39404", "right": "issue:44671", "accept": false, - "reason": "Whisper pipeline regression and CamemBERT masked-LM regression are unrelated." + "reason": "Whisper pipeline language-return regression and CamemBERT masked LM predictions are unrelated model-specific bugs." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "extra_state serialization is unrelated to the Qwen VL video regression." + "reason": "`extra_state` serialization and Qwen VL video-input regression are distinct code paths." }, { "left": "issue:44488", "right": "issue:45081", "accept": false, - "reason": "sleng-bert loading failure is unrelated to the Mistral tokenizer backend-tokenizer crash." - }, - { - "left": "issue:44737", - "right": "issue:45859", - "accept": false, - "reason": "Both mention device/offload behavior, but they affect different models and different mechanisms." - }, - { - "left": "issue:44661", - "right": "issue:44683", - "accept": false, - "reason": "TOKENIZER_MAPPING_NAMES handling is unrelated to flex_attention compilation." - }, - { - "left": "issue:44488", - "right": "issue:44987", - "accept": false, - "reason": "Two separate model-loading regressions with different models and failure modes." - }, - { - "left": "issue:43582", - "right": "issue:43632", - "accept": false, - "reason": "Apple Silicon allocator TypeError and _is_hf_initialized flag breakage are unrelated." + "reason": "Loading `cjvt/sleng-bert` and Mistral tokenizer regex patching are different tokenizer/model compatibility issues." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Trainer GPU broadcast TypeError is unrelated to Whisper return_language behavior." + "reason": "Trainer GPU broadcast NoneType error and Whisper pipeline `return_language` regression are unrelated." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is non-persistent buffer corruption." + "reason": "Both are v5 regressions, but one is model loading via `base_model_prefix` and the other is non-persistent buffer initialization; different underlying bugs." }, { "left": "issue:43716", "right": "issue:45084", "accept": false, - "reason": "Image preprocessor/model dtype mismatch is unrelated to the template-node compilation error." + "reason": "Image preprocessor dtype mismatch and template compilation errors do not share a concrete failing code path." }, { "left": "issue:46285", "right": "issue:46402", "accept": false, - "reason": "Inconsistent CLIP structure and Sam3Model text-encoder loading are different model-specific regressions." + "reason": "CLIP structural inconsistency and Sam3 text encoder weights not loading are separate model-architecture regressions." }, { "left": "issue:40990", "right": "issue:43421", "accept": false, - "reason": "Perplexity results on gpt-oss are unrelated to runtime special-token post-processor updates." + "reason": "Perplexity anomaly on gpt-oss-20b and tokenizer runtime post-processor updates are unrelated." }, { "left": "issue:36032", "right": "issue:42915", "accept": false, - "reason": "Tokenizer method conflict and Qwen3 MoE FP8 config failure are unrelated." - }, - { - "left": "issue:43643", - "right": "issue:44530", - "accept": false, - "reason": "AutoConfig trust_remote_code field loss is unrelated to the PagedAttentionCache linear_attention crash." + "reason": "Tokenizer method conflict and Qwen3Moe fine-grained FP8 failure are distinct issues." } ] }, @@ -2877,6 +2887,7 @@ "issue:43482", "issue:43540", "issue:43541", + "issue:43643", "issue:43644", "issue:43645", "issue:43646", @@ -2896,7 +2907,6 @@ "issue:44530", "issue:44534", "issue:44610", - "issue:44683", "issue:44811", "issue:44849", "issue:44987", @@ -2912,11 +2922,12 @@ "issue:46326" ], "soft_pairs": [ + "issue:43643|issue:44530", "issue:34567|issue:40444", "issue:44451|issue:44987", "issue:43541|issue:45084", - "issue:40444|issue:46326", "issue:42175|issue:46178", + "issue:40444|issue:46326", "issue:42915|issue:44530", "issue:44811|issue:45584", "issue:43644|issue:43646", @@ -2934,8 +2945,8 @@ "issue:43645|issue:44534", "issue:43646|issue:44534", "issue:44403|issue:45081", - "issue:43404|issue:44466", "issue:42915|issue:45084", + "issue:43404|issue:44466", "issue:43883|issue:44991", "issue:43994|issue:44610", "issue:43322|issue:43540", @@ -2945,43 +2956,48 @@ "issue:44987|issue:45081", "issue:44265|issue:44610", "issue:44991|issue:46315", - "issue:42915|issue:44291", - "issue:44220|issue:44683" + "issue:42915|issue:44291" ], - "prepared_review_unit_hash": "49a696d3e9103c66670ef10f03449f309ac8023a3b089fab2f3c2660b3a1abc1", + "prepared_review_unit_hash": "b82d389e8bc4bf26da74f810d7731125a8e5137c1423a00f52d66f0d67634bc4", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 35, - "serialized_chars": 38325, - "estimated_input_tokens": 9582, - "estimated_eval_tokens": 19420 + "serialized_chars": 38378, + "estimated_input_tokens": 9595, + "estimated_eval_tokens": 19446 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items in this cluster are unrelated model- or backend-specific issues. The only clear duplicate-style pairs are the GGUF loading regressions and the audio feature-extraction crash pair; the rest are too broad or too different to merge safely.", - "confidence": 0.44, - "canonical_issue_reason": "issue:43644 is the broadest, clearest regression report here: it names the v5 behavior change, states a concrete symptom, and points at a likely shared initialization/code-path problem.", + "summary": "This cluster is mostly a grab-bag of unrelated Transformers regressions and model-loading bugs. A few pairs are superficially similar (especially v5/model-loading complaints), but none are strong enough to treat as the same underlying bug with mergeable fixes.", + "confidence": 0.74, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43644 is a strong standalone triage anchor because it is version-specific, symptom-focused, and describes a reproducible regression rather than a model-specific one-off.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "Different failures: AutoConfig fields missing vs Qwen3.5 PagedAttentionCache linear_attention crash." + }, { "left": "issue:34567", "right": "issue:40444", "accept": false, - "reason": "Different subsystems: TrainerState token accounting vs multimodal iterable-dataset failure." + "reason": "TrainerState token-count tracking bug vs Qwen2.5-VL IterableDataset multi-image finetuning failure." }, { "left": "issue:44451", @@ -2993,209 +3009,200 @@ "left": "issue:43541", "right": "issue:45084", "accept": false, - "reason": "Torch dynamo grouped_mm tracing error is unrelated to template compilation failure." + "reason": "Grouped_mm torchdynamo tracing error is unrelated to template compilation TypeError." }, { - "left": "issue:40444", - "right": "issue:46326", + "left": "issue:42175", + "right": "issue:46178", "accept": false, - "reason": "Multimodal dataset loading and mistral_common tokenizer chat-template handling are different bugs." + "reason": "Backend packaging omission vs BACKENDS_MAPPING error for tensorflow_text are not the same bug." }, { - "left": "issue:42175", - "right": "issue:46178", + "left": "issue:40444", + "right": "issue:46326", "accept": false, - "reason": "One is missing tensorflow in an extra install, the other is a backend-mapping error for tensorflow_text." + "reason": "Multi-image video/audio dataset bug vs mistral_common continue_final_message tokenizer bug." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "Qwen3Moe FP8 config failure and PagedAttentionCache linear_attention crash are different code paths." + "reason": "Qwen3Moe FP8 config failure is distinct from Qwen3.5 cache group-type crash." }, { "left": "issue:44811", "right": "issue:45584", "accept": false, - "reason": "Both involve Whisper tokens, but one is batch_decode skip_special_tokens and the other is empty-transcription generation after align_special_tokens." + "reason": "Both involve Whisper, but one is batch_decode token skipping and the other is empty-transcription generation after alignment." }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are v5 regressions, but one is junk non-persistent buffers and the other is custom model initialization; not clearly the same bug." + "reason": "Non-persistent buffer initialization corruption is not the same as custom model initialization breakage." }, { "left": "issue:43452", "right": "issue:43482", - "accept": true, - "reason": "Both report GGUF loading failures in transformers v5, with one being the general API path and the other a specific model manifestation of the same loader problem." + "accept": false, + "reason": "GGUF loader breakage is plausible, but one is generic from_pretrained/gguf_file and the other is a specific model loading report." }, { "left": "issue:44206", "right": "issue:44220", - "accept": true, - "reason": "These point to the same audio feature-extraction path: the unsupported center arg and the _torch_extract_fbank_features() issue look like one underlying crash." + "accept": false, + "reason": "Both audio-feature related, but the titles point to different failing layers and symptoms." }, { "left": "issue:43883", "right": "issue:44488", "accept": false, - "reason": "Missing tied-weights metadata and inability to load a specific model are not the same issue." + "reason": "Missing all_tied_weights_keys attribute vs specific model loading failure are unrelated." }, { "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Device-map dequantized offload failure is unrelated to GGUF tokenizer/model loading." + "reason": "Dequantized offload device_map failure is unrelated to Qwen2.5-GGUF loading." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer attribute conflict and custom model initialization regression are different failures." + "reason": "Tokenizer method-name conflict vs notebook custom-model init regression are different bugs." }, { "left": "issue:45084", "right": "issue:45507", "accept": false, - "reason": "Template compilation error and GraniteMoEHybrid invalid-method call are not the same concrete bug." + "reason": "Template-node compilation TypeError and GraniteMoEHybrid invalid-method call are different code paths." }, { "left": "issue:44987", "right": "issue:44991", "accept": false, - "reason": "Different models and different failure points: model loading vs tokenizer loading." + "reason": "Both are v5 loading failures, but on different models/tokenizers and with no shared concrete root cause." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Qwen3.5 hidden-state output bug and Qwen2.5-VL RoPE indexing bug are separate model bugs." + "reason": "Qwen3.5 hidden-states bug is unrelated to Qwen2.5-VL rope-index scaling." }, { "left": "issue:44464", "right": "issue:45987", "accept": false, - "reason": "Chunked generation inconsistency and StaticCache return-type contract are unrelated." + "reason": "Chunked generation compiled-forward inconsistency is unrelated to StaticCache seq_length return-type bug." }, { "left": "issue:43746", "right": "issue:46326", "accept": false, - "reason": "PEFT checkpoint loading and mistral_common continue_final_message behavior are different areas." + "reason": "PEFT checkpoint loading issue and mistral_common continue_final_message bug are distinct." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and MIMI encoder padding-mask behavior are unrelated." + "reason": "Audio sampling-rate defaulting and MIMI padding-mask batching bug are different issues." }, { "left": "issue:43645", "right": "issue:44534", "accept": false, - "reason": "Custom model init in notebooks and non-persistent-buffer junk are not the same underlying bug." + "reason": "Notebook custom model init regression is not the same as non-persistent buffer junk values." }, { "left": "issue:43646", "right": "issue:44534", "accept": false, - "reason": "Custom model initialization regression and junk non-persistent buffers are different symptoms and code paths." + "reason": "Custom model initialization breakage and buffer-junk regression are separate failure modes." }, { "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "Logging noise during loading and a tokenizer regex patch crash are not duplicates." + "reason": "Loading noise issue is unrelated to the mistral regex patch crash." }, { - "left": "issue:43404", - "right": "issue:44466", + "left": "issue:42915", + "right": "issue:45084", "accept": false, - "reason": "lm_head tying missing in one model class and device-dependent serialization inconsistency are related conceptually but not the same bug." + "reason": "Qwen3Moe FP8 config failure and non-template node compilation failure do not match." }, { - "left": "issue:42915", - "right": "issue:45084", + "left": "issue:43404", + "right": "issue:44466", "accept": false, - "reason": "FP8 configuration failure and template compilation failure are unrelated." + "reason": "Both mention lm_head/tied weights, but one is incorrect tying and the other is inconsistent serialization by device." }, { "left": "issue:43883", "right": "issue:44991", "accept": false, - "reason": "Missing all_tied_weights_keys and tokenizer loading failure are different issues." + "reason": "Missing all_tied_weights_keys attribute is unrelated to tokenizer loading failure for a specific model." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 nonsensical outputs and OmDet-Turbo processor size mismatch are different model/processor problems." + "reason": "SigLIP2 bad results and OmDet-Turbo processor size mismatch are different model/processor bugs." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Llava Next segfault and Qwen3OmniMoe video-input ValueError are unrelated multimodal loading errors." + "reason": "Segmentation fault loading Llava Next and Qwen3OmniMoe video ValueError are unrelated multimodal failures." }, { "left": "issue:43089", "right": "issue:45987", "accept": false, - "reason": "Generation performance overhead and StaticCache seq-length type bug are different concerns." + "reason": "Generation performance overhead is unrelated to StaticCache return-type contract violation." }, { "left": "issue:43377", "right": "issue:43819", "accept": false, - "reason": "MIMI padding-mask mismatch and DAC STE mismatch are separate numerical correctness bugs." + "reason": "Both are audio-codec correctness bugs, but they involve different models and different missing pieces." }, { "left": "issue:44451", "right": "issue:45081", "accept": false, - "reason": "General model-load failure for one checkpoint and Mistral regex patch crash are unrelated." + "reason": "Specific model load failure vs Mistral tokenizer regex crash are not the same underlying bug." }, { "left": "issue:44987", "right": "issue:45081", "accept": false, - "reason": "Loading a specific model checkpoint and fixing Mistral regex tokenization are different code paths." + "reason": "Loading a specific model and crashing in fix_mistral_regex are different issues." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export export failure with torch_compilable_check is unrelated to image processor size mismatch." + "reason": "torch.export/exportability failure is unrelated to OmDet-Turbo processor input-size mismatch." }, { "left": "issue:44991", "right": "issue:46315", "accept": false, - "reason": "Tokenizer load regression and slow convert_tokens_to_ids regression are different tokenizer bugs." + "reason": "Tokenizer load failure and convert_tokens_to_ids performance regression are different problems." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "FP8 config failure and init_empty_weights _is_hf_initialized argument error are unrelated." - }, - { - "left": "issue:44220", - "right": "issue:44683", - "accept": false, - "reason": "Audio fbank feature extraction and compiled flex_attention on torch>=2.9 are separate failures." + "reason": "FP8 config failure and init_empty_weights unexpected-argument failure are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary stays close to the packet and the two accepted pairings are plausibly the same underlying bugfixes rather than overreaching merges. The rejected pairs are also framed conservatively. One minor note: the canonical-issue rationale for issue:43644 leans a bit on an inferred shared initialization/code-path problem, so keeping that wording tentative would be safest, but it does not rise to a rejection." - }, - "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 }, @@ -3233,6 +3240,7 @@ "issue:43792", "issue:44062", "issue:44079", + "issue:44220", "issue:44291", "issue:44403", "issue:44462", @@ -3261,22 +3269,23 @@ "issue:46326" ], "soft_pairs": [ + "issue:44220|issue:44683", "issue:45216|issue:45357", - "issue:36032|issue:43742", "issue:44062|issue:44987", + "issue:36032|issue:43742", "issue:42915|issue:44589", "issue:46209|issue:46326", "issue:44479|issue:45325", "issue:43632|issue:44756", "issue:36032|issue:41720", "issue:43531|issue:44877", - "issue:43404|issue:45127", "issue:43296|issue:43531", + "issue:43404|issue:45127", "issue:40444|issue:45325", - "issue:43611|issue:44534", "issue:43425|issue:44403", - "issue:40990|issue:44811", + "issue:43611|issue:44534", "issue:43792|issue:44589", + "issue:40990|issue:44811", "issue:44403|issue:44991", "issue:43632|issue:44534", "issue:44079|issue:45663", @@ -3285,220 +3294,232 @@ "issue:43421|issue:44568", "issue:43381|issue:45923", "issue:44291|issue:44589", - "issue:35707|issue:40444", "issue:43482|issue:45538", + "issue:35707|issue:40444", "issue:44829|issue:45479", "issue:44462|issue:45812", "issue:43482|issue:43525" ], - "prepared_review_unit_hash": "cb75b7b84044a4abd7f605656ec179d094b2455ffd00bdd682d0780994bab408", + "prepared_review_unit_hash": "f9546b8a7f01efe082c75d8100b87658dd3b00dc7bb66c822a65400a633236b4", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35799, - "estimated_input_tokens": 8950, - "estimated_eval_tokens": 18156 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36644, + "estimated_input_tokens": 9161, + "estimated_eval_tokens": 18578 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most pairs are unrelated issues that only share a broad subsystem or version-regression theme. I would only merge the clearly same Qwen3.5 save_pretrained regression, and the closely matching Qwen2.5-VL/Qwen3 video-position regression. No PRs are present.", - "confidence": 0.84, - "canonical_issue_reason": "issue:45812 is the strongest issue anchor here: it has multiple inbound references, a concrete regression description, and broad impact across Granite models.", + "summary": "This cluster is mostly a noisy mix of unrelated Transformers regressions. Only one soft pair looks like a true duplicate: `AutoTokenizer` ignoring `tokenizer.json` appears to explain the Granite wrong-token-ID regression.", + "confidence": 0.63, + "canonical_issue_reason": "`issue:44462` is the clearest root-cause report: it describes `AutoTokenizer` ignoring `tokenizer.json`, which is broad enough to cover the Granite token-ID failure in `issue:45812`.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the best global tracker candidate because it is the broadest umbrella for recent transformers regressions affecting remote_code models.", + "best_issue_reason": "`issue:44462` is the best representative issue because it states the underlying loading bug directly and is broader than the downstream Granite-specific symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45216", - "right": "issue:45357", - "accept": true, - "reason": "Same Qwen3.5 save_pretrained regression family: both report incorrect checkpoint saving/keys after newer transformers versions." + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction vs compiled flex-attention; unrelated code paths." }, { - "left": "issue:36032", - "right": "issue:43742", + "left": "issue:45216", + "right": "issue:45357", "accept": false, - "reason": "Different failures and model families: T5 tokenizer method conflict vs MobileLLM load key error." + "reason": "Both touch Qwen3.5 save_pretrained, but one is a broad checkpoint regression and the other is a specific visual-encoder key issue; not enough to prove the same bug." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "Both are tokenizer-load errors, but they affect different models and the reported failure modes are not the same bug." + "reason": "Tokenizer `AddedToken` argument conflict vs a model-loading failure for a specific repo; different failure modes." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5 tokenizer method-name conflict vs MobileLLM loading key error; unrelated." }, { "left": "issue:42915", "right": "issue:44589", "accept": false, - "reason": "FP8-related only at a high level; one is a model/config failure, the other is a Torch storage-type lookup error." + "reason": "FP8 config failure on Qwen3MoE vs a missing Float8 storage type error; different underlying issues." }, { "left": "issue:46209", "right": "issue:46326", "accept": false, - "reason": "Completely different code paths: FP8 inference on B200 vs tokenizer continue_final_message behavior." + "reason": "FP8 inference on B200 vs tokenizer `continue_final_message`; unrelated." }, { "left": "issue:44479", "right": "issue:45325", - "accept": true, - "reason": "Likely the same Qwen-VL temporal-position/rope-index regression underlying video and still-image handling in 5.3.0." + "accept": false, + "reason": "General video-input regression across Qwen VL models vs a still-image temporal \u0564\u056b\u0580position-id scaling bug; distinct multimodal paths." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "Unrelated problems: _is_hf_initialized regression vs mmap/OOM on Strix Halo." + "reason": "HF initialization flag regression vs mmap/OOM behavior on Strix Halo; unrelated." }, { "left": "issue:36032", "right": "issue:41720", "accept": false, - "reason": "Different model and symptom: T5 tokenizer load bug vs Qwen3 auto-device-map CUDA assert." + "reason": "Tokenizer load conflict vs CUDA assert during auto device mapping; unrelated." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Both mention model configs, but one is a Qwen3-MoE sliding_window issue and the other is Granite Speech strict-config loading." - }, - { - "left": "issue:43404", - "right": "issue:45127", - "accept": false, - "reason": "Both touch tied weights, but one is a missing lm_head tie in Mistral3ForConditionalGeneration and the other is LoRA-merge collapse with extended vocab." + "reason": "Qwen3-MoE sliding-window behavior vs Granite speech config loading; unrelated." }, { "left": "issue:43296", "right": "issue:43531", "accept": false, - "reason": "Different areas and failures: PaddleOCR-VL load failure in vLLM vs Qwen3-MoE sliding_window behavior." + "reason": "PaddleOCR-VL load failure in vLLM vs Qwen3-MoE sliding-window issue; unrelated." }, { - "left": "issue:40444", - "right": "issue:45325", + "left": "issue:43404", + "right": "issue:45127", "accept": false, - "reason": "Both are Qwen-VL-related, but iterable-dataset multi-image finetuning is not the same bug as rope-index temporal scaling." + "reason": "Both involve tied embeddings, but one is a missing tie in Mistral3 image-text generation and the other is LoRA merge collapse after vocab extension; different bugs." }, { - "left": "issue:43611", - "right": "issue:44534", + "left": "issue:40444", + "right": "issue:45325", "accept": false, - "reason": "Different regressions: base_model_prefix loading breakage vs non-persistent buffer initialization junk." + "reason": "Multi-image IterableDataset finetuning failure vs still-image rope scaling; different bugs." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch version incompatibility is unrelated to the transformer-loading noise issue." + "reason": "Torch version incompatibility vs noisy loading output; unrelated." }, { - "left": "issue:40990", - "right": "issue:44811", + "left": "issue:43611", + "right": "issue:44534", "accept": false, - "reason": "Unrelated model/evaluation problems: perplexity on gpt-oss-20b vs Whisper batch_decode skip_special_tokens." + "reason": "Base-model-prefix loading regression vs non-persistent buffer initialization junk; different code paths." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Different failure surfaces: Whisper runtime failure vs missing Float8 storage type." + "reason": "Whisper runtime failure vs Float8 storage lookup error; unrelated." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "High perplexity on GPT-OSS vs Whisper batch_decode skip_special_tokens behavior; unrelated." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Generic loading noise vs a tokenizer-loading regression for a specific model." + "reason": "Load-time noise vs tokenizer loading failure for EMBEDDIA/est-roberta; not the same bug." }, { "left": "issue:43632", "right": "issue:44534", "accept": false, - "reason": "Different v5 regressions affecting separate initialization paths." + "reason": "Both are v5 regressions, but `_is_hf_initialized` and non-persistent buffer corruption are separate issues." }, { "left": "issue:44079", "right": "issue:45663", "accept": false, - "reason": "ModelOutput key assignment bug is unrelated to Gemma-4 FSDP2 shared_kv_states construction." + "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 shared_kv_states KeyError; unrelated." }, { "left": "issue:33666", "right": "issue:40444", "accept": false, - "reason": "Both are Qwen-VL training topics, but multi-GPU training and iterable-dataset multi-image finetuning are different problems." + "reason": "Multi-GPU Qwen2-VL training vs Qwen2.5-VL multi-image IterableDataset failure; related model family, but different concrete problems." }, { "left": "issue:44683", "right": "issue:45020", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 is a different issue than remote_code model loading regressions." + "reason": "Flex-attention compilation on torch>=2.9 vs remote_code loading regressions; unrelated." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Both involve special tokens, but one is runtime post-processor sync and the other is add_special_tokens not adding BOS/EOS for mdeberta." + "reason": "Runtime special-token post-processor update vs add_special_tokens not adding BOS/EOS; same general area, but different concrete bugs." }, { "left": "issue:43381", "right": "issue:45923", "accept": false, - "reason": "Gradient checkpointing in eval mode is not the same as a missing supports_gradient_checkpointing flag in a trust_remote_code model." + "reason": "Eval-mode checkpointing restriction vs missing supports_gradient_checkpointing flag; different issues." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "Different low-level breakages: init_empty_weights/_is_hf_initialized vs Float8 storage lookup." + "reason": "Unexpected `_is_hf_initialized` arg with init_empty_weights vs Float8 storage lookup failure; unrelated." }, { - "left": "issue:35707", - "right": "issue:40444", + "left": "issue:43482", + "right": "issue:45538", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to multi-image IterableDataset finetuning." + "reason": "GGUF loading failure vs CLIPTokenizer max-length constant; unrelated." }, { - "left": "issue:43482", - "right": "issue:45538", + "left": "issue:35707", + "right": "issue:40444", "accept": false, - "reason": "Qwen2.5-GGUF loading under transformers v5 and CLIPTokenizer model_max_length are different tokenizer/model issues." + "reason": "Progressive generation with inputs_embeds/past_key_values vs Qwen2.5-VL multi-image finetuning; unrelated." }, { "left": "issue:44829", "right": "issue:45479", "accept": false, - "reason": "Different training bugs: flash_attention_3 degenerate training vs zero-loss classification with num_labels=1." + "reason": "Flash-attention-3 training degeneration vs single-label num_labels=1 zero-loss behavior; different failure causes." }, { "left": "issue:44462", "right": "issue:45812", - "accept": false, - "reason": "Both are AutoTokenizer regressions, but one ignores tokenizer.json while the other produces wrong Granite token IDs; not the same concrete bug." + "accept": true, + "reason": "Same tokenizer-loading path: ignoring `tokenizer.json` can directly produce the wrong Granite token IDs, so these look like one underlying bug." }, { "left": "issue:43482", "right": "issue:43525", "accept": false, - "reason": "GGUF loading failure and missing pad_token_id are unrelated compatibility errors." + "reason": "GGUF loading failure vs missing `pad_token_id` on Llama4Config; unrelated." + }, + { + "left": "issue:44987", + "right": "issue:43525", + "accept": false, + "reason": "Specific model load failure under transformers>=5.1 vs missing `pad_token_id`; different issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: 45020 is a reasonable broad tracker, 45812 is the strongest anchor by inbound refs and concrete Granite regression, and the rejected pairs are mostly conservative and model/symptom-specific. The two accepted pairs are also plausibly mergeable within the same bugfix family: the Qwen3.5 save_pretrained regressions, and the Qwen-VL temporal/rope-index regression affecting video vs still-image handling. No PRs are present, so the issue-pair focus is appropriate." + "feedback": "The summary is grounded in the packet and stays mostly conservative. The only positive soft pair, `issue:44462` / `issue:45812`, is a reasonable duplicate candidate based on the titles, and the rest of the soft-edge rejections avoid overclaiming. The canonical issue choice is also defensible as the broader tokenizer-loading report." }, "evaluator_used": true, "retried": false, @@ -3573,17 +3594,17 @@ "issue:41553|issue:43582", "issue:43818|issue:46402", "issue:44079|issue:44534", - "issue:43818|issue:45357", "issue:43531|issue:44291", + "issue:43818|issue:45357", "issue:41720|issue:45081", "issue:43377|issue:45491", "issue:40444|issue:45381", "issue:44265|issue:45072", "issue:43644|issue:44079", "issue:43749|issue:44079", + "issue:42175|issue:43296", "issue:43611|issue:43646", "issue:45325|issue:45381", - "issue:42175|issue:43296", "issue:43792|issue:43957", "issue:43994|issue:44898", "issue:43632|issue:45072", @@ -3593,219 +3614,219 @@ "issue:43701|issue:45325", "issue:44291|issue:44877", "issue:44479|issue:45823", - "issue:43873|issue:45701", "issue:43296|issue:45800", + "issue:43873|issue:45701", "issue:43606|issue:44265", "issue:43262|issue:43746" ], - "prepared_review_unit_hash": "6078f8aa508ee727bfbf8c6b526ab0c0002ec7cc83fde162e383d9bd86b78772", + "prepared_review_unit_hash": "4ebaabb18230cac542a3256b27a08e7f82e59e0a047c2f3553d9542686e16e2c", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35891, - "estimated_input_tokens": 8973, - "estimated_eval_tokens": 18202 + "serialized_chars": 35894, + "estimated_input_tokens": 8974, + "estimated_eval_tokens": 18204 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is mostly a noisy bundle of unrelated closed issues (model loading, quantization/offload, backend packaging, and several v5 regressions). The only pair that looks like a true duplicate is the Qwen2.5-VL position-id regression; the rest are similar only at a broad subsystem level.", - "confidence": 0.64, - "canonical_issue_reason": "issue:45325 is the most specific/root-cause report in the only plausible duplicate pair: it names the shared Qwen2.5-VL temporal position-id bug in `get_rope_index` under transformers 5.3.0.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers bugs and regressions. Only the two Qwen2.5-VL position-id reports look like the same underlying issue; the rest are thematic but not true duplicates.", + "confidence": 0.68, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45325 is the best representative of the closest duplicate subtopic because it pinpoints a concrete code path and regression version, while most other items are adjacent but distinct bugs.", + "best_issue_reason": "issue:45325 is the clearest concrete report of the Qwen2.5-VL rope/position-id regression and best represents the only likely duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43317", "right": "issue:43856", "accept": false, - "reason": "Different bugs: dequantized model loading/offload failure vs Qwen3 MoE training memory usage." + "reason": "Offloading/quantization load failure is unrelated to Qwen3 MoE training memory usage." }, { "left": "issue:42175", "right": "issue:45542", "accept": false, - "reason": "Both involve TF/backend packaging, but one is missing TensorFlow in the torch extra and the other is a tensorboard-only backend error; not the same concrete bug." + "reason": "Both mention TensorFlow/backend packaging, but one is missing TF in the install extra and the other is a tensorboard-only environment error." }, { "left": "issue:41553", "right": "issue:43582", "accept": false, - "reason": "Unrelated: tokenizer loading error for Voxtral vs an Apple Silicon `caching_allocator_warmup` TypeError." + "reason": "Tokenizer loading error messaging and Apple Silicon allocator TypeError are different bugs." }, { "left": "issue:43818", "right": "issue:46402", "accept": false, - "reason": "Different model-specific loading bugs: Video-LLaVA temporal attention/weight sharing vs Sam3Model text encoder weights not loaded." + "reason": "Video-LLaVA temporal-attention/weight-sharing bug is unrelated to Sam3 text encoder weight loading." }, { "left": "issue:44079", "right": "issue:44534", "accept": false, - "reason": "Different code paths: `ModelOutput` key assignment vs non-persistent buffer serialization in v5." + "reason": "ModelOutput key assignment and non-persistent buffer initialization are separate code paths." }, { - "left": "issue:43818", - "right": "issue:45357", + "left": "issue:43531", + "right": "issue:44291", "accept": false, - "reason": "Both are multimodal vision issues, but one is Video-LLaVA architecture/weights and the other is Qwen3.5 save_pretrained key serialization." + "reason": "Qwen3-MoE sliding-window behavior is unrelated to the init_empty_weights argument regression." }, { - "left": "issue:43531", - "right": "issue:44291", + "left": "issue:43818", + "right": "issue:45357", "accept": false, - "reason": "Both are v5 regressions around loading/init, but the concrete failures differ: sliding-window handling vs unexpected `_is_hf_initialized` arg." + "reason": "Video-LLaVA weight sharing and Qwen3.5 visual-key serialization are different model families and failures." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Unrelated: Qwen3 device-mapping CUDA assert vs Mistral tokenizer regex patch crashing." + "reason": "CUDA assert in Qwen3 auto device mapping is unrelated to the Mistral tokenizer regex patch crash." }, { "left": "issue:43377", "right": "issue:45491", "accept": false, - "reason": "Both are sequence/batching bugs, but they affect different models and mechanisms: missing padding masks vs sliding-window NaNs." + "reason": "Both are batching issues, but one is a missing padding-mask bug in MIMI and the other is a Gemma3 sliding-window NaN issue." }, { "left": "issue:40444", "right": "issue:45381", - "accept": true, - "reason": "Likely the same Qwen2.5-VL positional-encoding regression: both describe broken image/video position IDs in the multimodal vision path." + "accept": false, + "reason": "Qwen2.5-VL multi-image finetuning failure and video vision_position_ids regression are related only loosely, not the same bug." }, { "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "Different failures: `torch.export` incompatibility with `torch_compilable_check` vs bfloat16 dtype mismatches in inference." + "reason": "torch.export failures from torch_compilable_check are unrelated to bfloat16 dtype mismatches." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Related to v5 internals, but one is `_is_hf_initialized`/buffer state and the other is `ModelOutput` key handling." + "reason": "Non-persistent buffer junk and ModelOutput key assignment are different regressions." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "Different bugs: FSDP CPU RAM efficient loading vs `ModelOutput` key assignment." + "reason": "FSDP CPU RAM efficient loading and ModelOutput key handling do not share the same failure mode." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "TF backend packaging and PaddleOCR-VL loading in vLLM are separate dependency/model-loading problems." }, { "left": "issue:43611", "right": "issue:43646", "accept": false, - "reason": "Both are v5 loading/init regressions, but one is `base_model_prefix` loading and the other is custom model initialization; not the same concrete bug." + "reason": "Both are v5 loading regressions, but one is about base_model_prefix and the other about custom model initialization; not the same concrete bug." }, { "left": "issue:45325", "right": "issue:45381", "accept": true, - "reason": "Strong duplicate candidate: both point to Qwen2.5-VL temporal/vision position-id handling being wrong in v5.3.0." - }, - { - "left": "issue:42175", - "right": "issue:43296", - "accept": false, - "reason": "Both touch backend/model loading areas, but one is install-time TF backend packaging and the other is vLLM loading PaddleOCR-VL." + "reason": "Both point to the same Qwen2.5-VL 5.3.0 position-id/rope-index regression affecting image and video inputs." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Different model-loading failures: Whisper runtime failure vs `torch.device('meta')` model-loading regression." + "reason": "Whisper-large-v2 runtime failure and meta-device loading regression are unrelated." }, { "left": "issue:43994", "right": "issue:44898", "accept": false, - "reason": "Both are vision model issues, but one is SigLIP2 inference quality and the other is Perceiver interpolation failure." + "reason": "Siglip2 nonsensical outputs and Perceiver non-default-resolution failure are different model-specific issues." }, { "left": "issue:43632", "right": "issue:45072", "accept": false, - "reason": "Distinct issues: `_is_hf_initialized` regression vs inference dtype mismatch." + "reason": "_is_hf_initialized flag regression is unrelated to dtype mismatch during inference." }, { "left": "issue:45325", "right": "issue:45834", "accept": false, - "reason": "Different Qwen2.5-VL/Kosmos2.5 bugs: temporal position IDs vs OCR long-input index error." + "reason": "Qwen2.5-VL position ids and Kosmos2.5 OCR index errors are different bugs." }, { "left": "issue:43454", "right": "issue:44466", "accept": false, - "reason": "Both involve tied LM head weights, but one is a loading/tie failure and the other is inconsistent serialization by device; not clearly the same bug." + "reason": "Missing lm_head tie and inconsistent lm_head.weight serialization are related only broadly, not the same bug." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Unrelated model-specific failures: Qwen3Moe fine-grained FP8 vs Qwen2.5-GGUF loading under transformers v5." + "reason": "Qwen3MoE FP8 failure and Qwen2.5-GGUF loading failure are separate regressions." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "Different regressions: checkpoint key mismatch on resume vs Qwen2.5-VL position-id scaling." + "reason": "Checkpoint key mismatch and Qwen2.5-VL rope-index bug are unrelated." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Related only at a high level of initialization/config strictness; one is `_is_hf_initialized` in `init_empty_weights`, the other is strict config loading for granite_speech." + "reason": "Unexpected _is_hf_initialized argument and strict granite_speech config loading are different compatibility issues." }, { "left": "issue:44479", "right": "issue:45823", "accept": false, - "reason": "Both are model/device regressions, but one is Qwen VL video-input behavior and the other is Gemma4 PLE device mismatch." + "reason": "Qwen video-input regression and Gemma4 device mismatch are unrelated model-specific failures." }, { - "left": "issue:43873", - "right": "issue:45701", + "left": "issue:43296", + "right": "issue:45800", "accept": false, - "reason": "Too broad: quantization offloading behavior vs tokenization changes across transformers versions." + "reason": "PaddleOCR-VL loading in vLLM and torch 2.4.1 vs transformers 5.8.0 incompatibility are not the same bug." }, { - "left": "issue:43296", - "right": "issue:45800", + "left": "issue:43873", + "right": "issue:45701", "accept": false, - "reason": "Different compatibility issues: PaddleOCR-VL load failure in vLLM vs torch 2.4.1/transformers 5.8.0 incompatibility." + "reason": "Quantization offloading behavior and tokenization changes across versions are separate concerns." }, { "left": "issue:43606", "right": "issue:44265", "accept": false, - "reason": "Different bug classes: CPU offload device mismatch for a model vs `torch.export` failures with `torch_compilable_check`." + "reason": "CPU-offload device mismatch for bark-small is unrelated to torch.export and torch_compilable_check." }, { "left": "issue:43262", "right": "issue:43746", "accept": false, - "reason": "Related to loading/processor behavior, but one is audio chat-template sampling rate and the other is PEFT adapter checkpoint loading." + "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are different issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The two accepted Qwen2.5-VL pairs are reasonably supported by the issue titles/body themes, and the rejected soft edges are mostly cautious distinctions between related but non-identical bugs." + "feedback": "The analysis is grounded in the packet and stays conservative. The only accepted duplicate, issue:45325 vs issue:45381, is plausibly the same Qwen2.5-VL position-id/rope regression based on the titles and version context. The remaining soft-edge rejections are generally cautious and do not overclaim duplicate status from thematic similarity alone." }, "evaluator_used": true, "retried": false, @@ -3884,8 +3905,8 @@ "issue:41720|issue:45823", "issue:43606|issue:43819", "issue:43454|issue:43550", - "issue:39692|issue:40444", "issue:43881|issue:44488", + "issue:39692|issue:40444", "issue:46305|issue:46315", "issue:43377|issue:44610", "issue:43454|issue:45072", @@ -3898,29 +3919,29 @@ "issue:43454|issue:43720", "issue:45325|issue:46315", "issue:41720|issue:43322", - "issue:43749|issue:44534", "issue:44991|issue:46291", + "issue:43749|issue:44534", "issue:42617|issue:42915", "issue:41720|issue:44977", "issue:44466|issue:45357", "issue:43526|issue:43883", - "issue:44387|issue:45704", - "issue:43881|issue:44991" + "issue:43881|issue:44991", + "issue:44387|issue:45704" ], - "prepared_review_unit_hash": "032ac03698ecd5919230867907949a0bf0e3b57d83fdbe894be458d1b05b1599", + "prepared_review_unit_hash": "03963af0848c8adeca403a34eefeb2c47278f2b73007e3183c4aab7e226d6654", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36856, + "serialized_chars": 36855, "estimated_input_tokens": 9214, "estimated_eval_tokens": 18684 }, @@ -3929,192 +3950,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These are mostly unrelated, model-specific bug reports rather than true duplicates. A few pairs share broad themes like offloading, loading, or tokenizer regressions, but the concrete failure modes differ. No PRs are present, so the best anchor is a broad issue in the offloading/quantization area.", - "confidence": 0.77, - "canonical_issue_reason": "issue:43873 is the broadest representative of the small offloading/quantization subtheme and is open with strong discussion/references, making it the best cluster anchor.", + "summary": "This cluster is thematically broad but not duplicate-tight: it mixes tokenizer regressions, model-loading/device-mapping failures, quantization/offloading bugs, and a few unrelated model-specific issues. None of the proposed soft edges look like the same underlying bug/change.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43873 is the broadest representative because it centers on offloading plus quantization, which is the nearest shared theme across several neighboring reports. It is still only a cluster representative, not a true duplicate root.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 is the most general and reusable issue in the set; it can plausibly subsume nearby device-map/offload complaints better than the narrowly-scoped model-specific reports.", + "best_issue_reason": "Issue 43873 is the best global issue pick for this cluster because it captures the widest common failure mode (offloading/quantization) without being overly specific to one model.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40990", "right": "issue:45584", "accept": false, - "reason": "Different models and different symptoms: perplexity regression vs Whisper empty-transcription generation failure." + "reason": "Unrelated symptoms: high perplexity on one model vs empty-transcription generation failure on Whisper." }, { "left": "issue:34634", "right": "issue:35707", "accept": false, - "reason": "Unrelated areas: Bark voice preset processing vs progressive generation with inputs_embeds/past_key_values." + "reason": "Different subsystems and failure modes: Bark voice preset vs progressive generation with inputs_embeds/past_key_values." }, { "left": "issue:44568", "right": "issue:45920", "accept": false, - "reason": "Both tokenizer-related, but one is missing BOS/EOS insertion for a specific tokenizer while the other is wrong token IDs across many models; different bugs." + "reason": "Both are tokenizer regressions, but they affect different code paths and model families; not the same concrete bug." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Generation-cache behavior is unrelated to Qwen3 auto device-mapping cuda assertions." + "reason": "Generation API issue vs device-mapping CUDA assert; no shared root cause." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer padding/truncation crash is unrelated to torch.export failures from torch_compilable_check." + "reason": "Tokenizer NER/padding crash and torch.export torch_compilable_check failure are distinct issues." }, { "left": "issue:41720", "right": "issue:45823", "accept": false, - "reason": "Both involve device_map=\"auto\", but the concrete failures are different models and different device-mismatch/assertion paths." + "reason": "Both involve device mapping, but the models and failure mechanisms differ too much to merge." }, { "left": "issue:43606", "right": "issue:43819", "accept": false, - "reason": "Both mention model behavior issues, but CPU offload device mismatch is unrelated to DAC.from_latents / missing STE." + "reason": "CPU offload device mismatch is not the same as DAC missing STE behavior." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "Weight-tying/generation garbage issue is unrelated to a torch.compile SDPA failure." + "reason": "Weight-tying serialization bug and torch.compile SDPA failure are unrelated." }, { - "left": "issue:39692", - "right": "issue:40444", + "left": "issue:43881", + "right": "issue:44488", "accept": false, - "reason": "Documentation example/model-processor mismatch is unrelated to fine-tuning failures with multiple images per prompt." + "reason": "Model load failure and tokenizer load failure are different bugs." }, { - "left": "issue:43881", - "right": "issue:44488", + "left": "issue:39692", + "right": "issue:40444", "accept": false, - "reason": "Two loading failures, but they affect different models and likely different root causes." + "reason": "Documentation example errors and training with multiple images per prompt are not the same underlying bug." }, { "left": "issue:46305", "right": "issue:46315", "accept": false, - "reason": "Sliding-window off-by-one is unrelated to tokenizer conversion performance regression." + "reason": "Off-by-one sliding window masking and slow token conversion are unrelated regressions." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI padding-mask inconsistency is unrelated to OmDet-Turbo processor/model input-size mismatch." + "reason": "MIMI padding-mask correctness and OmDet processor image-size mismatch are unrelated." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "Weight tying / garbage text generation is not the same bug as dtype mismatches in bfloat16 inference." + "reason": "lm_head tying/serialization and dtype mismatch in bfloat16 inference are different code paths." }, { "left": "issue:43257", "right": "issue:45923", "accept": false, - "reason": "MoE weight conversion on accelerate+deepspeed is unrelated to a missing supports_gradient_checkpointing flag." + "reason": "Qwen MOE weight conversion and a missing gradient-checkpointing flag are unrelated." }, { "left": "issue:43572", "right": "issue:45759", "accept": false, - "reason": "Config pad_token_idx regression is unrelated to from_config text_config unwrapping for composite multimodal models." + "reason": "Config field missing in StableLM versus composite Qwen text_config unwrapping; not the same bug." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both are offloading-related, but one is dequantized-model loading on gpu+cpu offload and the other is a broad quantization-offloading complaint; not clearly the same concrete defect." + "reason": "Both are offload/quantization themed, but one is dequantized-model loading on gpu+cpu offload and the other is a broader quantization offloading report." }, { "left": "issue:40990", "right": "issue:42947", "accept": false, - "reason": "Perplexity regression is unrelated to gradient-checkpointing ineffectiveness." + "reason": "High perplexity on a specific model vs gradient checkpointing not taking effect are unrelated." }, { "left": "issue:43377", "right": "issue:44898", "accept": false, - "reason": "Both are vision-model bugs, but one is batching/padding-mask correctness and the other is non-default-resolution interpolation failure." + "reason": "Batched-vs-single MIMI output mismatch and Perceiver non-default-resolution failure are different model bugs." }, { "left": "issue:43322", "right": "issue:46050", "accept": false, - "reason": "Segmentation fault while loading Llava Next is unrelated to quantized model save/load corruption." + "reason": "Llava Next segfault on load and sinq save/load breakage are not the same code-path problem." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "lm_head weight tying/generation issue is unrelated to packed-weight unpacking during accelerate loading." + "reason": "lm_head tie failure and packed-weight unpacking during accelerate loading are different issues." }, { "left": "issue:45325", "right": "issue:46315", "accept": false, - "reason": "Qwen2.5-VL rope/position-id bug is unrelated to tokenizer conversion slowdown." + "reason": "Qwen2.5-VL rope/indexing bug and tokenizer conversion slowdown are unrelated." }, { "left": "issue:41720", "right": "issue:43322", "accept": false, - "reason": "Different model-loading failures with different error modes; no shared concrete bug path." + "reason": "CUDA assert during device-mapped Qwen3 loading and Llava Next segfault are different failures." }, { - "left": "issue:43749", - "right": "issue:44534", + "left": "issue:44991", + "right": "issue:46291", "accept": false, - "reason": "FSDP CPU RAM efficient loading is not the same as non-persistent buffer initialization junk." + "reason": "Tokenizer loading regression and kernels version/revision import error are unrelated." }, { - "left": "issue:44991", - "right": "issue:46291", + "left": "issue:43749", + "right": "issue:44534", "accept": false, - "reason": "Tokenizer loading regression is unrelated to evaluate importing kernels with missing version/revision." + "reason": "FSDP RAM-efficient loading breakage and non-persistent buffer junk initialization are not the same bug." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "3d_parallel.py runtime failure and Qwen3MoE FP8 failure are unrelated." + "reason": "3d_parallel.py execution failure and Qwen3Moe fine-grained FP8 failure are unrelated." }, { "left": "issue:41720", "right": "issue:44977", "accept": false, - "reason": "Both mention Qwen3, but auto device-mapping cuda asserts and flash-attention generation issues are different code paths." + "reason": "Device-mapping CUDA assert and flash-attention generation behavior are distinct model/runtime problems." }, { "left": "issue:44466", "right": "issue:45357", "accept": false, - "reason": "Both touch saving/serialization, but tied lm_head serialization and visual encoder key saving are different defects." + "reason": "Both touch serialization, but one is lm_head.weight inconsistency and the other is visual encoder key saving; not the same fix." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "BEiT label reduction bug is unrelated to Molmo tied-weights attribute error." + "reason": "BEiT label reduction bug and Molmo tied-weights attribute error are unrelated." }, { - "left": "issue:44387", - "right": "issue:45704", + "left": "issue:43881", + "right": "issue:44991", "accept": false, - "reason": "Memory/OOM symptoms are too broad here; int4 reserved-memory growth and apex FusedRMSNorm leak are different causes." + "reason": "Loading a model versus loading a tokenizer; different failure points." }, { - "left": "issue:43881", - "right": "issue:44991", + "left": "issue:44387", + "right": "issue:45704", "accept": false, - "reason": "Model loading failure and tokenizer loading failure are different subsystems and root causes." + "reason": "Quantization memory/OOM regression and apex RMSNorm memory leak are separate memory problems with different causes." } ] }, @@ -4154,16 +4175,17 @@ "issue:43450", "issue:43525", "issue:43531", + "issue:43540", "issue:43550", "issue:43572", "issue:43577", "issue:43606", "issue:43708", "issue:43720", + "issue:43782", "issue:43873", "issue:44008", "issue:44186", - "issue:44361", "issue:44423", "issue:44610", "issue:44671", @@ -4195,8 +4217,8 @@ "issue:36331|issue:42915", "issue:43450|issue:45701", "issue:43577|issue:43708", - "issue:43873|issue:44756", "issue:44008|issue:45823", + "issue:43873|issue:44756", "issue:43550|issue:43720", "issue:42175|issue:46291", "issue:43404|issue:45357", @@ -4216,210 +4238,210 @@ "issue:44671|issue:44743", "issue:45081|issue:45812", "issue:43316|issue:44913", - "issue:44186|issue:44361" + "issue:43540|issue:43782" ], - "prepared_review_unit_hash": "f1f50094f188d24758763b8d32d647a5eeaa1844214a20c4c7c91799c0149891", + "prepared_review_unit_hash": "25216e03998ed2720acd6f50131de9c1a8059263a11a0044b3ebbf19111f5c99", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36201, - "estimated_input_tokens": 9051, - "estimated_eval_tokens": 18358 + "serialized_chars": 36780, + "estimated_input_tokens": 9195, + "estimated_eval_tokens": 18646 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly a set of unrelated Transformers bug reports that only share broad subsystem similarity. None of the soft pairs look like the same underlying bug/change, so all soft edges should be rejected.", - "confidence": 0.96, - "canonical_issue_reason": "issue:43873 is the broadest, highest-signal bug report in the set, with active discussion and inbound references, making it a reasonable representative issue for the cluster.", + "summary": "These items are mostly related by broad transformer-model regression themes, but the soft pairs are not true duplicates: they usually differ in model family, failure mode, or code path. No soft edge should be merged. The broadest representative issue is the generic tokenization-regression report.", + "confidence": 0.64, + "canonical_issue_reason": "issue:45701 is the broadest representative: it describes a version-wide tokenization regression and already has multiple inbound references, making it a better cluster anchor than the more model-specific reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43873 has the strongest combination of impact, specificity, and discussion activity among the issues shown, so it is the best single issue representative.", + "best_issue_reason": "issue:45701 is the most generally applicable issue in this set; its scope is wider than the other model- or feature-specific bugs, so it best fits a canonical representative for the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:40990", "right": "issue:45701", "accept": false, - "reason": "Different problems: perplexity quality regression vs tokenizer/version-dependent tokenization drift; no shared code-path bug." + "reason": "Perplexity on gpt-oss-20b is a concrete evaluation issue; the other is a broad tokenization regression. Different symptom and scope." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Trainer compute_loss signature regression and 3d_parallel.py execution failure are unrelated." + "reason": "CustomTrainer.compute_loss kwarg breakage is a trainer API mismatch; 3d_parallel.py is a separate execution failure." }, { "left": "issue:43065", "right": "issue:46402", "accept": false, - "reason": "Both involve Sam3, but one is a dummy Conv2d in the pixel decoder and the other is missing text-encoder weight loading; different fixes." + "reason": "Both are Sam3-related, but one is a dummy Conv2d in the pixel decoder while the other is missing text encoder weights. Different components and fixes." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for granite_speech is unrelated to tokenizer loading for est-roberta." + "reason": "Different model families and different failure points: strict config loading vs tokenizer loading." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Tokenizer crash on NER/batched padding is unrelated to bfloat16 dtype mismatches in model inference." + "reason": "Tokenizer crashes on NER/padding versus dtype mismatches in inference; unrelated bugs." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "Unexpected trainer keyword argument and Qwen3MoE FP8 failure are different bug classes." + "reason": "Trainer custom loss signature breakage is unrelated to Qwen3Moe FP8 loading failure." }, { "left": "issue:43450", "right": "issue:45701", "accept": false, - "reason": "Video processor batching shape bug is unrelated to tokenization drift." + "reason": "Video processor batch-shape bug is a multimodal preprocessing issue, not the tokenization regression described in 45701." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "Model dtype not propagating and checkpoint resume max_steps calculation are separate trainer/model-loading issues." + "reason": "One is dtype remaining float32 after load; the other is incorrect max_steps on resume. Different trainer/model-state bugs." }, { - "left": "issue:43873", - "right": "issue:44756", + "left": "issue:44008", + "right": "issue:45823", "accept": false, - "reason": "Quantized offloading behavior and Strix Halo mmap OOM are different underlying problems." + "reason": "Gemma 3n variable-name collision in forward is unrelated to Gemma4 device-map placement mismatch." }, { - "left": "issue:44008", - "right": "issue:45823", + "left": "issue:43873", + "right": "issue:44756", "accept": false, - "reason": "Gemma3n variable collision in forward is unrelated to Gemma4 PLE device placement mismatch." + "reason": "Both touch loading/offload behavior, but one is quantization offloading and the other is mmap-induced OOM on Strix Halo. Not the same bug." }, { "left": "issue:43550", "right": "issue:43720", "accept": false, - "reason": "torch.compile SDPA failure in Bamba and packed-weight unpacking in BitNet are distinct code paths." + "reason": "torch.compile with SDPA on Bamba and packed-weight unpacking during accelerate loading are distinct code paths." }, { "left": "issue:42175", "right": "issue:46291", "accept": false, - "reason": "Missing tensorflow backend in torch install and kernels import-version requirement are unrelated packaging issues." + "reason": "Packaging extras missing TensorFlow backend is different from kernels v0.15 requiring explicit version/revision metadata." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Mistral lm_head tying bug and Qwen3.5 save_pretrained key regression affect different models and mechanisms." + "reason": "Weight tying in Mistral3 and incorrect visual encoder keys in Qwen3.5 save_pretrained are different regressions." }, { "left": "issue:41720", "right": "issue:42617", "accept": false, - "reason": "Qwen3 auto device mapping cuda assert and 3d_parallel.py startup failure are not the same bug." + "reason": "cudaErrorAssert with auto device mapping is not the same as failing to run 3d_parallel.py." }, { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Continuous-batching multimodal crash and Gemma4Processor missing _tokenizer are different serve-side failures." + "reason": "Serve continuous-batching multimodal crash and Gemma4Processor missing _tokenizer are related to serving, but they are different concrete failures." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Processor output size mismatch for OmDet-Turbo and Perceiver non-default-resolution failure are similar in theme but not the same concrete bug." + "reason": "Both concern image sizing, but one is OmDet processor output size and the other is Perceiver non-default resolution handling. Different models and fixes." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer NER/batching crash and image processor resolution mismatch are unrelated." + "reason": "Tokenizer/NER padding crash is unrelated to image processor resolution mismatch." }, { "left": "issue:43295", "right": "issue:45701", "accept": false, - "reason": "Processor.tokenizer regression is unrelated to tokenization drift across versions." + "reason": "Processor.tokenizer regression in v4.57.5 is not the same as a generic tokenization-change report, even if both mention tokenization." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "Both involve loading/quantization, but one is FineGrainedFP8 failure on Qwen3Moe and the other is device_map auto offload for dequantized models; not the same bug." + "reason": "Qwen3Moe FP8 failure and dequantized model offload failure are different loading problems." }, { "left": "issue:45632", "right": "issue:45684", "accept": false, - "reason": "Cache-path collision and read-only permission propagation are separate save/load issues." + "reason": "Cache-path collision for local model leaf directories is unrelated to save_pretrained permission propagation." }, { "left": "issue:43425", "right": "issue:46291", "accept": false, - "reason": "Torch 2.10 incompatibility and kernels import version/revision requirement are unrelated dependency problems." + "reason": "Torch 2.10 incompatibility is a version-compatibility issue, not the kernels dependency-revision error." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import error and 3d_parallel.py execution failure are different issues." + "reason": "GenerationMixin import error is unrelated to the 3d_parallel.py runtime failure." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Qwen3 auto device mapping assert and GPT-OSS tensor-parallel-only issue are separate model-parallelism bugs." + "reason": "Both involve GPT-style model execution, but one is an auto device-map CUDA assert and the other is a tensor-parallel-only GPT-OSS MoE issue." }, { "left": "issue:43525", "right": "issue:46291", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is unrelated to kernels import failure." + "reason": "Missing pad_token_id in Llama4Config is a model config bug; kernels version/revision requirements are a dependency-loading issue." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig missing pad_token_idx are different config regressions." + "reason": "These are different config-field regressions in different model families: sliding_window vs pad_token_idx." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked-LM prediction regression and Qwen3.5 recurrent-state reset are unrelated model-specific bugs." + "reason": "CamemBERT masked-LM prediction errors and recurrent-state reset with cache are distinct behavior bugs." }, { "left": "issue:45081", "right": "issue:45812", "accept": false, - "reason": "Mistral regex patch crash and Granite token ID regression are both tokenizer-related but not the same underlying defect." + "reason": "Mistral regex patch crash and Granite token-ID regression are both tokenizer-related but not the same underlying issue." }, { "left": "issue:43316", "right": "issue:44913", "accept": false, - "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct reload bug are distinct config issues." + "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct not persisting are separate config consistency bugs." }, { - "left": "issue:44186", - "right": "issue:44361", + "left": "issue:43540", + "right": "issue:43782", "accept": false, - "reason": "LayoutLMv2Tokenizer NER crash and MLukeTokenizer AttributeError are different tokenizer bugs affecting different model families." + "reason": "Qwen3OmniMoe video-input ValueError and Qwen3VL weight_only load error are different multimodal loading/processing failures." } ] }, @@ -4460,15 +4482,15 @@ "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:44387", "issue:44451", "issue:44492", @@ -4492,21 +4514,21 @@ "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:43388|issue:45059", "issue:43550|issue:43653", "issue:43606|issue:43720", "issue:44530|issue:46291", "issue:43873|issue:45704", - "issue:41720|issue:43746", "issue:45406|issue:45464", + "issue:41720|issue:43746", "issue:44387|issue:45701", "issue:45357|issue:46402", "issue:43606|issue:43653", @@ -4520,202 +4542,202 @@ "issue:43531|issue:44589", "issue:44514|issue:44545" ], - "prepared_review_unit_hash": "fa607b1b10a7e23ba6f7e3b0d34a910e12587426115453d34c8e422a3736fb55", + "prepared_review_unit_hash": "7a74483ff537125371ed7dc015b453c8f4ac9f5b2a95f8c3d5788d3775fbee41", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35379, - "estimated_input_tokens": 8845, - "estimated_eval_tokens": 17946 + "serialized_chars": 35386, + "estimated_input_tokens": 8847, + "estimated_eval_tokens": 17950 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Almost all items are unrelated, model-specific bugs or regressions. The only clear duplicate pair is the Qwen2_5_VLProcessor apply_chat_template batched-input crash; the rest should stay separate.", - "confidence": 0.95, - "canonical_issue_reason": "issue:44514 is the earlier, more canonical report for the exact same Qwen2_5_VLProcessor batched-input crash when padding=False.", + "summary": "This cluster is mostly a loose mix of unrelated bug reports connected only by weak textual similarity. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template crash report; all other soft edges should stay separate.", + "confidence": 0.97, + "canonical_issue_reason": "issue:44514 is the best canonical issue because it is the earlier report of the exact same Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash described in issue:44545.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44514 is the best issue to keep as the representative duplicate target: precise title, earlier filing, and a direct same-bug counterpart.", + "best_issue_reason": "issue:44514 is the strongest global issue representative since it is the only true duplicate-match in the cluster and has the clearest, most specific reproduction target.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43540", - "right": "issue:43782", + "left": "issue:44186", + "right": "issue:44361", "accept": false, - "reason": "Different model paths and failure modes: video input processing vs from_pretrained weight_only loading error." + "reason": "Different tokenizers, different failure modes, and no shared code path beyond generic tokenizer handling." }, { "left": "issue:42915", "right": "issue:43322", "accept": false, - "reason": "Unrelated bugs: FineGrainedFP8Config failure for Qwen3Moe vs Llava Next loading segfault." + "reason": "Different model families and errors; one is FP8 config loading, the other is a Llava Next segfault." }, { "left": "issue:43421", "right": "issue:45812", "accept": false, - "reason": "One is runtime post-processor sync in TokenizersBackend; the other is wrong token IDs for Granite models." + "reason": "Both are tokenizer-related, but one is runtime post-processor refresh and the other is a Granite token-ID regression; not the same bug." }, { "left": "issue:36331", "right": "issue:41720", "accept": false, - "reason": "Trainer API signature mismatch is unrelated to a Qwen3 auto device-mapping cudaErrorAssert." + "reason": "Completely different subsystems: trainer API signature vs device-mapping CUDA assert." }, { "left": "issue:43526", "right": "issue:43881", "accept": false, - "reason": "Different subsystems and symptoms: BeitImageProcessorFast label reduction vs glm-4v-9b loading failure." + "reason": "Unrelated: Beit image-label reduction bug versus a glm-4v loading failure." }, { "left": "issue:29159", "right": "issue:36032", "accept": false, - "reason": "General tokenizer inconsistency is not the same as a T5Tokenizer load-time add_special_tokens conflict." + "reason": "Both mention tokenizers, but the bugs are distinct: slow/fast inconsistency vs T5 method-name conflict." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS bus error in CLIP loading is distinct from progressive generation with inputs_embeds/past_key_values." + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." }, { - "left": "issue:43388", - "right": "issue:45059", + "left": "issue:30064", + "right": "issue:34634", "accept": false, - "reason": "Metric gathering bug is unrelated to SAM3 PCS behavior with text and bounding boxes." + "reason": "Image processor void-segmentation-map handling is unrelated to BarkProcessor voice_preset." }, { - "left": "issue:30064", - "right": "issue:34634", + "left": "issue:43388", + "right": "issue:45059", "accept": false, - "reason": "Void segmentation map processing and BarkProcessor voice_preset are different processor bugs." + "reason": "Different components and symptoms: metric gathering batch trimming vs SAM3 text/bbox behavior." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "torch.compile/SDPA failure is not the same as BigBirdTokenizer special-token registration." + "reason": "Different bugs in different areas: torch.compile/SDPA versus tokenizer special-token registration." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "CPU offload device mismatch and BitNet packed-weight unpacking are separate accelerate/loading issues." + "reason": "Both involve loading/offload-ish paths, but the concrete problems and models are different." }, { "left": "issue:44530", "right": "issue:46291", "accept": false, - "reason": "PagedAttentionCache linear_attention crash is unrelated to kernels v0.15 import-version requirements." + "reason": "PagedAttentionCache Qwen3.5 crash and kernels dependency import failure are unrelated." }, { "left": "issue:43873", "right": "issue:45704", "accept": false, - "reason": "Quantization offloading behavior is not the same as T5's apex.FusedRMSNorm memory leak." + "reason": "Quantization offload behavior and T5 apex RMSNorm memory leak are distinct issues." }, { - "left": "issue:41720", - "right": "issue:43746", + "left": "issue:45406", + "right": "issue:45464", "accept": false, - "reason": "Qwen3 auto device-mapping cuda assert and GraniteSpeech PEFT checkpoint loading are unrelated." + "reason": "Different crashes in different APIs/models: Gemma4Processor missing _tokenizer versus Qwen3.5 streaming inference failure." }, { - "left": "issue:45406", - "right": "issue:45464", + "left": "issue:41720", + "right": "issue:43746", "accept": false, - "reason": "Gemma4Processor missing _tokenizer and Qwen3.5 streaming chat/completions failure are different bugs." + "reason": "Different model/load paths and different errors; no evidence of the same underlying bug." }, { "left": "issue:44387", "right": "issue:45701", "accept": false, - "reason": "Int4 quantization OOM is not the same as a general version-dependent tokenization change." + "reason": "CUDA memory regression under int4 quantization is unrelated to a tokenization-version-change report." }, { "left": "issue:45357", "right": "issue:46402", "accept": false, - "reason": "Incorrect save_pretrained visual encoder keys differs from missing text encoder weights on load." + "reason": "Both concern multimodal config/weight handling, but one is incorrect visual encoder keys and the other is missing text encoder weights; not the same bug." }, { "left": "issue:43606", "right": "issue:43653", "accept": false, - "reason": "CPU offload device mismatch is unrelated to BigBirdTokenizer decode output from mask-token registration." + "reason": "One is a CPU-offload device mismatch during model loading; the other is a tokenizer special-token decode issue." }, { "left": "issue:43531", "right": "issue:45759", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is different from unwrapping text_config in composite multimodal configs." + "reason": "Both touch Qwen configs, but they address different problems: sliding_window behavior versus unwrapping text_config in from_config." }, { "left": "issue:44492", "right": "issue:44509", "accept": false, - "reason": "A typo in cache strategies is unrelated to docs still mentioning removed pipeline tasks." + "reason": "Pure documentation/typo cleanup items, but they do not describe the same change or bug." }, { "left": "issue:34634", "right": "issue:41720", "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping are different code paths." + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping failures are unrelated." }, { "left": "issue:42490", "right": "issue:43329", "accept": false, - "reason": "Trailing-slash import warning is unrelated to the multimodal token-count video-branch undefined-variable bug." + "reason": "Local-path import warning and multimodal token-counting undefined names are unrelated." }, { "left": "issue:34567", "right": "issue:35707", "accept": false, - "reason": "TrainerState token counter update issue is unrelated to progressive generation with inputs_embeds/past_key_values." + "reason": "TrainerState token-count tracking and progressive generation with embeddings/past KV are different code paths." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Cannot load one specific model in v5 is not the same as tokenizer loading failure for EMBEDDIA/est-roberta." + "reason": "Both are model-loading regressions, but they affect different models and likely different tokenizer configs." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Qwen3.5 num_labels propagation is unrelated to the single-label classification zero-loss regression." + "reason": "Both involve classification configuration, but one is Qwen3.5 num_labels propagation and the other is a generic zero-loss problem." }, { "left": "issue:43531", "right": "issue:44589", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and Float8 storage lookup failure are unrelated." + "reason": "Sliding-window Qwen3-MoE issue is unrelated to a Float8 storage lookup TypeError." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "Same processor, same crash, same condition: batched apply_chat_template with padding=False. These are clear duplicates." + "reason": "Exact same Qwen2_5_VLProcessor.apply_chat_template batched-padding=False crash; clearly the same underlying bug and mergeable as one fix." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only clearly duplicate pair in the provided data is issue:44514/issue:44545, which have identical titles, while the rest are plausibly distinct by subsystem and symptom. No soft-edge verdict overstates the evidence." + "feedback": "Grounded overall: the cluster summary stays conservative, and the one accepted pair (issue:44514 / issue:44545) is clearly the same reported crash by title. The other soft-edge rejections are suitably cautious given the packet." }, "evaluator_used": true, "retried": false, @@ -4794,244 +4816,241 @@ "issue:43421|issue:43927", "issue:44987|issue:45042", "issue:44368|issue:45357", - "issue:44918|issue:44936", "issue:39692|issue:44610", + "issue:44918|issue:44936", "issue:43782|issue:45381", "issue:43746|issue:46050", "issue:42886|issue:44843", - "issue:45704|issue:45923", "issue:43421|issue:43653", + "issue:45704|issue:45923", "issue:43381|issue:43844", "issue:45704|issue:45859", "issue:45663|issue:45923", "issue:41720|issue:45704", "issue:43531|issue:43645", - "issue:43421|issue:44186", "issue:43994|issue:44671", + "issue:43421|issue:44186", "issue:43746|issue:43873", - "issue:39692|issue:43873", "issue:45081|issue:45356", + "issue:39692|issue:43873", "issue:43596|issue:44220", - "issue:42915|issue:43746", "issue:43540|issue:46294", + "issue:42915|issue:43746", "issue:45356|issue:45907", "issue:30064|issue:36010" ], - "prepared_review_unit_hash": "8294f26ddb9afcdedd7822c8de18423e0ba444e81d029154cc84c1fb37700000", + "prepared_review_unit_hash": "a3cffb914d346bc00d2b68e6ed2fea068d4501d617417e189029c5de56575fb2", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 37073, - "estimated_input_tokens": 9269, - "estimated_eval_tokens": 18794 + "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 cluster is highly heterogeneous; the soft pairs share superficial vocabulary but do not look like the same underlying bug/change. I would not merge any of the candidate pairs as duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "No clear canonical issue: the items span unrelated tokenizer, vision, training, quantization, and config bugs, so none is a strong duplicate anchor.", + "summary": "This is mostly a heterogeneous cluster of unrelated bug reports across tokenizers, vision, multimodal, training, and config/loading paths. Only one pair looks like a plausible duplicate around offline tokenizer loading; the rest appear distinct.", + "confidence": 0.78, + "canonical_issue_reason": "No single issue is a clean canonical for the whole cluster; the items span unrelated bugs, with only the offline-tokenizer pair looking duplicate-like.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good global representative for deduplication because the cluster does not center on one defect family.", + "best_issue_reason": "No globally representative issue stands out for the full cluster. If anything, 44843 is the closest anchor for the one clear duplicate-like pair, but it does not cover the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44898", "right": "issue:45072", "accept": false, - "reason": "Different subsystems and failures: Perceiver resolution handling vs dtype mismatches in SwitchTransformers/TimmWrapperModel." + "reason": "Different models and failure modes; image resolution interpolation vs bfloat16 dtype mismatch." }, { "left": "issue:44077", "right": "issue:44625", "accept": false, - "reason": "Unrelated problems: PatchTSTMixer post_init policy vs Qwen3.5 num_labels propagation." + "reason": "Unrelated bugs: optional post_init enforcement vs num_labels propagation in Qwen3.5." }, { "left": "issue:45759", "right": "issue:45859", "accept": false, - "reason": "Different code paths: composite config unwrapping vs CPU-offload incompatibility in rotary embedding." + "reason": "Different code paths: composite config unwrapping vs CPU offload incompatibility." }, { "left": "issue:45588", "right": "issue:45759", "accept": false, - "reason": "Flash-attention sinkless-model crash is unrelated to Qwen multimodal config unwrapping." + "reason": "Flash-attention s_aux crash is unrelated to the Qwen config unwrapping issue." }, { "left": "issue:44062", "right": "issue:44451", "accept": false, - "reason": "Both tokenizer-related, but one is an AddedToken argument bug and the other is a model loading failure." + "reason": "Both tokenizer-related, but one is AddedToken argument handling and the other is a model-specific load failure." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Special-token post-processor updates and DiaConfig save/load token IDs are distinct bugs." + "reason": "Both touch special tokens, but one is runtime post-processor updates and the other is save/load losing custom IDs." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Model loading regression vs image processor torchvision dependency; no shared underlying defect." + "reason": "Different regressions: model loading failure vs image processor requiring torchvision." }, { "left": "issue:44368", "right": "issue:45357", "accept": false, - "reason": "Warning about tie_word_embeddings is unrelated to incorrect visual encoder keys in save_pretrained." + "reason": "Same model family, but warning noise vs incorrect saved visual encoder keys are different bugs." }, { - "left": "issue:44918", - "right": "issue:44936", + "left": "issue:39692", + "right": "issue:44610", "accept": false, - "reason": "Trainer failures, but one is SFT embedding unpacking and the other is evaluate-after-train state handling." + "reason": "Different models and symptoms; docs/example mismatch vs processor/model input-size mismatch." }, { - "left": "issue:39692", - "right": "issue:44610", + "left": "issue:44918", + "right": "issue:44936", "accept": false, - "reason": "Both involve vision models, but one is a doc example/model mismatch and the other is processor output size vs model expectation." + "reason": "Trainer SFT embedding unpacking and train/evaluate lifecycle failure are distinct issues." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Different multimodal bugs: weight_only loading error vs wrong video vision_position_ids." + "reason": "Different Qwen VL problems: weight_only load failure vs video position-id bug." }, { "left": "issue:43746", "right": "issue:46050", "accept": false, - "reason": "PEFT local checkpoint loading is not the same as quantized sinq save/load breakage." + "reason": "PEFT checkpoint loading and quantized save/load corruption are not the same bug." }, { "left": "issue:42886", "right": "issue:44843", - "accept": false, - "reason": "Related offline symptoms, but one is generic cache loading and the other is a specific unconditional model_info() call in _patch_mistral_regex." + "accept": true, + "reason": "Both report offline tokenizer loading failing from cache; 44843 pinpoints the likely root cause in _patch_mistral_regex/model_info()." }, { "left": "issue:45704", "right": "issue:45923", "accept": false, - "reason": "RMSNorm memory leak and missing supports_gradient_checkpointing are unrelated." - }, - { - "left": "issue:43421", - "right": "issue:43653", - "accept": false, - "reason": "Both token-related, but one is runtime special-token post-processing and the other is BigBirdTokenizer mask-token registration." + "reason": "T5 RMSNorm memory leak and missing gradient-checkpointing flag are unrelated." }, { "left": "issue:43381", "right": "issue:43844", "accept": false, - "reason": "Eval-mode gradient checkpointing restriction is unrelated to ZeRO-3 gradient inflation." + "reason": "Eval-mode checkpointing restriction is not the same as ZeRO-3 gradient explosion." }, { "left": "issue:45704", "right": "issue:45859", "accept": false, - "reason": "Apex RMSNorm leak has nothing to do with Qwen3_5Moe rotary embedding offload compatibility." + "reason": "T5 memory leak and Qwen rotary offload incompatibility are separate problems." }, { "left": "issue:45663", "right": "issue:45923", "accept": false, - "reason": "FSDP2 shared_kv_states KeyError is unrelated to a missing gradient-checkpointing flag." + "reason": "FSDP2 shared_kv_states KeyError is unrelated to a missing supports_gradient_checkpointing flag." }, { "left": "issue:41720", "right": "issue:45704", "accept": false, - "reason": "Qwen3 auto device-map CUDA assert is unrelated to T5 RMSNorm memory leak." + "reason": "Qwen3 device-map assert and T5 norm leak have different root causes." }, { "left": "issue:43531", "right": "issue:43645", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is unrelated to Jupyter custom-model initialization in Transformers 5.0.0." + "reason": "Qwen3-MoE sliding window bug and Jupyter custom-model initialization breakage are unrelated." }, { - "left": "issue:43421", - "right": "issue:44186", + "left": "issue:43994", + "right": "issue:44671", "accept": false, - "reason": "Different tokenizer bugs: runtime special-token updates vs LayoutLMv2 NER/padding crashes." + "reason": "Different models and failure types; bad SigLIP2 outputs vs CamemBERT masked-LM regression." }, { - "left": "issue:43994", - "right": "issue:44671", + "left": "issue:43421", + "right": "issue:44186", "accept": false, - "reason": "Incorrect SigLIP2 outputs and CamemBERT masked-LM predictions are separate model-specific regressions." + "reason": "Runtime special-token updates are unrelated to LayoutLMv2 NER/padding crashes." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "Local checkpoint loading with PEFT and quantization/offloading behavior are different failures." + "reason": "PEFT checkpoint loading and quantization offloading do not appear to share a code-path bug." }, { - "left": "issue:39692", - "right": "issue:43873", + "left": "issue:45081", + "right": "issue:45356", "accept": false, - "reason": "A broken documentation example is not the same bug as quantization offloading behavior." + "reason": "Both involve tokenizer patching, but one is a backend_tokenizer AttributeError and the other is a codec-handling regression." }, { - "left": "issue:45081", - "right": "issue:45356", + "left": "issue:39692", + "right": "issue:43873", "accept": false, - "reason": "Both touch fix_mistral_regex, but one crashes on backend_tokenizer access while the other is a codec/regression warning issue." + "reason": "SigLIP2 docs/example mismatch is unrelated to offloading not working with quantization." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Zero-3/BertModel indexing error is unrelated to _torch_extract_fbank_features()." + "reason": "Zero3/BertModel init error and audio fbank feature extraction are unrelated." }, { - "left": "issue:42915", - "right": "issue:43746", + "left": "issue:43540", + "right": "issue:46294", "accept": false, - "reason": "Qwen3Moe FineGrainedFP8Config failure is unrelated to GraniteSpeech PEFT checkpoint loading." + "reason": "Both are video-related, but they describe different Qwen models and different failure points." }, { - "left": "issue:43540", - "right": "issue:46294", + "left": "issue:42915", + "right": "issue:43746", "accept": false, - "reason": "Both are video-input issues, but one is Qwen3OmniMoe processing ValueError and the other is chat-template token/feature mismatch in Qwen3.5-9B." + "reason": "Qwen3Moe FP8 failure is unrelated to GraniteSpeech local checkpoint loading." }, { "left": "issue:45356", "right": "issue:45907", "accept": false, - "reason": "Kimi tokenizer codec/fix_mistral_regex regression is unrelated to list_repo_files revision forwarding." + "reason": "Tokenizer regression and list_repo_files revision forwarding are unrelated." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Image processor void-segmentation handling is unrelated to the GenerationMixin import error." + "reason": "Image segmentation processing bug and GenerationMixin import error are unrelated." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative. The only accepted soft edge, 42886\u219444843, is supported by the shared offline tokenizer/cache failure and the explicit _patch_mistral_regex/model_info() root cause in 44843. The rest of the pairwise rejections are appropriately cautious and do not overstate duplication." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -5114,9 +5133,9 @@ "issue:43322|issue:45685", "issue:43716|issue:44008", "issue:43643|issue:44877", - "issue:43388|issue:46032", - "issue:43335|issue:43756", "issue:44843|issue:45081", + "issue:43335|issue:43756", + "issue:43388|issue:46032", "issue:30064|issue:36331", "issue:43950|issue:45987", "issue:41720|issue:44991", @@ -5128,208 +5147,208 @@ "issue:43540|issue:45381", "issue:45584|issue:45834" ], - "prepared_review_unit_hash": "0240b3b582ff06b5fde8afd5b47762ff8e853c79cef0767d0ddae22535baec22", + "prepared_review_unit_hash": "83ac19c87d1772fed70b3238c5119c803fc50da6020c237ec1cb3d771a1ee9ac", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36826, - "estimated_input_tokens": 9207, - "estimated_eval_tokens": 18670 + "serialized_chars": 36833, + "estimated_input_tokens": 9209, + "estimated_eval_tokens": 18674 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: most items are independent bugs/features across unrelated models and APIs, with no true duplicate group. The closest thematic overlap is a few Mistral/tokenizer-related reports, but even those describe different failure modes.", - "confidence": 0.88, - "canonical_issue_reason": "Issue #44843 is the best single anchor only because it is a clear, high-impact core API regression in tokenizer loading; however, the cluster overall is not a duplicate set and most items do not belong under one canonical issue.", + "summary": "This cluster is heterogeneous: it spans unrelated tokenizer, vision, training, cache, and model-loading regressions. None of the soft-edge pairs look like true duplicates of the same underlying bug/change, so this should not be deduped as one cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No clear canonical issue: the items describe many different failure modes across unrelated subsystems, so any single issue would be a poor representative.", "canonical_pr_reason": null, - "best_issue_reason": "Issue #44843 has the strongest standalone triage value: it is reproducible, affects offline mode, and sits in a widely used code path (`AutoTokenizer.from_pretrained`).", + "best_issue_reason": "No single issue is suitable as a global representative because the cluster is not centered on one concrete bug or code path.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "Different models and different failure modes: accelerate loading of BitNet packed weights vs DAC latent reconstruction/STE mismatch." + "reason": "Both are numerical/model bugs, but one is BitNet accelerate-loading and the other is DAC latent decoding; different code paths and fixes." }, { "left": "issue:43653", "right": "issue:43720", "accept": false, - "reason": "Tokenizer special-token registration/decoding bug vs BitNet accelerate-loading failure; unrelated code paths." + "reason": "Tokenizer special-token decoding bug vs BitNet packed-weight loading bug; unrelated subsystems." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio feature extraction issue vs tokenizer-mapping/model-registration bug; no shared underlying defect." + "reason": "Audio fbank feature extraction issue vs tokenizer mapping failure; no shared underlying defect." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "Both are model/runtime regressions, but one is torch.compile+SDPA on Bamba and the other is dtype mismatch on SwitchTransformers/TimmWrapperModel." + "reason": "Both mention dtype/SDPA-like behavior, but one is Bamba torch.compile with SDPA and the other is generic inference dtype mismatch; not the same bug." }, { "left": "issue:43295", "right": "issue:45834", "accept": false, - "reason": "Processor/tokenizer regression in one model family vs OCR index error in Kosmos2.5; different root causes." + "reason": "Processor/tokenizer regression vs long OCR input index error; different failure mechanisms." }, { "left": "issue:43540", "right": "issue:44008", "accept": false, - "reason": "Both are multimodal model bugs, but they affect different models and different code paths (video processing vs forward variable collision)." + "reason": "Qwen3OmniMoe video-input processing error vs Gemma3n variable-name collision in forward(); unrelated code paths." }, { "left": "issue:43618", "right": "issue:43697", "accept": false, - "reason": "CLIP output field regression vs RTDetrV2 output drift in v5; not the same bug." + "reason": "CLIP attentions field regression vs RTDetrV2 output drift in v5; different models and symptoms." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken constructor argument collision is not the same as a tokenizer-loading compatibility failure for a specific model." + "reason": "AddedToken keyword conflict vs failure to load a specific tokenizer model; not the same defect." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Distinct model families and distinct defects; no plausible single fix would cover both." + "reason": "Different model families and bugs: Bamba SDPA/compile issue vs DAC missing-STE mismatch." }, { "left": "issue:45183", "right": "issue:45479", "accept": false, - "reason": "Input validation error messaging vs zero-loss classification bug; different logic and symptoms." + "reason": "General input validation error messaging vs classification-loss degeneracy; only loosely related by model validation." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "Tokenizer batching/NER crash vs Perceiver positional encoding/interpolation failure; unrelated." + "reason": "LayoutLMv2 tokenizer crash vs Perceiver image classification resolution bug; distinct components and fixes." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration bug vs DAC latent mismatch; no common underlying issue." + "reason": "BigBird mask-token registration issue vs DAC latent/forward mismatch; no shared root cause." }, { "left": "issue:43531", "right": "issue:43643", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE vs config-field loss in trust_remote_code; unrelated." + "reason": "Qwen3-MoE sliding-window behavior vs AutoConfig trust_remote_code missing fields; unrelated areas." }, { "left": "issue:43322", "right": "issue:45685", "accept": false, - "reason": "Segfault while loading Llava Next vs MPS histogram error in MoE; different subsystems and causes." + "reason": "Llava Next loading segfault vs MPS histogram op not implemented for Int; completely different failure modes." }, { "left": "issue:43716", "right": "issue:44008", "accept": false, - "reason": "Both touch multimodal inputs, but one is a dtype mismatch and the other is a variable-name collision in forward()." + "reason": "Mistral image-preprocessor dtype mismatch vs Gemma3n audio mask attribute collision; unrelated issues." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "Both concern config loading, but one is missing fields with trust_remote_code and the other is strict config rejection; not the same defect." + "reason": "Both are config-loading regressions, but one is missing fields from trust_remote_code and the other is strict config blocking granite_speech; different causes." }, { - "left": "issue:43388", - "right": "issue:46032", + "left": "issue:44843", + "right": "issue:45081", "accept": false, - "reason": "Metric gathering truncation bug vs Mamba2 cache correctness bug; completely different code paths." + "reason": "Offline model_info call in tokenizer regex patch vs backend_tokenizer attribute crash in the same patch family, but the described failures are different bug manifestations and fixes." }, { "left": "issue:43335", "right": "issue:43756", "accept": false, - "reason": "Both are model configuration issues, but they involve different architectures and different misconfigurations." + "reason": "SwitchTransformers sparse-layer creation bug vs Smollm3 RoPE-layer mismatch; not the same model logic." }, { - "left": "issue:44843", - "right": "issue:45081", + "left": "issue:43388", + "right": "issue:46032", "accept": false, - "reason": "Both are in `_patch_mistral_regex`, but they are different bugs: offline model_info call vs backend_tokenizer attribute crash." + "reason": "Metric-gathering label truncation vs Mamba2 use_cache correctness; different runtime paths and symptoms." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image processor segmentation-map failure vs custom trainer loss signature mismatch; unrelated." + "reason": "Image processor segmentation-map handling vs CustomTrainer compute_loss signature mismatch; no overlap in underlying bug." }, { "left": "issue:43950", "right": "issue:45987", "accept": false, - "reason": "One is non-persistent buffer corruption on load; the other is a cache-length return-type regression. Different mechanisms." + "reason": "from_pretrained buffer corruption vs StaticCache get_seq_length return-type bug; different APIs and fix scopes." }, { "left": "issue:41720", "right": "issue:44991", "accept": false, - "reason": "A800 cudaErrorAssert in Qwen3 auto device mapping vs tokenizer loading failure for EMBEDDIA/est-roberta; unrelated." + "reason": "Qwen3 auto device-mapping cuda assert vs tokenizer load failure for a specific model; unrelated." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Both are v5 output regressions, but on different models with different symptoms and likely different fixes." + "reason": "Different-output regression for Pegasus vs incorrect masked LM predictions for CamemBERT; separate model-specific behaviors." }, { "left": "issue:44821", "right": "issue:45084", "accept": false, - "reason": "AutoImageProcessor URL loading vs Jinja/template compilation error; unrelated." + "reason": "Loading AutoImageProcessor from URL vs template compilation error; different subsystems." }, { "left": "issue:43421", "right": "issue:44811", "accept": false, - "reason": "Special-token/post-processor update bug vs Whisper batch_decode skip_special_tokens handling; different APIs and behavior." + "reason": "Runtime special-token/post-processor update feature vs Whisper batch_decode skip_special_tokens bug; not the same change." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "BLIP2 dtype propagation bug vs Video-LLaVA temporal-attention/weight-sharing issue; unrelated." + "reason": "Blip2 dtype propagation issue vs Video-LLaVA temporal-attention/weight-sharing bug; distinct models and defects." }, { "left": "issue:39692", "right": "issue:45834", "accept": false, - "reason": "SigLIP2 doc example errors vs Kosmos2.5 OCR long-input index error; no shared underlying bug." + "reason": "SigLIP2 doc/example errors vs Kosmos2.5 long OCR index error; only broadly vision/OCR-related, not duplicate." }, { "left": "issue:43540", "right": "issue:45381", "accept": false, - "reason": "Qwen3OmniMoe video-input processing failure vs Qwen2.5-VL vision_position_ids bug; different model families and defects." + "reason": "Qwen3OmniMoe video-processing ValueError vs qwen2.5-vl video vision_position_ids bug; same general family but different concrete failure points and fixes." }, { "left": "issue:45584", "right": "issue:45834", "accept": false, - "reason": "Whisper empty-transcription generation regression vs Kosmos2.5 long OCR index error; unrelated." + "reason": "Whisper empty transcription after token alignment vs Kosmos2.5 long OCR input index error; unrelated." } ] }, @@ -5432,208 +5451,208 @@ "issue:45003|issue:46285", "issue:43653|issue:44912" ], - "prepared_review_unit_hash": "46bf2703470dee32c04820b5f0537e6ed8ebdf77737f36079a8d03b291c3637e", + "prepared_review_unit_hash": "e4015acdf0211e2711709779979662ed417a05000e73072e40d3a9b92abf667d", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36282, - "estimated_input_tokens": 9071, - "estimated_eval_tokens": 18398 + "serialized_chars": 36285, + "estimated_input_tokens": 9072, + "estimated_eval_tokens": 18400 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated bug reports that happen to share broad tokenizer/model-loading vocabulary. I found one likely duplicate pair around added-token spacing vs slow/fast tokenizer inconsistency; the rest look like distinct bugs in different models, loaders, or subsystems.", + "summary": "The cluster is mostly heterogeneous tokenizer, model-loading, multimodal, and docs bugs. Only one soft edge looks like a plausible duplicate pair; the rest are too broad or clearly different code paths.", "confidence": 0.71, - "canonical_issue_reason": "issue:29159 is the broadest representative of the only likely duplicate theme here: tokenizer behavior inconsistency, especially around added tokens and spacing differences.", + "canonical_issue_reason": "issue:28218 is the clearest concrete symptom report in the only plausible duplicate pair; issue:29159 is broader and reads like a more general description of the same tokenizer inconsistency.", "canonical_pr_reason": null, - "best_issue_reason": "issue:29159 is the clearest and most general issue to represent the tokenizer added-token inconsistency thread; it subsumes the narrower spacing symptom in issue:28218.", + "best_issue_reason": "issue:28218 is the best representative issue here because it is open, specific, and user-facing, while the rest of the cluster does not form a coherent duplicate set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "Different tokenizer failures: NER/batched padding crash vs loading FastSpeech2ConformerTokenizer." + "reason": "Different bugs: LayoutLMv2 tokenizer NER/padding crash vs AutoTokenizer failing to load a specific tokenizer." }, { "left": "issue:43782", "right": "issue:44977", "accept": false, - "reason": "Different models and failure modes: weight_only load error vs flash-attention generation regression." + "reason": "Different model/runtime failures: weight_only loading error vs flash-attention generation issue." }, { "left": "issue:43644", "right": "issue:43749", "accept": false, - "reason": "Both are loading-related, but one is non-persistent buffer corruption and the other is FSDP CPU RAM efficient loading; not the same bug." + "reason": "Both involve loading, but one is non-persistent buffers after v5.0 and the other is FSDP CPU RAM efficient loading; not the same code-path." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "Different code paths and symptoms: CPU offload device mismatch vs bfloat16 dtype mismatch." + "reason": "Both are CI/model inference mismatches, but they affect different models and different failure modes (CPU offload device mismatch vs bf16 dtype mismatch)." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "One is a v5 pipeline error-message issue; the other is a model output correctness bug in AutoModel/pipeline." + "reason": "One is an error-message/docs issue about removed translation pipelines; the other is a nonsensical output bug for SigLIP2." }, { "left": "issue:44479", "right": "issue:45834", "accept": false, - "reason": "Different regressions in different models: video input handling vs long OCR input indexing." + "reason": "Both are multimodal regressions, but they are distinct models and distinct input-processing failures." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "Unrelated issues: BarkProcessor voice_preset vs GenerationMixin import failure." + "reason": "Unrelated: BarkProcessor voice_preset bug vs GenerationMixin import failure." }, { "left": "issue:44805", "right": "issue:44936", "accept": false, - "reason": "Different failures: mask/index shape mismatch vs trainer.evaluate() after train." + "reason": "Different failures: mask shape IndexError vs trainer.evaluate() breaking after train()." }, { "left": "issue:43577", "right": "issue:43582", "accept": false, - "reason": "Different problems: dtype propagation in BLIP2 vs Apple Silicon allocator warmup TypeError." + "reason": "Different issues: BLIP2 dtype not propagating vs Apple Silicon warmup TypeError." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "Related to v5 discussion, but one is a RoPE implementation question and the other is stale docs about removed pipeline tasks." + "reason": "GLM-5 RoPE implementation discussion vs docs mentioning removed pipeline tasks." }, { "left": "issue:28218", "right": "issue:29159", "accept": true, - "reason": "Likely the same tokenizer bug: added-token spacing/slow-fast tokenizer inconsistency around token insertion behavior." + "reason": "Both point at tokenizer behavior inconsistencies around token insertion/spacing; the broader slow-vs-fast inconsistency report plausibly covers the specific extra-space symptom." }, { "left": "issue:44561", "right": "issue:45020", "accept": false, - "reason": "Both mention remote-code breakage, but they appear to involve different removed APIs and different failure surfaces." + "reason": "Related to remote-code breakage, but one is is_torch_fx_available removal and the other is broader sys.modules/modeling_utils behavior changes." }, { "left": "issue:45987", "right": "issue:46192", "accept": false, - "reason": "Different API bugs: StaticCache return type vs return_tensors being ignored with explicit text_kwargs." + "reason": "Unrelated API issues: StaticCache.get_seq_length return type vs return_tensors being ignored with text_kwargs." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Different inference/loading bugs: packed weights unpacking vs dtype mismatches in bfloat16." + "reason": "Different model-loading problems: packed-weight unpacking in BitNet vs bf16 dtype mismatches in unrelated models." }, { "left": "issue:44610", "right": "issue:45072", "accept": false, - "reason": "Different model-specific regressions: OmDet-Turbo processor size mismatch vs dtype mismatch in another subsystem." + "reason": "Different symptoms and paths: processor output size mismatch vs inference dtype mismatch." }, { "left": "issue:44671", "right": "issue:45701", "accept": false, - "reason": "Different symptoms and scope: CamemBERT MLM regression vs generic tokenization changes across versions." + "reason": "One is a CamemBERT masked-LM regression; the other is a broad complaint about tokenization changing across versions." }, { "left": "issue:43825", "right": "issue:43827", "accept": false, - "reason": "Same general v5 pipeline-removal area, but one is a runtime message bug and the other is docs drift; not the same underlying bug." + "reason": "Both mention removed pipeline tasks, but one is an error message and the other is documentation; not the same underlying bug." }, { "left": "issue:43782", "right": "issue:45325", "accept": false, - "reason": "Different Qwen-VL bugs: weight_only loading vs still-image temporal position scaling." + "reason": "Different Qwen multimodal bugs with different failure points." }, { "left": "issue:43749", "right": "issue:44291", "accept": false, - "reason": "Both are loading problems, but the causes differ: FSDP CPU RAM efficient loading vs unexpected _is_hf_initialized argument." + "reason": "Both concern model loading, but one is FSDP CPU RAM efficient loading and the other is init_empty_weights argument handling." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Different hardware/backend crashes: Apple Silicon caching allocator vs CUDA AMP loss crash." + "reason": "Different hardware/backend crashes with different stacks and failure causes." }, { "left": "issue:45081", "right": "issue:45406", "accept": false, - "reason": "Both involve Mistral/Gemma processor loading, but the concrete failures and affected objects differ." + "reason": "Different AttributeErrors in different components: tokenizer regex patching vs Gemma4Processor missing _tokenizer." }, { "left": "issue:42175", "right": "issue:45081", "accept": false, - "reason": "Unrelated: backend dependency installation vs tokenizer regex patch crash." + "reason": "Unrelated issues: pip extras dependency omission vs Mistral tokenizer regex patch crash." }, { "left": "issue:43906", "right": "issue:46209", "accept": false, - "reason": "No shared bug: isolated reproduction of an old issue vs FP8 inference failure on B200." + "reason": "An isolated reproduction of a prior issue is not the same as an FP8 inference failure on B200." }, { "left": "issue:45372", "right": "issue:46326", "accept": false, - "reason": "Different mistral_common-related breakages: ReasoningEffort import vs continue_final_message handling." + "reason": "Both involve mistral_common, but one is a missing import for Gemma 4 loading and the other is a continue_final_message tokenizer bug." }, { "left": "issue:43262", "right": "issue:43819", "accept": false, - "reason": "Different audio/model bugs: chat-template sampling-rate default vs DAC latent reconstruction mismatch." + "reason": "Different audio/DAC code paths: chat-template sampling-rate default vs from_latents forward mismatch." }, { "left": "issue:34634", "right": "issue:36331", "accept": false, - "reason": "Unrelated: BarkProcessor voice preset vs Trainer.compute_loss signature change." + "reason": "Unrelated: BarkProcessor bug vs CustomTrainer compute_loss signature mismatch." }, { "left": "issue:45003", "right": "issue:46285", "accept": false, - "reason": "Different API regressions: unsafe sys.modules access vs CLIP text model structure inconsistency." + "reason": "Different regressions: unsafe sys.modules access vs CLIP text model structure change." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "Different bugs in different areas: tokenizer special-token registration vs MXFP4 quantization fallback." + "reason": "Different bugs: BigBirdTokenizer special-token registration vs MXFP4 quantization falling back to bf16." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well-grounded in the packet and stays appropriately conservative. The one accepted pair (issue:28218 / issue:29159) is plausibly the same tokenizer spacing/inconsistency thread, while the rest of the verdicts read as careful title-based separations rather than overclaims. No mergeability concerns apply since there are no PR pairs here." + "feedback": "The summary is well-grounded in the packet and stays conservative overall. The accepted pair (issue:28218 / issue:29159) is a plausible tokenizer-behavior duplicate based on the titles, and the other soft-edge rejections are appropriately cautious rather than overstated." }, "evaluator_used": true, "retried": false, @@ -5716,218 +5735,218 @@ "issue:43550|issue:44912", "issue:43746|issue:44811", "issue:43819|issue:43927", - "issue:36010|issue:43645", "issue:44387|issue:44617", + "issue:36010|issue:43645", "issue:43257|issue:43329", "issue:43720|issue:44610", "issue:43854|issue:44977", "issue:38175|issue:44610", - "issue:36010|issue:36331", "issue:43531|issue:45042", + "issue:36010|issue:36331", "issue:46421|issue:46424", "issue:39692|issue:44912", - "issue:42490|issue:43257", "issue:45459|issue:45803", + "issue:42490|issue:43257", "issue:43262|issue:44811", - "issue:43257|issue:43696", - "issue:43454|issue:43577" + "issue:43454|issue:43577", + "issue:43257|issue:43696" ], - "prepared_review_unit_hash": "8a424f88ce2a273df8f00a65695e6b3070593b934c31a08b528c2b5c28be7fda", + "prepared_review_unit_hash": "33da1165d3715326c22b1d98776812ec2498ed972f2a651f57dc99d3e1d3b2bd", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 35711, - "estimated_input_tokens": 8928, - "estimated_eval_tokens": 18112 + "serialized_chars": 35718, + "estimated_input_tokens": 8930, + "estimated_eval_tokens": 18116 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: it mixes unrelated loading, trainer, processor, quantization, cache, and model-specific bugs. None of the soft pairs look like the same underlying issue, so I would not merge any of them as duplicates.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue stands out; the items are too diverse to represent one underlying bug or regression.", + "summary": "This cluster is a loose collection of mostly unrelated Transformers bug reports spanning loading, quantization, cache, processors, and model-specific regressions. I do not see a true duplicate/merge group among the soft pairs; most are only superficially similar by subsystem or symptom.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45020 is the broadest, most representative regression report in the set: it describes a general breakage in recent Transformers versions affecting models that rely on `remote_code`.", "canonical_pr_reason": null, - "best_issue_reason": "No issue is a strong global representative for the cluster; the closest broad items are still model- or path-specific and do not cover the rest.", + "best_issue_reason": "issue:45020 is the best single anchor for the cluster because it captures a high-level, user-facing compatibility regression rather than a model-specific one-off.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "Both are loading-related, but one is a MobileLLM key error and the other is a broad remote_code regression; different code paths." + "reason": "Both are loading regressions, but one is a MobileLLM key error and the other is a broad remote_code regression; different failure modes and likely different fixes." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "Both are vision-model bugs, but BEiT label reduction and RTDetr output drift are unrelated failures." + "reason": "Unrelated bugs: one is a Beit label-reduction bug, the other is RTDetrV2 output divergence." }, { "left": "issue:45084", "right": "issue:45538", "accept": false, - "reason": "Compiler/template-node errors and CLIPTokenizer max length are different subsystems with no shared bug." + "reason": "Different issues in different layers: compilation error vs tokenizer max_length default." }, { "left": "issue:44060", "right": "issue:44368", "accept": false, - "reason": "Both mention tied-weights warnings, but one is a specific bad tie mapping and the other is a LoRA warning from ms-swift; not the same fix." + "reason": "Both mention Qwen tied-weights warnings, but one is a specific incorrect parameter tying bug and the other is a generic warning during LoRA fine-tuning; not clearly the same fix." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Processor label reduction and MPS attention correctness are unrelated issues." + "reason": "No overlap beyond being correctness bugs; label reduction and MPS attention correctness are unrelated." }, { "left": "issue:43329", "right": "issue:45600", "accept": false, - "reason": "Multimodal token counting in the video branch and stale auto_mappings entries are different bugs." + "reason": "Different models and code paths: video token counting bug vs stale auto_mappings reference." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Float8 storage loading and PIL backend torchvision requirements are unrelated." + "reason": "Different concerns: float8 storage lookup vs unwanted torchvision dependency." }, { "left": "issue:43650", "right": "issue:43906", "accept": false, - "reason": "One is a placeholder title; the other is an isolated reproduction of a different bug." + "reason": "One is a placeholder issue title, the other is a reproduction of a specific upstream bug; not the same underlying problem." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both involve from_pretrained loading, but gguf_file handling and a tokenizer-loading regression are different code paths." + "reason": "Both affect loading, but one is gguf handling across AutoTokenizer/AutoModel and the other is a specific tokenizer load failure for EMBEDDIA/est-roberta." }, { "left": "issue:43638", "right": "issue:44936", "accept": false, - "reason": "Trainer/deepspeed index errors after init and evaluate-after-train behavior are distinct failures." + "reason": "Different trainer failures: ZeRO-3 index error during training setup vs evaluate-after-train failure." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile with SDPA on Bamba and MXFP4 quantization fallback on git-oss are unrelated model-specific bugs." + "reason": "Unrelated quantization/model bugs: SDPA compile failure in Bamba vs MXFP4 load fallback in git-oss-20b." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "PEFT checkpoint loading for GraniteSpeech and Whisper batch_decode special-token handling are different components." + "reason": "Different processor/model behaviors: local PEFT checkpoint loading vs Whisper batch_decode option handling." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "DAC latent decoding mismatch and DiaConfig token-ID persistence are unrelated." + "reason": "Different bugs: DAC latent reconstruction mismatch vs DiaConfig losing custom token IDs on save/load." }, { - "left": "issue:36010", - "right": "issue:43645", + "left": "issue:44387", + "right": "issue:44617", "accept": false, - "reason": "Both touch custom model compatibility, but GenerationMixin import failure and Jupyter initialization breakage are different regressions." + "reason": "Both can lead to OOM, but the causes are different: int4 reserved-memory growth vs a specific video model consuming too much memory." }, { - "left": "issue:44387", - "right": "issue:44617", + "left": "issue:36010", + "right": "issue:43645", "accept": false, - "reason": "Both are OOMs, but one is an int4 reserved-memory regression and the other is a model-specific Sam3Video OOM." + "reason": "One is an import path change for GenerationMixin, the other is a notebook custom-model initialization regression; not the same code-path." }, { "left": "issue:43257", "right": "issue:43329", "accept": false, - "reason": "Qwen3-MoE loading/conversion and multimodal video token counting are not the same bug." + "reason": "Same family, different defects: weight conversion under accelerate+deepspeed vs sliding-window handling in Qwen3-MoE." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "BitNet packed-weight loading and OmDet processor/model input-size mismatch are unrelated." + "reason": "Different subsystems entirely: BitNet packed-weight loading vs OmDet-Turbo image size mismatch." }, { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Unit-test loading failure for GLM-4.7 and flash-attention generation issues for Qwen3.5 are different code paths." + "reason": "Both are model-specific load/generation problems, but they concern different models and different failure mechanisms." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 zero-probabilities and OmDet processor sizing are unrelated model bugs." + "reason": "Unrelated model behavior: zero probabilities in SigLIP2 vs OmDet-Turbo processor/model size mismatch." }, { - "left": "issue:36010", - "right": "issue:36331", + "left": "issue:43531", + "right": "issue:45042", "accept": false, - "reason": "GenerationMixin import failure and compute_loss signature incompatibility are different API breakages." + "reason": "Different bugs in different components: Qwen3-MoE sliding-window handling vs PIL backend dependency detection." }, { - "left": "issue:43531", - "right": "issue:45042", + "left": "issue:36010", + "right": "issue:36331", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and PIL/torchvision dependency handling have no shared underlying defect." + "reason": "ImportError for GenerationMixin and Trainer compute_loss signature mismatch are separate API regressions." }, { "left": "issue:46421", "right": "issue:46424", "accept": false, - "reason": "Both are cache-related, but lazy initialization causing recompilation and max_cache_length semantics are separate issues." + "reason": "Both are cache-related, but one is about lazy init causing recompilation and the other about `max_cache_length` only applying to quantized cache; different fixes." }, { "left": "issue:39692", "right": "issue:44912", "accept": false, - "reason": "Documentation example errors and quantization fallback loading are not the same bug." + "reason": "Different model/documentation/quantization issues; one is a SigLIP2 example bug, the other is MXFP4 loading behavior for git-oss-20b." }, { - "left": "issue:42490", - "right": "issue:43257", + "left": "issue:45459", + "right": "issue:45803", "accept": false, - "reason": "Trailing-slash path warnings and Qwen3-MoE weight conversion failures are unrelated." + "reason": "Both involve exception handling, but in different functions and with different failure semantics; not the same bug." }, { - "left": "issue:45459", - "right": "issue:45803", + "left": "issue:42490", + "right": "issue:43257", "accept": false, - "reason": "Both involve exception handling, but protobuf decode fallback and Fuyu tensor conversion are different code paths." + "reason": "Local-path import warning with trailing slash is unrelated to Qwen3 MoE weight conversion under accelerate+deepspeed." }, { "left": "issue:43262", "right": "issue:44811", "accept": false, - "reason": "Audio chat-template sample-rate defaults and Whisper batch_decode skipping special tokens are unrelated processor behaviors." + "reason": "Different audio processor issues: sampling-rate defaulting vs batch_decode honoring `skip_special_tokens`." }, { - "left": "issue:43257", - "right": "issue:43696", + "left": "issue:43454", + "right": "issue:43577", "accept": false, - "reason": "Qwen3-MoE conversion/loading and GPT-oss CUDA OOM are different model/runtime problems." + "reason": "Different model-loading bugs in different architectures: AyaVision weight tying vs Blip2 dtype handling." }, { - "left": "issue:43454", - "right": "issue:43577", + "left": "issue:43257", + "right": "issue:43696", "accept": false, - "reason": "AyaVision lm_head tying and Blip2 dtype propagation are unrelated model-loading bugs." + "reason": "Completely different issues: Qwen3 MoE weight conversion vs GPT-oss-20b out-of-memory." } ] }, @@ -6014,8 +6033,8 @@ "issue:44291|issue:45042", "issue:42915|issue:44821", "issue:44987|issue:45538", - "issue:43927|issue:44912", "issue:42915|issue:45538", + "issue:43927|issue:44912", "issue:43452|issue:43582", "issue:43697|issue:45357", "issue:44849|issue:45701", @@ -6026,235 +6045,226 @@ "issue:43317|issue:45538", "issue:44479|issue:45381", "issue:44062|issue:44991", - "issue:28282|issue:34689", "issue:43540|issue:44479", + "issue:28282|issue:34689", "issue:45137|issue:45405", "issue:43540|issue:44560", "issue:43742|issue:44291", "issue:24643|issue:29159", "issue:44987|issue:45907" ], - "prepared_review_unit_hash": "e59c4b1aba85db10714c14d14755ceadc761b91673a17ce2b49dc65eac4f274e", + "prepared_review_unit_hash": "844fbe8fbcf65634585b0bdc91ddf2083fc26a77db94fad55f5a8f9bf37dcc6f", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37323, - "estimated_input_tokens": 9331, - "estimated_eval_tokens": 18918 + "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": "The cluster is thematically related around Transformers regressions, but the soft-paired issues mostly describe distinct bugs, models, or code paths rather than true duplicates. I would not merge any of the proposed soft edges; the set is too heterogeneous for duplicate collapsing.", - "confidence": 0.44, - "canonical_issue_reason": "issue:45701 is the broadest representative in the set: it is an open, cross-version tokenization regression with multiple inbound references, and it best matches the general theme of version-dependent behavior changes.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers bugs, with one clear subcluster around Qwen video-input regressions in v5.3.0. I only merged the pairs that appear to describe the same underlying video-processing bug family; the rest are false positives across different code paths and models.", + "confidence": 0.79, + "canonical_issue_reason": "issue:44479 is the best representative because it describes the broad Qwen video-input regression affecting multiple model families, and it matches the most convincing soft-edge neighbors in this set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the best single issue to anchor the cluster because it is the most general and widely referenced regression report, even though the overall set is not a tight duplicate cluster.", + "best_issue_reason": "issue:44479 is the strongest cluster anchor: it is a concrete, versioned regression with the broadest scope and the clearest overlap to other video-input reports in the set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43643", "right": "issue:45759", "accept": false, - "reason": "Different APIs and failure modes: missing fields from AutoConfig with trust_remote_code versus not unwrapping text_config in AutoModelForCausalLM.from_config." + "reason": "Different loading paths and failures: missing fields from trust_remote_code config vs. unwrapping text_config for composite multimodal CausalLM config." }, { "left": "issue:43329", "right": "issue:44479", - "accept": false, - "reason": "Both are video-related, but one is an undefined-variable bug in _get_num_multimodal_tokens while the other is a broader v5.3.0 regression across several Qwen video models." + "accept": true, + "reason": "Both describe the same Qwen video-input regression class in the multimodal token/vision-video path, and the titles point to the same code area rather than separate bugs." }, { "left": "issue:43329", "right": "issue:43450", "accept": false, - "reason": "The first is a multimodal token-counting helper bug; the second is an incorrect batched video processor output shape. Different code paths." + "reason": "Both are video-related, but one is an internal token-counting bug and the other is a batched video-processor shape problem; these are different failure modes." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "TrainerState token counters and perplexity on GPT-OSS are unrelated problems." + "reason": "Unrelated issues: trainer token accounting vs. unusually high perplexity on a specific model/dataset." }, { "left": "issue:43606", "right": "issue:45859", "accept": false, - "reason": "Both mention offload/device issues, but they affect different models and code paths: suno/bark-small CPU offload versus Qwen3_5MoeTextRotaryEmbedding.forward." + "reason": "Both involve CPU offload, but they concern different models and different concrete errors; not enough evidence of one underlying bug." }, { "left": "issue:29159", "right": "issue:34567", "accept": false, - "reason": "Tokenizer parity between slow/fast implementations is unrelated to TrainerState.num_input_tokens_seen not updating." + "reason": "Tokenizer consistency issue vs. trainer state token counting; different subsystems and symptoms." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "Both are memory/loading regressions, but FSDP CPU RAM-efficient loading and Strix Halo mmap OOM are different failures and fixes." + "reason": "Different problems: FSDP CPU RAM efficient loading vs. disabling mmap on Strix Halo to avoid OOM." }, { "left": "issue:43257", "right": "issue:43931", "accept": false, - "reason": "One is Qwen3 MoE loading with accelerate+deepspeed conversion issues; the other is a Qwen3-VL weight-shape mismatch during model load." + "reason": "Both are Qwen loading issues, but one is accelerate+deepspeed MOE weight conversion and the other is a shape mismatch for a specific VL model." }, { "left": "issue:44568", "right": "issue:45356", "accept": false, - "reason": "Special-token insertion for mdeberta-v3 is a tokenizer behavior bug, while Kimi-K2.5 is a codec/regex regression with a warning issue." + "reason": "Distinct tokenizer regressions affecting different models and different mechanisms (special token insertion vs codec handling/warning behavior)." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "FineGrainedFP8Config failure in Qwen3Moe and a key error loading MobileLLM-125M are different loading/configuration bugs." + "reason": "Different loading failures: FineGrainedFP8Config vs. a key error for MobileLLM-125M." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "Unexpected _is_hf_initialized in init_empty_weights is unrelated to PIL backend processors requiring torchvision." + "reason": "Different areas: init_empty_weights argument handling vs. PIL/image processor dependency requirements." }, { "left": "issue:42915", "right": "issue:44821", "accept": false, - "reason": "These are unrelated: Qwen3Moe FP8 loading versus AutoImageProcessor URL loading." + "reason": "No meaningful overlap: Qwen3Moe FP8 failure vs. loading AutoImageProcessor from URL." }, { "left": "issue:44987", "right": "issue:45538", "accept": false, - "reason": "Model loading failure for physical-intelligence/fast is unrelated to CLIPTokenizer's model_max_length default." + "reason": "Different tokenizer/model-loading bugs with no shared concrete code path." }, { - "left": "issue:43927", - "right": "issue:44912", + "left": "issue:42915", + "right": "issue:45538", "accept": false, - "reason": "DiaConfig save/load losing custom token IDs is a config persistence bug, not the MXFP4 quantization fallback problem." + "reason": "Different models and failures; one is FP8 config handling, the other is an extreme model_max_length issue." }, { - "left": "issue:42915", - "right": "issue:45538", + "left": "issue:43927", + "right": "issue:44912", "accept": false, - "reason": "A Qwen3Moe FP8 loading issue does not match a tokenizer max-length bug." + "reason": "Custom token IDs being dropped on save/load is unrelated to MXFP4 quantization falling back to bf16." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "gguf_file loading failures and Apple Silicon allocator warmup TypeError are different subsystems and symptoms." + "reason": "GGUF loading breakage and an AppleSilicon caching_allocator_warmup TypeError are unrelated." }, { "left": "issue:43697", "right": "issue:45357", "accept": false, - "reason": "RTDetrV2 output drift in v5 versus Qwen3.5 save_pretrained key corruption are separate regressions." + "reason": "Different regression classes: RTDetr output differences vs. Qwen3.5 save_pretrained key serialization." }, { "left": "issue:44849", "right": "issue:45701", "accept": false, - "reason": "Qwen3.5 output_hidden_states behavior is a model forward-pass issue, not the tokenizer version-regression reported in issue:45701." + "reason": "Qwen3.5 hidden-state output bug and a broad tokenization-version-change report are not the same underlying issue." }, { "left": "issue:44568", "right": "issue:45488", "accept": false, - "reason": "Both are tokenizer-related, but add_special_tokens behavior and LlamaTokenizer overriding tokenizer.json's pre-tokenizer are distinct bugs." + "reason": "Both are tokenizer regressions, but they affect different tokenizers and different behaviors; too broad to call the same bug." }, { "left": "issue:33357", "right": "issue:34567", "accept": false, - "reason": "MacOS bus error on a CLIP model and TrainerState token counting are not the same problem." + "reason": "Bus error on MacOS CLIP model loading vs. trainer token counting; unrelated." }, { "left": "issue:41720", "right": "issue:43540", "accept": false, - "reason": "cudaErrorAssert with auto device mapping is a device-placement/runtime issue, while the other is a video-input ValueError in Qwen3OmniMoe." + "reason": "Different Qwen3 issues and different failure points: auto device mapping CUDA assert vs. video input processing ValueError." }, { "left": "issue:42175", "right": "issue:43824", "accept": false, - "reason": "Missing TensorFlow in a torch-only install and a missing Qwen2_5_VLForConditionalGeneration import are unrelated packaging/import issues." + "reason": "Backend packaging issue vs. import failure for a specific model class; unrelated." }, { "left": "issue:43317", "right": "issue:45538", "accept": false, - "reason": "Dequantized model loading with device_map=auto is unrelated to CLIPTokenizer model_max_length." + "reason": "Device-map/offload dequantized model loading is unrelated to CLIP tokenizer max length." }, { "left": "issue:44479", "right": "issue:45381", - "accept": false, - "reason": "Both are video regressions in the Qwen family, but one is a broad multi-model v5.3.0 regression and the other is a specific qwen2.5-vl vision_position_ids bug; not clearly the same underlying defect." + "accept": true, + "reason": "Both are the same Qwen video-input regression family in v5.3.0, with one report broad across Qwen video models and the other a specific vision_position_ids manifestation." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken special-argument duplication and tokenizer loading failure for EMBEDDIA/est-roberta are different tokenizer bugs." + "reason": "Generic AddedToken argument conflict and a tokenizer loading failure for a specific model do not look like the same bug." }, { - "left": "issue:28282", - "right": "issue:34689", + "left": "issue:43540", + "right": "issue:44560", "accept": false, - "reason": "A missing PyTorch import and a specific Llama 3.2 90B Vision loading regression are not the same issue." + "reason": "Both are video-related, but they concern different models and different errors; not enough evidence they share one fix." }, { - "left": "issue:43540", - "right": "issue:44479", + "left": "issue:28282", + "right": "issue:34689", "accept": false, - "reason": "Both concern video input failures, but the concrete failures differ: Qwen3OmniMoe raises ValueError, while the other is a broader regression across several Qwen video models." + "reason": "PyTorch-not-found import error and a Llama Vision model loading regression are unrelated." }, { "left": "issue:45137", "right": "issue:45405", "accept": false, - "reason": "FSDP CPU_RAM_EFFICIENT_LOADING and a bumped MIN_PEFT_VERSION are unrelated; one is a runtime bug, the other is a dependency-version problem." - }, - { - "left": "issue:43742", - "right": "issue:44291", - "accept": false, - "reason": "Key error loading MobileLLM-125M is unrelated to the init_empty_weights unexpected-argument regression." - }, - { - "left": "issue:24643", - "right": "issue:29159", - "accept": false, - "reason": "DeepSpeed training weight-shape error and slow/fast tokenizer inconsistency do not share a code path." + "reason": "DeepSpeed ZeRO3 deque error vs. a PEFT version requirement bump are unrelated." }, { "left": "issue:44987", "right": "issue:45907", "accept": false, - "reason": "Model loading failure for a specific repo is unrelated to list_repo_files not forwarding revision." + "reason": "Different CI/load failures with no shared code path: model loading regression vs. list_repo_files not forwarding revision." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted soft edges are plausible members of the same Qwen video-input regression family, while the rejections mostly avoid overclaiming shared root causes across different models and subsystems. The choice of issue:44479 as the canonical anchor is reasonable based on its broad v5.3.0 video regression title and its overlap with the accepted neighbors." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -6326,18 +6336,18 @@ "issue:42915|issue:43749", "issue:42175|issue:44488", "issue:38175|issue:44898", - "issue:43493|issue:43994", "issue:44291|issue:44821", + "issue:43493|issue:43994", "issue:43257|issue:43756", - "issue:34567|issue:36331", "issue:44737|issue:45588", + "issue:34567|issue:36331", "issue:43582|issue:43782", "issue:43526|issue:45479", "issue:44062|issue:45081", "issue:43404|issue:45987", "issue:27957|issue:28282", - "issue:44792|issue:44964", "issue:44913|issue:46121", + "issue:44792|issue:44964", "issue:43906|issue:45381", "issue:45042|issue:45081", "issue:43756|issue:46399", @@ -6349,206 +6359,203 @@ "issue:45405|issue:45600", "issue:43643|issue:45375" ], - "prepared_review_unit_hash": "7c6e4ce5b7789b821e85e3a4013f7369675acf955d4bed5ff3da5febb2c068be", + "prepared_review_unit_hash": "cf314d89686c5bc8a1ea099f7c31b90254424e2543c0980464b1abe6df12b91d", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 36187, - "estimated_input_tokens": 9047, - "estimated_eval_tokens": 18350 + "serialized_chars": 36190, + "estimated_input_tokens": 9048, + "estimated_eval_tokens": 18352 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is still mostly heterogeneous, with no strong PR material and only one plausible duplicate issue pair. For the issue anchor, 43749 is a solid representative, but not a clear dominant triage winner over 44488; the signal is split across several unrelated reports.", - "confidence": 0.79, - "canonical_issue_reason": "issue:43749 is a reasonable representative for the FSDP/loading subcluster, but the packet does not support calling it the single highest-signal issue; discussion activity is higher on issue:44488, so 43749 should be treated as one strong candidate rather than the clear top.", + "summary": "This cluster is a grab-bag of mostly unrelated Transformers bug reports across different models and subsystems. None of the soft pairs look like the same underlying defect, so they should stay split. If one issue must serve as the anchor, issue 43749 is the strongest representative because it has the broadest impact and the most discussion/references.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43749 is the most central and well-connected item in this mixed cluster, with multiple inbound references and substantial discussion.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44488 is the strongest overall standalone issue candidate by discussion activity and still has some inbound signal, making it a slightly better global representative than 43749 for the cluster as a whole.", + "best_issue_reason": "issue:43749 is the best single issue to anchor triage because it describes a cross-cutting loading failure and has the strongest evidence of broader impact.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:28282", "right": "issue:29127", "accept": false, - "reason": "Different problems: missing PyTorch import vs. LayoutLMv3 error-message clarity." + "reason": "Different problems: PyTorch import/dependency detection vs a LayoutLMv3 error-message clarity issue." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Both mention dtype, but they affect different models and failure modes; not the same bug." + "reason": "Both mention dtype, but they affect different models and code paths; not the same concrete bug." }, { "left": "issue:42915", "right": "issue:43749", "accept": false, - "reason": "Different subsystems and symptoms: FineGrainedFP8Config vs. FSDP CPU RAM efficient loading." + "reason": "FineGrainedFP8/Qwen3MoE loading failure is unrelated to FSDP CPU RAM efficient loading." }, { "left": "issue:42175", "right": "issue:44488", "accept": false, - "reason": "Packaging/backend dependency complaint vs. a specific model-loading failure; not the same underlying issue." + "reason": "Backend package dependency issue vs a model-loading failure for cjvt/sleng-bert." }, { "left": "issue:38175", "right": "issue:44898", "accept": false, - "reason": "SigLIP2 probability output issue vs. Perceiver interpolation/resolution behavior; different bugs." - }, - { - "left": "issue:43493", - "right": "issue:43994", - "accept": true, - "reason": "Same SigLIP2 model family and same underlying symptom class: incorrect/nonsensical outputs from the HF implementation." + "reason": "Separate vision-model bugs: unexpected zero probabilities in SigLIP2 vs Perceiver non-default resolution failure." }, { "left": "issue:44291", "right": "issue:44821", "accept": false, - "reason": "Unexpected `_is_hf_initialized` argument vs. AutoImageProcessor URL loading failure; different code paths." + "reason": "init_empty_weights/_is_hf_initialized TypeError is unrelated to AutoImageProcessor URL loading." }, { - "left": "issue:43257", - "right": "issue:43756", + "left": "issue:43493", + "right": "issue:43994", "accept": false, - "reason": "Different Qwen-family bugs with unrelated mechanics: weight conversion/loading vs. RoPE layer dropping." + "reason": "Same model family, but one is HF-vs-JAX parity and the other is bad AutoModel/pipeline outputs; not clearly one fix." }, { - "left": "issue:34567", - "right": "issue:36331", + "left": "issue:43257", + "right": "issue:43756", "accept": false, - "reason": "TrainerState token tracking and CustomTrainer signature mismatch are separate trainer issues." + "reason": "Different Qwen/Smollm issues: MOE weight conversion under accelerate+deepspeed vs RoPE-layer dropping." }, { "left": "issue:44737", "right": "issue:45588", "accept": false, - "reason": "XLNet CPU placement bug vs. flash_attention crash on sink-less models; unrelated paths." + "reason": "XLNet CPU placement bug is unrelated to flash_attention.py crashing on sink-less models." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token accounting and CustomTrainer compute_loss signature are different training bugs." }, { "left": "issue:43582", "right": "issue:43782", "accept": false, - "reason": "Apple Silicon allocator warmup TypeError vs. Qwen3VL weight_only load error; not the same bug." + "reason": "Apple Silicon allocator TypeError is unrelated to Qwen3VL weight_only loading failure." }, { "left": "issue:43526", "right": "issue:45479", "accept": false, - "reason": "Image label reduction bug vs. sequence-classification zero-loss degeneracy; different issues." + "reason": "BeitImageProcessorFast label reduction bug is separate from single-label-classification zero-loss behavior." }, { "left": "issue:44062", "right": "issue:45081", "accept": false, - "reason": "Both tokenizer-related, but the concrete crashes and affected paths differ." + "reason": "Tokenizer AddedToken argument collision and Mistral regex patch backend attribute crash are different failures." }, { "left": "issue:43404", "right": "issue:45987", "accept": false, - "reason": "Mistral lm_head tying bug vs. StaticCache return-type contract issue; unrelated." + "reason": "Mistral3 lm_head tying issue is unrelated to StaticCache.get_seq_length returning the wrong type." }, { "left": "issue:27957", "right": "issue:28282", "accept": false, - "reason": "Unrelated: Flash Attention 2 support for XLM-Roberta vs. missing PyTorch import error." + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to a PyTorch ImportError on AutoModel." }, { - "left": "issue:44792", - "right": "issue:44964", + "left": "issue:44913", + "right": "issue:46121", "accept": false, - "reason": "Both are model-loading/test failures, but for different multimodal models and different failure modes." + "reason": "Both involve RoPE/config handling, but one is reload losing rotary_pct and the other is JSON list type handling during validation." }, { - "left": "issue:44913", - "right": "issue:46121", + "left": "issue:44792", + "right": "issue:44964", "accept": false, - "reason": "Related to RoPE/config handling, but one is reload defaulting and the other is JSON-list validation; not the same bug." + "reason": "Janus test failure and Phi-4 multimodal loading failure are different multimodal issues." }, { "left": "issue:43906", "right": "issue:45381", "accept": false, - "reason": "Different models and failure modes; the reproduction issue is not the same as the Qwen2.5-VL position-id bug." + "reason": "Generic reproduction of another issue is unrelated to Qwen2.5-VL video vision_position_ids miscalculation." }, { "left": "issue:45042", "right": "issue:45081", "accept": false, - "reason": "Image-processor backend dependency issue vs. Mistral tokenizer regex patch crash." + "reason": "Image-processor torchvision dependency regression is unrelated to tokenizer regex patching." }, { "left": "issue:43756", "right": "issue:46399", "accept": false, - "reason": "Both Qwen-related, but the underlying code paths differ: RoPE layer handling vs. rotary-kernel registration." + "reason": "Both touch RoPE/Qwen3.5-adjacent areas, but one is missing RoPE layers and the other is a registration/decorator crash." }, { "left": "issue:43756", "right": "issue:45325", "accept": false, - "reason": "Both involve Qwen, but one is RoPE-layer count and the other is video temporal position IDs." + "reason": "Smollm3 RoPE-layer dropping and Qwen2.5-VL temporal position scaling are different models and different bugs." }, { "left": "issue:43742", "right": "issue:44589", "accept": false, - "reason": "MobileLLM key error vs. missing Float8 storage object; different loading failures." + "reason": "MobileLLM loading key error is unrelated to Float8 storage lookup failure." }, { "left": "issue:43452", "right": "issue:44488", "accept": false, - "reason": "Different failures: gguf_file breaks multiple from_pretrained entrypoints, while the other is a specific checkpoint load problem." + "reason": "GGUF file handling for tokenizer/CausalLM is unrelated to sleng-bert loading failure." }, { "left": "issue:43531", "right": "issue:44821", "accept": false, - "reason": "Qwen3-MoE sliding-window bug vs. AutoImageProcessor URL loading failure; unrelated." + "reason": "Qwen3-MoE sliding_window behavior is unrelated to loading AutoImageProcessor from URL." }, { "left": "issue:45357", "right": "issue:45920", "accept": false, - "reason": "Incorrect visual encoder keys on save vs. wrong token IDs from AutoTokenizer; different regressions." + "reason": "Qwen3.5 visual encoder save_pretrained bug is unrelated to AutoTokenizer producing wrong token IDs across multiple models." }, { "left": "issue:45405", "right": "issue:45600", "accept": false, - "reason": "Dependency version bump vs. auto_mappings CI breakage; not the same issue." + "reason": "Unreleased PEFT version bump is unrelated to removed Sam3LiteText config references breaking CI." }, { "left": "issue:43643", "right": "issue:45375", "accept": false, - "reason": "Missing fields with trust_remote_code vs. missing deepstack_visual_indexes under strict config handling; different serialization problems." + "reason": "AutoConfig trust_remote_code missing fields is a different config-loading problem than Qwen3_5MoeVisionConfig strict-field dropping." } ] }, - "evaluator_result": { - "accept": false, - "feedback": "The cluster-level summary is mostly conservative, but the canonical issue claim overstates the evidence. `issue:43749` is a reasonable anchor, yet the packet does not support saying it has the \"highest triage signal\" or the \"most discussion and inbound references\": `issue:44488` has more discussion activity, and the evidence is not strong enough to rank `43749` as the clear top across those dimensions. Reframe it as one of the stronger candidates rather than the single highest. The accepted SigLIP2 duplicate edge is plausible, but the issue-anchor overclaim is enough to reject." - }, - "evaluator_used": true, - "retried": true, - "accepted_nontrivial_soft_edge": true, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -6629,36 +6636,36 @@ "issue:43819|issue:44811", "issue:43688|issue:43749", "issue:43452|issue:43577", + "issue:45372|issue:46170", "issue:28282|issue:33357", "issue:44936|issue:45823", - "issue:45372|issue:46170", "issue:43906|issue:46202", "issue:44361|issue:44442", "issue:43653|issue:44361", "issue:43295|issue:45356", - "issue:38175|issue:43295", "issue:44484|issue:44704", + "issue:38175|issue:43295", "issue:45542|issue:46178", "issue:24643|issue:27957", "issue:43653|issue:44568", - "issue:29159|issue:33357", "issue:43577|issue:43697", + "issue:29159|issue:33357", "issue:44464|issue:44743" ], - "prepared_review_unit_hash": "106c8d68eea7af31c6e4a0c05c37e2356006734bc0c8b9260af232275e6e3bd4", + "prepared_review_unit_hash": "06578af5df1362ab1ec15925683c06397b5855fcc12b5cb1c99a58221cf04572", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36661, + "serialized_chars": 36663, "estimated_input_tokens": 9166, "estimated_eval_tokens": 18588 }, @@ -6667,186 +6674,180 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: most items are unrelated Transformers bugs/regressions across loading, tokenizers, processors, training, and serialization. I found no safe duplicate merges; the only close theme is tied-weights/v5, but even that is not clearly the same concrete bug.", - "confidence": 0.27, - "canonical_issue_reason": "issue:45005 is the broadest and most representative report here, centered on tied-weights regressions in v5, so it is the closest umbrella issue for this noisy cluster.", + "summary": "The items are mostly heterogeneous, with a loose cluster around v5/tokenizer/save_pretrained regressions, but none of the soft pairs look like the same concrete bug. No PRs are present.", + "confidence": 0.61, + "canonical_issue_reason": "45005 is the broadest representative issue in the list, centered on a v5 tied-weights regression; the others are either more specific manifestations or unrelated failures.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45005 is the best single anchor for the cluster because it captures the most reusable high-level theme (tied weights in v5) without overfitting to one narrow symptom.", + "best_issue_reason": "45005 is the most generally framed and likely to subsume related tied-weights regressions if any duplicate exists.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43720", "right": "issue:44898", "accept": false, - "reason": "Different bugs and code paths: BitNet packed-weight loading in accelerate vs Perceiver image classification with interpolate_pos_encoding." + "reason": "Different models and failure modes: BitNet packed-weight loading vs Perceiver interpolation/resolution handling." }, { "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 strict loading for granite_speech; not the same defect." + "reason": "Both involve config/loading, but the bugs are different: num_labels propagation vs strict config rejection for granite_speech." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "DeepSpeed training weight-shape error and LayoutLMv3 box-validation messaging are unrelated failures." + "reason": "Completely different problems: DeepSpeed training weight-shape error vs LayoutLMv3 box validation messaging." }, { "left": "issue:43697", "right": "issue:45216", "accept": false, - "reason": "Both are v5 regressions, but one is RTDetrV2 output drift and the other is Qwen3.5 save_pretrained checkpoint corruption; different paths." + "reason": "Different regressions: RTDetrV2 output changes vs Qwen3.5 checkpoint save_pretrained correctness." }, { "left": "issue:41762", "right": "issue:44936", "accept": false, - "reason": "Gemma3 ZeRO-3 loading IndexError and trainer.evaluate-after-train behavior are distinct bugs." + "reason": "Different paths: Gemma3 ZeRO-3 loading crash vs trainer evaluate-after-train behavior." }, { "left": "issue:44493", "right": "issue:46050", "accept": false, - "reason": "Unexpected position-id keys and sinq quantization save/load breakage are not the same underlying issue." + "reason": "Unexpected key/position-id serialization issue is not the same as sinq quantized save/load breakage." }, { "left": "issue:44514", "right": "issue:45381", "accept": false, - "reason": "Both involve Qwen2.5-VL, but one is a batched processor chat-template crash and the other is a video position-id bug; different failures." + "reason": "Both are Qwen2.5-VL related, but one is batched chat template padding and the other is video vision_position_ids." }, { "left": "issue:44466", "right": "issue:45005", "accept": false, - "reason": "Close topic overlap around tied weights, but 44466 is a specific lm_head serialization inconsistency while 45005 is a broader translation-model tied-weights report; not clearly one bug." + "reason": "Related theme of tied weights, but not clearly the same concrete bug; one is device-dependent lm_head serialization, the other is broader translation-model tied-weight issues." }, { "left": "issue:30064", "right": "issue:33666", "accept": false, - "reason": "Void segmentation-map processing and Qwen2-VL multi-GPU training are unrelated." + "reason": "Unrelated: void segmentation map processing vs Qwen2-VL multi-GPU training." }, { "left": "issue:43452", "right": "issue:44843", "accept": false, - "reason": "gguf_file loading breakage and an offline model_info() call in the mistral regex patch are different from different code paths." + "reason": "Different loader failures: gguf_file handling vs unconditional model_info() call in mistral regex patch." }, { "left": "issue:28282", "right": "issue:32090", "accept": false, - "reason": "A missing-PyTorch import error and a NoneType broadcast error in Trainer are unrelated." + "reason": "Different errors and layers: missing PyTorch dependency vs Trainer GPU broadcast NoneType failure." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "DAC from_latents/STE mismatch and Whisper batch_decode skip_special_tokens behavior are unrelated." + "reason": "Different functionality: DAC latent reconstruction vs Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:43688", "right": "issue:43749", "accept": false, - "reason": "Auxiliary-loss normalization in MoE models and FSDP CPU RAM efficient loading are different bugs." + "reason": "Distinct bugs: auxiliary loss normalization vs FSDP CPU RAM efficient loading." }, { "left": "issue:43452", "right": "issue:43577", "accept": false, - "reason": "gguf_file loading failures and BLIP2 dtype propagation are unrelated." + "reason": "Both concern model loading, but one is gguf/AutoTokenizer and the other is BLIP2 dtype propagation." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Different dependency failures: mistral_common ReasoningEffort import vs Windows/ROCm torch.distributed dependency." }, { "left": "issue:28282", "right": "issue:33357", "accept": false, - "reason": "Generic AutoModel import error and MacOS bus error on a CLIP community model do not share the same root cause." + "reason": "Unrelated: PyTorch missing ImportError vs MacOS bus error with CLIP." }, { "left": "issue:44936", "right": "issue:45823", "accept": false, - "reason": "trainer.evaluate lifecycle failure and Gemma4 PLE device mismatch are distinct issues." - }, - { - "left": "issue:45372", - "right": "issue:46170", - "accept": false, - "reason": "Gemma 4 processor import failure and Windows ROCm torch.distributed dependency errors are unrelated." + "reason": "Different runtime contexts and symptoms: trainer lifecycle bug vs Gemma4 PLE device mismatch." }, { "left": "issue:43906", "right": "issue:46202", "accept": false, - "reason": "An isolated reproduction of an earlier issue is not the same as the Gemma pipeline roundtrip failure on \u2581." + "reason": "Both mention Gemma/pipeline behavior, but there is no clear evidence they are the same underlying roundtrip bug." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "MLukeTokenizer task AttributeError and FastSpeech2ConformerTokenizer loading failure are different tokenizer bugs." + "reason": "Different tokenizer failures: MLuke task AttributeError vs FastSpeech2ConformerTokenizer loading." }, { "left": "issue:43653", "right": "issue:44361", "accept": false, - "reason": "BigBird mask-token special-token registration and MLukeTokenizer task handling are unrelated." + "reason": "Different tokenizer issues: BigBird special-token registration vs MLuke task AttributeError." }, { "left": "issue:43295", "right": "issue:45356", "accept": false, - "reason": "Both are regressions around tokenizer/processor behavior, but one is processor.tokenizer access/images passthrough and the other is Kimi-K2.5 codec handling plus warning text; not the same bug." - }, - { - "left": "issue:38175", - "right": "issue:43295", - "accept": false, - "reason": "SigLIP2 zero-probability outputs and processor.tokenizer regression are unrelated." + "reason": "Both are regressions involving processor/tokenizer behavior, but the concrete breakages differ." }, { "left": "issue:44484", "right": "issue:44704", "accept": false, - "reason": "A save_pretrained shard-size question and an AutoProcessor kwargs-forwarding bug are not duplicates." + "reason": "Different topics: save_pretrained shard size question vs missing kwargs passed to cached_file." }, { - "left": "issue:45542", - "right": "issue:46178", + "left": "issue:38175", + "right": "issue:43295", "accept": false, - "reason": "Both are backend/import problems, but tensorboard-without-TensorFlow and tensorflow_text backend registration are distinct failure modes." + "reason": "Unrelated: zero probabilities in SigLIP2 vs processor.tokenizer regression." }, { - "left": "issue:24643", - "right": "issue:27957", + "left": "issue:45542", + "right": "issue:46178", "accept": false, - "reason": "DeepSpeed weight-shape training error and XLMRoberta Flash Attention 2 are unrelated." + "reason": "Both concern backend availability, but they are different missing-backend errors with different libraries." }, { "left": "issue:43653", "right": "issue:44568", "accept": false, - "reason": "BigBird special-token registration and mdeberta-v3 add_special_tokens BOS/EOS behavior are different tokenizer issues." + "reason": "Both involve tokenizer special-token behavior, but the specific bugs and models differ." }, { - "left": "issue:29159", - "right": "issue:33357", + "left": "issue:43577", + "right": "issue:43697", "accept": false, - "reason": "Slow-vs-fast tokenizer inconsistency and a MacOS CLIP bus error are unrelated." + "reason": "Different issues: BLIP2 dtype staying float32 vs RTDetrV2 output changes." }, { - "left": "issue:43577", - "right": "issue:43697", + "left": "issue:29159", + "right": "issue:33357", "accept": false, - "reason": "BLIP2 dtype propagation and RTDetrV2 output changes are different bugs." + "reason": "Unrelated: slow/fast tokenizer inconsistency vs MacOS bus error." }, { "left": "issue:44464", "right": "issue:44743", "accept": false, - "reason": "Chunked generation with compiled forward and recurrent state reset in modular_qwen3_5 are different code paths and symptoms." + "reason": "Both touch generation/cache behavior, but one is compiled forward chunking and the other is Qwen3.5 recurrent-state reset." } ] }, @@ -6924,8 +6925,8 @@ "issue:43531|issue:43742", "issue:44514|issue:45290", "issue:43646|issue:43749", - "issue:43653|issue:43927", "issue:43688|issue:45440", + "issue:43653|issue:43927", "issue:43688|issue:45704", "issue:43756|issue:46402", "issue:45491|issue:46002", @@ -6935,245 +6936,251 @@ "issue:44479|issue:44625", "issue:44464|issue:44977", "issue:43643|issue:44561", - "issue:32090|issue:34567", "issue:41720|issue:44464", + "issue:32090|issue:34567", "issue:44977|issue:46209", - "issue:43577|issue:44554", + "issue:43742|issue:44821", "issue:43504|issue:43720", + "issue:43577|issue:44554", "issue:33666|issue:36010", + "issue:43577|issue:45588", "issue:44466|issue:46050", - "issue:32090|issue:33666", - "issue:43742|issue:44821", "issue:43504|issue:43550", - "issue:43577|issue:45588", + "issue:32090|issue:33666", "issue:43504|issue:44610", + "issue:43452|issue:44451", "issue:43720|issue:43927", - "issue:43262|issue:43295", "issue:44554|issue:46365", - "issue:43452|issue:44451", + "issue:43262|issue:43295", "issue:43606|issue:43927", "issue:45823|issue:46050" ], - "prepared_review_unit_hash": "b32b8d0b19a7b18c6de47fe0977574e2c079fee81a6e8cb231f297f01ab75944", + "prepared_review_unit_hash": "abe7bd906f64f1cebcf2efc5226a862020c0b98e69a1155d4083ef10de364f64", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37596, - "estimated_input_tokens": 9399, - "estimated_eval_tokens": 19054 + "serialized_chars": 37598, + "estimated_input_tokens": 9400, + "estimated_eval_tokens": 19056 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is highly heterogeneous: most items are unrelated issue reports that only share a broad subsystem (tokenization, chat templates, loading, MoE, etc.). I found no soft-edge pair that is clearly the same underlying bug/change, so no merges are justified from the provided evidence.", - "confidence": 0.64, - "canonical_issue_reason": "issue:45701 is the broadest representative in the set and has the strongest cross-reference signal, but this is only a weak fit because the cluster is not a true duplicate set.", + "summary": "This cluster is highly heterogeneous: tokenizer/config loading regressions, multimodal processor crashes, training bugs, and model-specific inference issues. None of the soft pairs looks like the same underlying defect with a single mergeable fix, so there is no safe duplicate canonicalization here.", + "confidence": 0.92, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:45701 is the best single issue to represent the cluster by breadth and inbound references, though the underlying reports are too diverse for a clean canonical duplicate.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Both involve tokenizer loading/patching, but one is an offline Mistral regex issue and the other is a specific tokenizer load failure for EMBEDDIA/est-roberta." + "reason": "Both are tokenizer-loading regressions, but they fail for different reasons and code paths." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Different models and failures: Qwen3-MoE sliding_window behavior vs a key error while loading MobileLLM-125M." + "reason": "Different models and symptoms: Qwen3-MoE sliding-window behavior vs MobileLLM key error." }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Both hit apply_chat_template, but the triggers differ: batched padding=False vs assistant tool-call messages with no content." + "reason": "Same `apply_chat_template` area, but distinct crash triggers and inputs." }, { "left": "issue:43646", "right": "issue:43749", "accept": false, - "reason": "These are unrelated regressions: custom model initialization under v5.0 vs FSDP CPU RAM efficient loading." + "reason": "Custom model initialization breakage and FSDP CPU RAM-efficient loading are unrelated." }, { - "left": "issue:43653", - "right": "issue:43927", + "left": "issue:43688", + "right": "issue:45440", "accept": false, - "reason": "Tokenizer special-token registration bug vs config save/load losing custom token IDs; similar area, different defect." + "reason": "Separate MoE correctness bugs in different architectures and computations." }, { - "left": "issue:43688", - "right": "issue:45440", + "left": "issue:43653", + "right": "issue:43927", "accept": false, - "reason": "Both are MoE-related, but one is auxiliary-loss normalization and the other is a DeepSeekV3MoE implementation divergence." + "reason": "Tokenizer special-token registration bug vs config field persistence bug." }, { "left": "issue:43688", "right": "issue:45704", "accept": false, - "reason": "Completely different problems: MoE loss normalization vs a T5/apex RMSNorm memory leak." + "reason": "Aux-loss normalization and RMSNorm memory leak are unrelated failures." }, { "left": "issue:43756", "right": "issue:46402", "accept": false, - "reason": "Different model families and failures: Smollm3 RoPE-layer handling vs Sam3 text encoder weights not loading." + "reason": "Different models and load-time behaviors; no shared underlying defect." }, { "left": "issue:45491", "right": "issue:46002", "accept": false, - "reason": "One is a Gemma3 sliding-window NaN bug; the other is NaNs in classification heads after checkout/init." + "reason": "Both involve NaNs, but in different components and with different triggers." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "One is a torch.compile/SDPA failure for Bamba-9B-v2; the other is a config/token-ID persistence bug." + "reason": "Bamba compile/SDPA failure is unrelated to DiaConfig token-ID loss." }, { "left": "issue:43688", "right": "issue:45701", "accept": false, - "reason": "Different issue types: MoE loss normalization vs a general tokenization-change report." + "reason": "Related broadly to tokenization/model behavior, but not the same bug." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "Unrelated failures: Bamba torch.compile+SDPA vs an MLukeTokenizer AttributeError on tasks." + "reason": "Different model/runtime issues with no common code path." }, { "left": "issue:44479", "right": "issue:44625", "accept": false, - "reason": "Video-input regression across Qwen VL variants vs num_labels propagation in Qwen3.5 config." + "reason": "Video-input regression and num_labels propagation are separate defects." }, { "left": "issue:44464", "right": "issue:44977", "accept": false, - "reason": "Different generation bugs: compiled-forward chunked generation inconsistency vs flash-attention behavior in Qwen3.5." + "reason": "Both affect generation, but compiled-forward inconsistency and flash-attention failure differ." }, { "left": "issue:43643", "right": "issue:44561", "accept": false, - "reason": "Both mention trust_remote_code, but one is missing fields in AutoConfig and the other is removal of is_torch_fx_available." + "reason": "Both hit `trust_remote_code` users, but they are different v5 breakages." }, { - "left": "issue:32090", - "right": "issue:34567", + "left": "issue:41720", + "right": "issue:44464", "accept": false, - "reason": "Trainer broadcast TypeError vs TrainerState.num_input_tokens_seen not updating; unrelated Trainer bugs." + "reason": "Auto device-map assert on A800 is unrelated to chunked-generation compilation issues." }, { - "left": "issue:41720", - "right": "issue:44464", + "left": "issue:32090", + "right": "issue:34567", "accept": false, - "reason": "A800 auto device-mapping cudaErrorAssert is unrelated to compiled-forward chunked generation inconsistency." + "reason": "Trainer broadcast TypeError and token-count tracking are separate training bugs." }, { "left": "issue:44977", "right": "issue:46209", "accept": false, - "reason": "Both are inference regressions, but one is Qwen3.5 flash-attention behavior and the other is FP8 inference on B200." + "reason": "Flash-attention generation bug vs FP8/B200 inference failure are distinct." }, { - "left": "issue:43577", - "right": "issue:44554", + "left": "issue:43742", + "right": "issue:44821", "accept": false, - "reason": "Different layers and devices: dtype propagation in Blip2 vs an MPS attention correctness issue." + "reason": "Model loading key error and AutoImageProcessor URL loading are different problems." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Both are loading-time issues, but one is Beit legacy-field handling and the other is BitNet packed-weight unpacking." + "reason": "Legacy-field load failure vs BitNet packed-weight handling are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype propagation and MPS attention correctness are different issues." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Qwen2-VL multi-GPU training is unrelated to a GenerationMixin import error." + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import error are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "Dtype-loading bug vs flash_attention aux-state crash are different code paths." }, { "left": "issue:44466", "right": "issue:46050", "accept": false, - "reason": "Both touch save/load behavior, but one is tied-weight serialization and the other is sinq quantization persistence." - }, - { - "left": "issue:43742", - "right": "issue:44821", - "accept": false, - "reason": "Different loader paths: MobileLLM key error vs AutoImageProcessor failing to load from URL." + "reason": "Both involve save/load, but tied-weight serialization and sinq quantization are different defects." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "Legacy-field loading bug vs torch.compile/SDPA failure are different code paths." + "reason": "Semantic-segmentation preset loading and Bamba compile/SDPA failures are unrelated." }, { - "left": "issue:43577", - "right": "issue:45588", + "left": "issue:32090", + "right": "issue:33666", "accept": false, - "reason": "Different failures: dtype not applied to Blip2 components vs flash_attention.py crashing on s_aux=None." + "reason": "Trainer GPU broadcast error and Qwen2-VL multi-GPU training request are not the same bug." }, { "left": "issue:43504", "right": "issue:44610", "accept": false, - "reason": "Unrelated preprocessing/model-metadata bugs: Beit legacy field vs OmDet-Turbo image size mismatch." + "reason": "Beit legacy-field loading and OmDet-Turbo processor size mismatch are separate issues." }, { - "left": "issue:43720", - "right": "issue:43927", + "left": "issue:43452", + "right": "issue:44451", "accept": false, - "reason": "Packed-weight loading failure vs custom token-ID persistence; different models and causes." + "reason": "GGUF/pretrained loading breakage and ScandiBERT tokenizer/model loading failure are not the same defect." }, { - "left": "issue:43262", - "right": "issue:43295", + "left": "issue:43720", + "right": "issue:43927", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting is unrelated to processor.tokenizer/image-passing regression." + "reason": "BitNet weight unpacking and DiaConfig token-ID persistence are unrelated." }, { "left": "issue:44554", "right": "issue:46365", "accept": false, - "reason": "MPS attention correctness bug vs Trainer import failure tied to datasets; no shared underlying defect." + "reason": "MPS attention correctness and Trainer import shadowing are different problems." }, { - "left": "issue:43452", - "right": "issue:44451", + "left": "issue:43262", + "right": "issue:43295", "accept": false, - "reason": "General model/tokenizer loading with gguf_file is not the same as a specific ScandiBERT loading regression." + "reason": "Audio chat-template sampling-rate handling and processor.tokenizer regression are separate bugs." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch for bark-small is unrelated to custom token-ID loss on config save/load." + "reason": "CPU-offload device mismatch in bark-small is unrelated to DiaConfig token-ID loss." }, { "left": "issue:45823", "right": "issue:46050", "accept": false, - "reason": "Device-map auto forward mismatch in Gemma4 PLE is unrelated to sinq quantized save/load breakage." + "reason": "Gemma4 device mismatch and sinq save/load breakage are different issues." } ] }, @@ -7215,7 +7222,6 @@ "issue:43606", "issue:43646", "issue:43688", - "issue:43701", "issue:43720", "issue:43824", "issue:43825", @@ -7241,16 +7247,17 @@ "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:44849|issue:46209", + "issue:33666|issue:36331", "issue:43525|issue:43572", + "issue:44849|issue:46209", "issue:43278|issue:43688", "issue:43720|issue:44361", "issue:43493|issue:43825", @@ -7268,231 +7275,237 @@ "issue:43278|issue:43844", "issue:43646|issue:44466", "issue:44464|issue:46424", - "issue:33357|issue:33666", "issue:43334|issue:43572", + "issue:33357|issue:33666", "issue:44589|issue:45520", "issue:43257|issue:43299", "issue:43334|issue:45759", "issue:43824|issue:44589", "issue:43927|issue:44361", "issue:43504|issue:43606", - "issue:43278|issue:43701", - "issue:43572|issue:44291" + "issue:43572|issue:44291", + "issue:45229|issue:45823" ], - "prepared_review_unit_hash": "6ee2d9e2d74b37ec63a15f3ac454f62d10552a5dfa1ad1e5422c53a07e7d8ec4", + "prepared_review_unit_hash": "f0f03347cf8c4bbd0b278d1da4f5ce59c1410e47420af8a8bc3dbb5133d681d2", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37079, - "estimated_input_tokens": 9270, - "estimated_eval_tokens": 18796 + "serialized_chars": 37116, + "estimated_input_tokens": 9279, + "estimated_eval_tokens": 18814 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a heterogeneous set of Transformers issue reports across tokenization, config loading, multimodal models, and training/runtime regressions. I don\u2019t see any soft pair that is clearly the same underlying bug/change, so no duplicates are accepted.", - "confidence": 0.74, - "canonical_issue_reason": "Issue 45812 is the closest thing to a canonical representative: it describes a broad, user-visible tokenizer regression that affects an entire model family, which is more central than the many narrow model-specific failures in the set.", + "summary": "This cluster is mostly heterogeneous: the items span many unrelated model, tokenizer, config, and loading regressions. None of the soft pairs looks like a safe duplicate merge, though a few share broad themes (for example, v5 config regressions or Qwen/SigLIP2 family issues).", + "confidence": 0.77, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 45812 is the strongest triage candidate overall because it is concrete, reproducible, and has the widest apparent blast radius among the listed reports.", + "best_issue_reason": "Issue 45812 is the broadest single regression here: it describes a silent tokenizer corruption affecting all Granite models and has clear user impact, so it is the strongest representative issue if one must be chosen.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:33666", - "right": "issue:36331", - "accept": false, - "reason": "Different failures: Qwen2-VL multi-GPU training vs a CustomTrainer compute_loss signature mismatch." - }, { "left": "issue:43994", "right": "issue:46192", "accept": false, - "reason": "Different bugs: SigLIP2 inference mismatch vs return_tensors being ignored when text_kwargs is passed." + "reason": "Different bugs: SigLIP2 wrong outputs versus a tokenizer kwargs/return_tensors handling issue." }, { - "left": "issue:44849", - "right": "issue:46209", + "left": "issue:33666", + "right": "issue:36331", "accept": false, - "reason": "Unrelated: Qwen3.5 hidden-state behavior vs FP8 inference on B200." + "reason": "Different areas and failures: Qwen2-VL multi-GPU training versus CustomTrainer.compute_loss signature mismatch." }, { "left": "issue:43525", "right": "issue:43572", "accept": false, - "reason": "Both are missing-config-attribute regressions, but for different model families and different fields; not the same bug." + "reason": "Both are v5 padding-token config regressions, but they affect different models and different missing fields; not the same concrete bug." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Unrelated issues: Qwen3.5 hidden-state behavior versus FP8 inference failure on B200." }, { "left": "issue:43278", "right": "issue:43688", "accept": false, - "reason": "Different code paths: dtype drift in evaluation vs auxiliary-loss normalization." + "reason": "Different failure modes: dtype drift in evaluation versus auxiliary-loss normalization." }, { "left": "issue:43720", "right": "issue:44361", "accept": false, - "reason": "Unrelated: packed-weight loading in BitNet vs a tokenizer AttributeError in MLukeTokenizer." + "reason": "Different subsystems: accelerate loading of packed BitNet weights versus tokenizer/task AttributeError." }, { "left": "issue:43493", "right": "issue:43825", "accept": false, - "reason": "Different subsystems: SigLIP2 implementation discrepancy vs pipeline translation-task messaging." + "reason": "Both mention model/pipeline behavior, but SigLIP2 implementation mismatch and pipeline translation messaging are unrelated." }, { "left": "issue:44625", "right": "issue:46002", "accept": false, - "reason": "Different symptoms and root causes: num_labels propagation vs NaNs in classification heads." + "reason": "Different bugs: missing num_labels propagation versus NaNs in classification heads after checkout/init." }, { "left": "issue:28218", "right": "issue:33357", "accept": false, - "reason": "No overlap beyond being issues in the library; tokenizer spacing and Mac bus error are unrelated." + "reason": "Tokenizer spacing bug versus a MacOS bus error when loading a CLIP model; no shared underlying code path is evident." }, { "left": "issue:39692", "right": "issue:43493", "accept": false, - "reason": "Same model family, but one is doc-example mistakes and the other is an implementation discrepancy; not the same bug." + "reason": "Both are SigLIP2-related, but one is documentation/example errors while the other is an implementation-vs-JAX discrepancy." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "Model loading failure vs a dependency version bump issue; unrelated." + "reason": "Different topics: Phi-4 multimodal loading versus an unreleased PEFT version bump." }, { "left": "issue:44849", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 bugs: hidden states vs incorrect visual encoder keys on save." + "reason": "Different Qwen3.5 regressions: hidden states output versus incorrect save_pretrained visual encoder keys." }, { "left": "issue:44568", "right": "issue:45812", "accept": false, - "reason": "Both tokenizer regressions, but different models and different behaviors (special token insertion vs wrong token IDs)." + "reason": "Tokenizer special-token insertion bug versus Granite token-id regression; same broad area, but not the same concrete issue." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "Unrelated: SigLIP2 discrepancy vs gradient growth under DeepSpeed ZeRO-3." + "reason": "Different bugs: SigLIP2 discrepancy versus gradient blow-up with HfDeepSpeedConfig + ZeRO-3." }, { "left": "issue:39692", "right": "issue:43295", "accept": false, - "reason": "Documentation/example errors are not the same as the processor.tokenizer regression." + "reason": "Different issue types: SigLIP2 doc-example problems versus processor/tokenizer regression in v4.57.5." }, { "left": "issue:44488", "right": "issue:44843", "accept": false, - "reason": "Different failures: model loading for cjvt/sleng-bert vs offline model_info lookup in tokenizer patching." + "reason": "Both concern loading/tokenizer behavior, but one is a specific model load failure and the other is an offline-mode regression in _patch_mistral_regex." }, { "left": "issue:45200", "right": "issue:45229", "accept": false, - "reason": "Different issues: missing mm_token_type_ids defaulting vs GPU OOM in Gemma4 inference." + "reason": "Different Gemma4 problems: missing mm_token_type_ids default versus multi-GPU CUDA OOM." }, { "left": "issue:43540", "right": "issue:44977", "accept": false, - "reason": "Different Qwen model bugs: video input processing vs flash-attention generation behavior." + "reason": "Different Qwen variants and different bugs: video-input ValueError versus flash-attention generation issue." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Different problem domains: evaluation dtype change vs DeepSpeed gradient instability." + "reason": "Different regressions: embedding dtype drift versus gradient instability under ZeRO-3." }, { "left": "issue:43646", "right": "issue:44466", "accept": false, - "reason": "Custom model initialization breakage and tied-weight serialization are separate failure modes." + "reason": "Both are v5-related, but one is custom model initialization and the other is serialization of tied weights; not the same code-path bug." }, { "left": "issue:44464", "right": "issue:46424", "accept": false, - "reason": "Chunked generation inconsistency is not the same as cache max length being ignored except for quantized cache." + "reason": "Different topics: compiled-forward chunked generation versus cache-length handling for quantized cache." }, { - "left": "issue:33357", - "right": "issue:33666", + "left": "issue:43334", + "right": "issue:43572", "accept": false, - "reason": "No shared underlying bug: Mac bus error on CLIP vs Qwen2-VL multi-GPU training." + "reason": "Both are missing config attributes after v5-style changes, but they concern different models and different fields." }, { - "left": "issue:43334", - "right": "issue:43572", + "left": "issue:33357", + "right": "issue:33666", "accept": false, - "reason": "Similar missing-attribute theme, but different models, fields, and load paths." + "reason": "Completely unrelated: MacOS bus error on CLIP versus Qwen2-VL training." }, { "left": "issue:44589", "right": "issue:45520", "accept": false, - "reason": "Unrelated low-level errors: float8 storage lookup vs flash_attn import on Python 3.13." + "reason": "Different runtime failures: missing Float8 storage type versus flash_attn import_utils KeyError on Python 3.13." }, { "left": "issue:43257", "right": "issue:43299", "accept": false, - "reason": "Both involve Qwen MoE, but one is accelerate/deepspeed weight conversion and the other is model loading in dev; not mergeable as one fix." + "reason": "Both are Qwen MoE loading issues, but one is accelerate+deepspeed conversion and the other is generic Qwen3VL loading; not a single duplicate bug." }, { "left": "issue:43334", "right": "issue:45759", "accept": false, - "reason": "Different root causes: missing pad_token_id on Qwen3-VL text config vs not unwrapping text_config in composite Qwen multimodal configs." + "reason": "Different config/model-init regressions: missing pad_token_id versus AutoModelForCausalLM not unwrapping text_config." }, { "left": "issue:43824", "right": "issue:44589", "accept": false, - "reason": "ImportError for a class and float8 storage resolution are unrelated." + "reason": "Unrelated: import failure for a Qwen2.5 VL class versus a float8 storage lookup error." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "Different issues: DiaConfig save/load token IDs vs MLukeTokenizer task AttributeError." + "reason": "Different model components and symptoms: DiaConfig save/load corruption versus MLukeTokenizer AttributeError." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "Legacy-field load failure and CPU offload device mismatch are unrelated." + "reason": "Different bugs: legacy-field loading failure versus CPU-offload device mismatch." }, { - "left": "issue:43278", - "right": "issue:43701", + "left": "issue:43572", + "right": "issue:44291", "accept": false, - "reason": "Different bugs: eval dtype regression vs resume_from_checkpoint key mismatch." + "reason": "Both are initialization/config regressions, but one is missing pad_token_idx and the other is an unexpected _is_hf_initialized argument; separate fixes." }, { - "left": "issue:43572", - "right": "issue:44291", + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different compatibility problems in different modules; no shared underlying bug is apparent." + }, + { + "left": "issue:45229", + "right": "issue:45823", "accept": false, - "reason": "Both are loading-related regressions, but the missing pad_token_idx problem is separate from the unexpected _is_hf_initialized argument issue." + "reason": "Both involve Gemma4 inference, but one is OOM in multi-GPU inference and the other is a PLE device mismatch during forward." } ] }, @@ -7526,6 +7539,7 @@ "issue:38175", "issue:42490", "issue:43065", + "issue:43278", "issue:43504", "issue:43540", "issue:43575", @@ -7533,6 +7547,7 @@ "issue:43606", "issue:43653", "issue:43696", + "issue:43701", "issue:43844", "issue:43873", "issue:43994", @@ -7550,7 +7565,6 @@ "issue:45084", "issue:45198", "issue:45200", - "issue:45229", "issue:45290", "issue:45405", "issue:45406", @@ -7565,11 +7579,11 @@ "issue:46369" ], "soft_pairs": [ - "issue:45229|issue:45823", + "issue:43278|issue:43701", "issue:43575|issue:45405", "issue:44792|issue:45405", - "issue:43504|issue:43653", "issue:44493|issue:46285", + "issue:43504|issue:43653", "issue:44464|issue:45584", "issue:43606|issue:44112", "issue:43994|issue:44493", @@ -7579,10 +7593,10 @@ "issue:44423|issue:44734", "issue:29942|issue:30333", "issue:42490|issue:43696", - "issue:43844|issue:44964", "issue:28218|issue:34567", - "issue:29159|issue:32090", + "issue:43844|issue:44964", "issue:45406|issue:46121", + "issue:29159|issue:32090", "issue:45084|issue:45520", "issue:44554|issue:45446", "issue:15354|issue:38175", @@ -7592,184 +7606,190 @@ "issue:45200|issue:45823", "issue:44521|issue:45290" ], - "prepared_review_unit_hash": "b90bc5648f01c4763c9d2abd47ec7c923b5146031c3476af13ae6da29ae32053", + "prepared_review_unit_hash": "c50cc905705dafa473a8007346fa398868ac8aa9bee0e2dd0d8077408ecea836", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 26, - "serialized_chars": 34947, - "estimated_input_tokens": 8737, - "estimated_eval_tokens": 17730 + "serialized_chars": 35505, + "estimated_input_tokens": 8877, + "estimated_eval_tokens": 18010 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is a heterogeneous grab bag of unrelated bugs across tokenization, serving, multimodal models, quantization, and CI. None of the soft pairs look like the same underlying issue/change, so no duplicate merges are justified.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "The cluster is mostly heterogeneous; none of the soft pairs look like true duplicates under a strict same-bug standard. The most representative issue is the multimodal chat-template bug (#44521), but it is not a duplicate of the other candidates.", + "confidence": 0.93, + "canonical_issue_reason": "#44521 is the broadest, most actively discussed issue here and has the clearest ongoing user impact (multimodal `apply_chat_template` behavior). It is the best representative issue for this otherwise mixed cluster.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "#44521 is the strongest cluster head candidate because it is open, has multiple inbound references, and concerns a concrete, user-facing multimodal template failure rather than a niche or one-off regression.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45229", - "right": "issue:45823", + "left": "issue:43278", + "right": "issue:43701", "accept": false, - "reason": "Both involve Gemma4, but one is a multi-GPU OOM and the other is a device-mapping forward mismatch; different failures and fixes." + "reason": "Different problems: dtype drift in evaluation vs a checkpoint-resume key mismatch." }, { "left": "issue:43575", "right": "issue:45405", "accept": false, - "reason": "Completely different problems: tensor-parallel OOM versus an unreleased PEFT version requirement." + "reason": "OOM during tensor-parallel model load is unrelated to a PEFT version bump / release-availability issue." }, { "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "A Janus test failure and a dependency/version bump are unrelated." + "reason": "A failing janus test is not the same bug as the PEFT minimum-version mismatch." }, { - "left": "issue:43504", - "right": "issue:43653", + "left": "issue:44493", + "right": "issue:46285", "accept": false, - "reason": "Different subsystems and bug types: legacy field loading for Beit versus BigBird tokenizer special-token registration." + "reason": "Unexpected position-id keys across many models is not the same as CLIP text-model structure inconsistency." }, { - "left": "issue:44493", - "right": "issue:46285", + "left": "issue:43504", + "right": "issue:43653", "accept": false, - "reason": "Both mention model structure/configuration regressions, but the concrete problems are different and not the same code-path bug." + "reason": "Legacy-field loading bug in BEiT segmentation is unrelated to BigBird tokenizer special-token registration." }, { "left": "issue:44464", "right": "issue:45584", "accept": false, - "reason": "Chunked generation with compiled forward is unrelated to Whisper empty-transcription handling after align_special_tokens." + "reason": "Compiled-forward chunked generation mismatch is a different code path from empty-transcription failure after `align_special_tokens`." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Both touch device behavior, but one is a real CPU offload mismatch and the other is a stale GraniteSpeech CI test." + "reason": "CPU offload device mismatch in Bark is not the same as a stale CI test in GraniteSpeech." }, { "left": "issue:43994", "right": "issue:44493", "accept": false, - "reason": "Siglip2 nonsensical outputs and unexpected position-id keys are different model/config regressions." + "reason": "Nonsensical SigLIP outputs are unrelated to unexpected position-id keys." }, { "left": "issue:43873", "right": "issue:46050", "accept": false, - "reason": "Quantization/offloading interaction is not the same bug as broken save/load round-tripping for sinq." + "reason": "Quantization/offloading behavior is different from broken save/load of a sinq-quantized model." }, { "left": "issue:45588", "right": "issue:46050", "accept": false, - "reason": "Flash-attention sinkless-model crash and quantized save/load corruption are unrelated." + "reason": "Flash-attention sinkless-model crash is not the same as quantized model persistence failure." }, { "left": "issue:43065", "right": "issue:46369", "accept": false, - "reason": "Dummy Conv2d implementation in Sam3PixelDecoder is not the same as a memory overflow in sam3-litetext." + "reason": "A dummy Conv2d implementation in Sam3PixelDecoder is different from a Sam3 LiteText memory overflow." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Both are serving crashes, but they occur in different request paths with different root causes." + "reason": "Both involve serve, but one is a multimodal `.to` crash and the other is KV-cache indexing; different bugs." }, { "left": "issue:29942", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 CI failures and MLFlow reporting on failed jobs are unrelated." + "reason": "Flash Attention 2 test failures are unrelated to MLFlow reporting after failed training jobs." }, { "left": "issue:42490", "right": "issue:43696", "accept": false, - "reason": "Local-path import warnings have nothing to do with GPT-oss CUDA OOM." - }, - { - "left": "issue:43844", - "right": "issue:44964", - "accept": false, - "reason": "DeepSpeed gradient growth and Phi-4 multimodal load failure are different bugs." + "reason": "Path-with-slash import warning is unrelated to GPU OOM for GPT-oss-20b." }, { "left": "issue:28218", "right": "issue:34567", "accept": false, - "reason": "Tokenizer spacing after added tokens is unrelated to TrainerState token counting." + "reason": "Tokenizer added-space behavior is not the same as trainer input-token accounting." }, { - "left": "issue:29159", - "right": "issue:32090", + "left": "issue:43844", + "right": "issue:44964", "accept": false, - "reason": "Slow/fast tokenizer inconsistency is unrelated to a Trainer GPU broadcast NoneType error." + "reason": "DeepSpeed ZeRO-3 gradient growth and Phi-4 multimodal loading failure are different issues." }, { "left": "issue:45406", "right": "issue:46121", "accept": false, - "reason": "Gemma4 serve processor missing _tokenizer and rope-params list handling are different issues." + "reason": "Gemma4Processor missing `_tokenizer` is unrelated to rope-parameter list type handling." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Slow-vs-fast tokenizer inconsistency is not the same as a trainer broadcast TypeError." }, { "left": "issue:45084", "right": "issue:45520", "accept": false, - "reason": "Non-template node compilation and Python 3.13 flash_attn import mapping are unrelated." + "reason": "Template-node compilation failure is unrelated to a Python 3.13 flash_attn import KeyError." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "MPS attention correctness and a PyTorch version guard for AuxRequest import are different concerns." + "reason": "MPS attention correctness when head dims differ is not the same as a PyTorch version check bug in flex attention import." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp export failure and Siglip2 zero probabilities do not share a code path." + "reason": "TorchScript export support for GeneratorExp is unrelated to SigLIP zero-probability outputs." }, { "left": "issue:43582", "right": "issue:44977", "accept": false, - "reason": "Apple Silicon allocator warmup TypeError and Qwen3.5 flash-attention generation issues are unrelated." + "reason": "Apple Silicon allocator TypeError is unrelated to Qwen3.5 flash-attention generation issues." }, { "left": "issue:43540", "right": "issue:45200", "accept": false, - "reason": "Qwen3OmniMoe video processing error and Gemma4 mm_token_type_ids defaulting are separate multimodal bugs." + "reason": "Qwen3OmniMoe video-processing ValueError is unrelated to Gemma4 token-type defaults for text-only fine-tuning." }, { "left": "issue:45072", "right": "issue:45198", "accept": false, - "reason": "bfloat16 dtype mismatches and Wav2Vec2 save/tokenization failures are unrelated." + "reason": "bfloat16 dtype mismatches in inference are unrelated to Wav2Vec2 save/tokenization failures." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Missing default `mm_token_type_ids` for text-only fine-tuning is not the same as a PLE device mismatch during forward." }, { "left": "issue:44521", "right": "issue:45290", "accept": false, - "reason": "Both involve apply_chat_template, but one is all-zero assistant masks for multimodal inputs while the other is a crash on tool-call assistant messages; different underlying bugs." + "reason": "Both involve `apply_chat_template`, but one is zero assistant masks for multimodal inputs and the other is a crash on tool-call messages without content; not the same bug." } ] }, @@ -7798,10 +7818,10 @@ "issue:37428", "issue:38175", "issue:39692", + "issue:41720", "issue:42490", "issue:42915", "issue:43232", - "issue:43278", "issue:43295", "issue:43334", "issue:43388", @@ -7815,7 +7835,6 @@ "issue:43701", "issue:43824", "issue:43844", - "issue:43856", "issue:43873", "issue:43950", "issue:43957", @@ -7851,17 +7870,17 @@ "issue:43295|issue:43450", "issue:45381|issue:46209", "issue:43295|issue:44610", - "issue:43295|issue:43873", "issue:43572|issue:44987", + "issue:43295|issue:43873", "issue:39692|issue:46192", "issue:44554|issue:45356", "issue:43572|issue:43957", - "issue:45520|issue:46178", "issue:37428|issue:43824", + "issue:45520|issue:46178", "issue:44466|issue:45127", - "issue:43550|issue:43844", "issue:43950|issue:46285", "issue:44898|issue:45412", + "issue:43550|issue:43844", "issue:43450|issue:43575", "issue:45200|issue:45325", "issue:43334|issue:43525", @@ -7872,216 +7891,216 @@ "issue:44792|issue:45563", "issue:38175|issue:43493", "issue:28218|issue:35141", - "issue:43278|issue:43856" + "issue:41720|issue:46209" ], - "prepared_review_unit_hash": "d11ef8187a582e3f4a86a5beba35e86c12993d339d276b0d8547616f638650a7", + "prepared_review_unit_hash": "814514b7bc1976e7083f8c4dd54813b4935ce5fc6863026158b00d255fe58c3b", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 30, - "serialized_chars": 37016, - "estimated_input_tokens": 9254, - "estimated_eval_tokens": 18764 + "serialized_chars": 36462, + "estimated_input_tokens": 9116, + "estimated_eval_tokens": 18488 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is not a duplicate cluster: it spans many unrelated bugs across tokenization, model loading, generation, multimodal processors, and config regressions. I rejected all soft edges because none clearly describe the same underlying fix or code path.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue is appropriate; the items are thematically broad but not the same bug, so there is no coherent duplicate target.", - "canonical_pr_reason": "No PRs are present in this set.", - "best_issue_reason": "No global best issue stands out as a representative duplicate source because the issues are unrelated and would not collapse into one fix.", - "best_pr_reason": "No PRs are present in this set.", + "summary": "This cluster is largely a false-positive similarity group: most items are unrelated regressions, import errors, tokenizer/processor bugs, or model-specific failures. I did not find any soft pair that clearly represents the same underlying issue, so no duplicate merge is warranted.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43295 is the closest thing to a representative bug report here because it is a concrete, well-scoped regression with decent detail/activity, but the overall set is too heterogeneous for a true canonical duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43295 is the most representative high-signal issue in the list, but it should be treated only as a loose exemplar rather than an umbrella for the cluster.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37428", "right": "issue:43957", "accept": false, - "reason": "Both are import/loading failures, but they involve different symbols and different code paths (flash attention helper vs model loading on meta device)." + "reason": "Different failures: a missing internal flash-attention symbol import versus a meta-device model loading regression." }, { "left": "issue:43493", "right": "issue:43701", "accept": false, - "reason": "SigLIP2 implementation discrepancy and resume_from_checkpoint key mismatch are unrelated bugs." + "reason": "SigLIP2 implementation/discrepancy bug and checkpoint resume key mismatch are unrelated code paths." }, { "left": "issue:44186", "right": "issue:46192", "accept": false, - "reason": "Both involve tokenizer/input handling, but one is a crash in LayoutLMv2 NER batching while the other is return_tensors being ignored when text_kwargs is passed." + "reason": "LayoutLMv2 NER/batched padding crash is not the same as return_tensors being ignored when text_kwargs is passed." }, { "left": "issue:43232", "right": "issue:44849", "accept": false, - "reason": "Generation kwargs after sync_gpus and Qwen3.5 output_hidden_states are different generation/model bugs." + "reason": "Generation state update after sync_gpus is unrelated to the Qwen3.5 output_hidden_states regression." }, { "left": "issue:45823", "right": "issue:46192", "accept": false, - "reason": "Gemma4 device-mismatch during forward is unrelated to return_tensors being ignored in processor kwargs." + "reason": "Gemma4 device_map=\"auto\" device mismatch is a different bug from text processor kwargs handling." }, { "left": "issue:43295", "right": "issue:43450", "accept": false, - "reason": "Processor.tokenizer/image passthrough regression and batched video shape handling are different multimodal processor issues." + "reason": "Processor/tokenizer regression with images is not the same as batched video shape handling." }, { "left": "issue:45381", "right": "issue:46209", "accept": false, - "reason": "Wrong Qwen2.5-VL vision position IDs and FP8 inference failure on B200 are unrelated." + "reason": "Qwen2.5-VL video position_ids bug and FP8 inference failure on B200 are distinct model/hardware issues." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Custom model processor regression and OmDet-Turbo resolution mismatch are not the same bug." + "reason": "These affect different multimodal paths: custom processor/tokenizer regression versus OmDet-Turbo processor input size mismatch." }, { - "left": "issue:43295", - "right": "issue:43873", + "left": "issue:43572", + "right": "issue:44987", "accept": false, - "reason": "Processor API regression and quantization/offloading behavior are unrelated." + "reason": "Missing pad_token_idx in StableLmConfig is unrelated to loading physical-intelligence/fast." }, { - "left": "issue:43572", - "right": "issue:44987", + "left": "issue:43295", + "right": "issue:43873", "accept": false, - "reason": "Missing pad_token_idx in StableLmConfig and a loading failure for physical-intelligence/fast do not share the same fix path." + "reason": "Processor/tokenizer regression and offloading/quantization behavior are different underlying problems." }, { "left": "issue:39692", "right": "issue:46192", "accept": false, - "reason": "SigLIP2 documentation errors are unrelated to the processor kwargs/return_tensors regression." + "reason": "Documentation example/model-processor mismatch is not the same as return_tensors precedence being ignored." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "MPS attention correctness and Kimi tokenizer codec handling are unrelated bugs." + "reason": "MPS attention correctness on mismatched head dims is unrelated to Kimi-K2.5 tokenizer codec/regex regression." }, { "left": "issue:43572", "right": "issue:43957", "accept": false, - "reason": "Both are model/config loading issues, but one is a missing config attribute and the other is a meta-device loading regression." - }, - { - "left": "issue:45520", - "right": "issue:46178", - "accept": false, - "reason": "Different import guard failures in different backends; not the same underlying issue." + "reason": "Both mention loading-related breakage, but one is a config field regression and the other is a meta-device model-loading bug." }, { "left": "issue:37428", "right": "issue:43824", "accept": false, - "reason": "Both are import errors, but they refer to different missing names and different subsystems/models." + "reason": "Both are import errors, but for different missing symbols/classes and different models." }, { - "left": "issue:44466", - "right": "issue:45127", + "left": "issue:45520", + "right": "issue:46178", "accept": false, - "reason": "Tied-weight serialization inconsistency and LoRA merge collapse are distinct failures." + "reason": "flash_attn backend import handling on Python 3.13 is unrelated to the tensorflow_text BACKENDS_MAPPING error." }, { - "left": "issue:43550", - "right": "issue:43844", + "left": "issue:44466", + "right": "issue:45127", "accept": false, - "reason": "torch.compile SDPA failure and gradient explosion with DeepSpeed ZeRO-3 are unrelated." + "reason": "Tied-weight serialization inconsistency and LoRA merge collapse after vocab extension are different symptoms and fixes." }, { "left": "issue:43950", "right": "issue:46285", "accept": false, - "reason": "Non-persistent buffer corruption in from_pretrained and CLIPTextModel structure inconsistency are different regression classes." + "reason": "Non-persistent buffer corruption in from_pretrained is unrelated to the CLIPTextModel structural change." }, { "left": "issue:44898", "right": "issue:45412", "accept": false, - "reason": "Perceiver positional interpolation failure and RT-DETR memory release are unrelated." + "reason": "Perceiver interpolation failure and RT-DETR memory release are unrelated model/runtime issues." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile+SDPA failure and DeepSpeed ZeRO-3 gradient inflation are different training/runtime bugs." }, { "left": "issue:43450", "right": "issue:43575", "accept": false, - "reason": "Video processor shape bug and tensor-parallel OOM on Qwen2 are not the same code path." + "reason": "Batched video processor shape bug is unrelated to tensor-parallel OOM during Qwen2-57B-A14B-Instruct loading." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Both concern multimodal inputs, but one is missing mm_token_type_ids defaults and the other is incorrect vision_position_ids scaling." + "reason": "Gemma 4 mm_token_type_ids defaults and Qwen2.5-VL still-image rope scaling are distinct multimodal bookkeeping issues." }, { "left": "issue:43334", "right": "issue:43525", "accept": false, - "reason": "Missing pad_token_id on different model configs is similar symptomatically but not the same concrete bug or fix." + "reason": "They share a missing pad_token_id symptom, but the affected model/config paths are different enough that this is not clearly one duplicate bug." }, { "left": "issue:42490", "right": "issue:43493", "accept": false, - "reason": "Path-trailing-slash import warning and SigLIP2 discrepancy are unrelated." + "reason": "Trailing-slash local path import warning and SigLIP2 JAX-vs-HF discrepancy are unrelated." }, { "left": "issue:42915", "right": "issue:43824", "accept": false, - "reason": "FineGrainedFP8Config failure and Qwen2.5-VL import error are different issues." + "reason": "FineGrainedFP8Config failure on Qwen3Moe is not the same as the Qwen2.5-VL import error." }, { "left": "issue:43824", "right": "issue:43975", "accept": false, - "reason": "Missing model import and incorrect detokenization are unrelated." + "reason": "Qwen2.5-VL import failure and DeepSeek tokenizer detokenization regression do not share the same underlying bug." }, { "left": "issue:43388", "right": "issue:43526", "accept": false, - "reason": "gather_for_metrics label truncation and BeitImageProcessorFast reduce_labels behavior are both label-related but not the same bug." + "reason": "gather_for_metrics label dropping and BeitImageProcessorFast reduce_labels truncation are different label-handling bugs." }, { "left": "issue:44792", "right": "issue:45563", "accept": false, - "reason": "A failing Janus image-generation test and a stale warning about num_return_sequences are unrelated." + "reason": "Janus image-generation test failure is unrelated to the stale num_return_sequences warning." }, { "left": "issue:38175", "right": "issue:43493", "accept": false, - "reason": "Unexpected zero probabilities in one model and a broader SigLIP2 implementation discrepancy are not clearly the same fix." + "reason": "Both involve SigLIP2, but one is a specific zero-probability symptom and the other is a broader implementation discrepancy; not clearly the same fix." }, { "left": "issue:28218", "right": "issue:35141", "accept": false, - "reason": "Tokenizer extra-space detokenization and embedding reinitialization after resize are unrelated." + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resizing with tie_word_embedding=False." }, { - "left": "issue:43278", - "right": "issue:43856", + "left": "issue:41720", + "right": "issue:46209", "accept": false, - "reason": "Embedding dtype drift in evaluate and Qwen3 MoE memory usage are different problems." + "reason": "cudaErrorAssert on A800 with auto device mapping is a different problem from FP8 inference failure on B200." } ] }, @@ -8109,8 +8128,8 @@ "issue:32090", "issue:37428", "issue:38175", - "issue:41720", "issue:43262", + "issue:43278", "issue:43335", "issue:43425", "issue:43450", @@ -8121,6 +8140,7 @@ "issue:43653", "issue:43819", "issue:43827", + "issue:43856", "issue:43901", "issue:44112", "issue:44206", @@ -8148,245 +8168,244 @@ "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:43278|issue:43856", "issue:43526|issue:46192", + "issue:43653|issue:44112", "issue:44442|issue:45198", "issue:43262|issue:43450", - "issue:45290|issue:46326", "issue:38175|issue:45412", + "issue:45290|issue:46326", "issue:45520|issue:46170", "issue:44610|issue:44857", "issue:45759|issue:46285", - "issue:43827|issue:43901", "issue:44514|issue:44521", "issue:44561|issue:46234", + "issue:43827|issue:43901", "issue:43819|issue:44112", "issue:44206|issue:45584", "issue:43819|issue:45083", "issue:44857|issue:45593", "issue:43335|issue:43550", "issue:37428|issue:43425", - "issue:43606|issue:45083", "issue:43611|issue:46402", + "issue:43606|issue:45083", "issue:29942|issue:32090", "issue:44448|issue:44743", - "issue:43335|issue:43653", "issue:45200|issue:45588", + "issue:43335|issue:43653", "issue:45127|issue:45245", - "issue:44855|issue:45083", "issue:37428|issue:44351", + "issue:44855|issue:45083", "issue:43550|issue:44112" ], - "prepared_review_unit_hash": "c917a97ad879a5a39a7b2d783bb2f73a6b8885f97068c777b94f7ceea54b3620", + "prepared_review_unit_hash": "daad55b14c05b72b690c9abb60a170c859b483ef4fabd1f06bbad275023545bf", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36619, - "estimated_input_tokens": 9155, - "estimated_eval_tokens": 18566 + "serialized_chars": 36632, + "estimated_input_tokens": 9158, + "estimated_eval_tokens": 18572 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: the pairings share only superficial vocabulary (flash attention, processor/template, model loading, or regression language) but describe different code paths and symptoms. None of the soft edges look like true duplicates or mergeable PR-equivalents.", - "confidence": 0.91, - "canonical_issue_reason": null, + "summary": "This is a heterogeneous cluster of distinct bugs and docs issues, not a true duplicate set. All soft pairs look superficially related by subsystem or API name, but they point to different code paths, models, or failure modes.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits: the items span unrelated bugs across import utils, chat templates, processors, model loading, training, and CI, so there is no real duplicate cluster to collapse.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "No global best issue: any one item would be an arbitrary representative rather than a true canonical duplicate target.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41720", - "right": "issue:46209", + "left": "issue:43278", + "right": "issue:43856", "accept": false, - "reason": "Both mention CUDA/FP8-style inference failures, but one is a Qwen3 auto device-mapping assert on A800 and the other is FP8 inference on B200; different hardware, failure mode, and likely code paths." + "reason": "Different bugs: dtype drift in evaluation vs Qwen3 MoE memory usage." }, { - "left": "issue:43653", - "right": "issue:44112", + "left": "issue:43526", + "right": "issue:46192", "accept": false, - "reason": "Both are CI-related token/special-token problems, but one is a tokenizer decode issue from a missing registered mask token and the other is a stale device override test failure in GraniteSpeech." + "reason": "Different processor behaviors: label reduction vs return_tensors handling with text_kwargs." }, { - "left": "issue:43526", - "right": "issue:46192", + "left": "issue:43653", + "right": "issue:44112", "accept": false, - "reason": "One is a BeitImageProcessorFast bug in reduce_labels, the other is return_tensors being ignored when text_kwargs is passed; unrelated processor behaviors." + "reason": "Tokenizer special-token registration vs a GraniteSpeech CI device-override test failure." }, { "left": "issue:44442", "right": "issue:45198", "accept": false, - "reason": "Both involve tokenization/saving, but one is AutoTokenizer failing to load FastSpeech2ConformerTokenizer and the other is Wav2Vec2 save_pretrained/tokenization failure; different models and mechanisms." + "reason": "Both touch speech tokenization, but they describe different failures in different model classes." }, { "left": "issue:43262", "right": "issue:43450", "accept": false, - "reason": "One concerns audio processor chat-template sampling-rate defaults; the other is batched video processor output shape. Same general processor area, but not the same bug." + "reason": "Different modalities and paths: audio chat-template sampling rate vs batched video output shape." }, { - "left": "issue:45290", - "right": "issue:46326", + "left": "issue:38175", + "right": "issue:45412", "accept": false, - "reason": "Both are chat-template edge cases, but one crashes on assistant tool-call messages with no content and the other is a continue_final_message bug for mistral_common tokenizers; different inputs and logic." + "reason": "Unrelated issues: zero probabilities in SigLIP2 vs RT-DETR memory not released." }, { - "left": "issue:38175", - "right": "issue:45412", + "left": "issue:45290", + "right": "issue:46326", "accept": false, - "reason": "Zero probabilities in SigLIP2 is a model output issue; RT-DETR memory not released is a lifecycle/memory management issue. No shared underlying defect." + "reason": "Both involve chat templates, but one is a tokenize=True crash with tool calls and the other is continue_final_message on mistral_common tokenizers." }, { "left": "issue:45520", "right": "issue:46170", "accept": false, - "reason": "Both are import/runtime dependency issues on specific platforms, but one is a Python 3.13 flash_attn import KeyError and the other is Windows/ROCm torch.distributed dependency breakage." + "reason": "Different import/platform problems: flash_attn key lookup on Python 3.13 vs Windows ROCm/torch.distributed dependency failure." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "OmDet-Turbo processor input-size mismatch and LwDetrImageLoss AMP crash are unrelated: preprocessing/configuration vs numeric instability in loss computation." + "reason": "Different vision bugs: wrong processor resize output vs float16 AMP crash in a loss function." }, { "left": "issue:45759", "right": "issue:46285", "accept": false, - "reason": "One is about AutoModelForCausalLM.from_config not unwrapping text_config for multimodal Qwen configs; the other is CLIP text-model structural inconsistency. Both touch model wrappers, but not the same code path." - }, - { - "left": "issue:43827", - "right": "issue:43901", - "accept": false, - "reason": "Both are docs issues around pipeline behavior, but one is v5 removal of pipeline references in summarization/translation docs and the other is outdated return_all_scores mention in text-classification docs." + "reason": "Different model-loading regressions: unwrapping text_config vs CLIP text-model structure mismatch." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Both involve apply_chat_template on multimodal/batched inputs, but one crashes with padding=False on batched Qwen2_5_VLProcessor input and the other returns all-zero assistant masks; different failure surfaces." + "reason": "Both are multimodal chat-template bugs, but one is a batched padding=False crash and the other is all-zero assistant masks; not the same defect." }, { "left": "issue:44561", "right": "issue:46234", "accept": false, - "reason": "One is trust_remote_code breakage from removing is_torch_fx_available; the other is a missing is_falcon_mamba_ssm_available symbol. Both are import-utils regressions, but not the same missing API." + "reason": "Different missing utilities in import_utils; they are not the same removed symbol or code path." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues around pipeline API changes, but they affect different docs and different deprecations." }, { "left": "issue:43819", "right": "issue:44112", "accept": false, - "reason": "DAC.from_latents mismatch vs forward pass is a model-method consistency bug; GraniteSpeech stale device override is a CI test issue. No common underlying bug." + "reason": "DAC forward/from_latents mismatch is unrelated to a GraniteSpeech CI stale device override test." }, { "left": "issue:44206", "right": "issue:45584", "accept": false, - "reason": "LasrFeatureExtractor unsupported center arg and Whisper empty transcription after align_special_tokens are both preprocessing/generation issues, but they affect different components and symptoms." + "reason": "Different audio bugs: unsupported center arg vs empty-transcription generation after align_special_tokens." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC latent conversion mismatch and qwen3_omni_moe feature-length helper behavior are unrelated internal logic bugs." + "reason": "Different model-specific helpers and failure modes; no shared underlying bug." }, { "left": "issue:44857", "right": "issue:45593", "accept": false, - "reason": "LwDetrImageLoss AMP crash is a numeric/precision failure; D-FINE missing auxiliary losses when denoising is off is a training-loss configuration bug." + "reason": "Both are training-related vision issues, but one is an AMP/CUDA crash and the other is missing auxiliary losses." }, { "left": "issue:43335", "right": "issue:43550", "accept": false, - "reason": "SwitchTransformersConfig creating a sparse layer at zero sparse layers and Bamba torch.compile SDPA failure are both model-architecture issues, but not the same code path or defect." + "reason": "Unrelated: SwitchTransformers config creates the wrong layer structure; Bamba fails under torch.compile with SDPA." }, { "left": "issue:37428", "right": "issue:43425", "accept": false, - "reason": "One is a specific flash-attention helper import error; the other is a broad Torch 2.10 incompatibility report. Too general and not the same bug." + "reason": "One is a missing flash-attention import symbol; the other is a broad Torch 2.10 compatibility issue." }, { - "left": "issue:43606", - "right": "issue:45083", + "left": "issue:43611", + "right": "issue:46402", "accept": false, - "reason": "CPU offload device mismatch in bark-small and qwen3_omni_moe feature-length helper behavior are unrelated; one is device placement, the other is shape/length calculation." + "reason": "Both are loading regressions, but they concern different mechanisms and models." }, { - "left": "issue:43611", - "right": "issue:46402", + "left": "issue:43606", + "right": "issue:45083", "accept": false, - "reason": "base_model_prefix loading regression and Sam3Model text encoder weights not loading are both model-loading issues, but they concern different mechanisms and models." + "reason": "Device-mismatch on CPU offload is unrelated to qwen3_omni_moe feature-length calculations." }, { "left": "issue:29942", "right": "issue:32090", "accept": false, - "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are unrelated training/runtime failures with no shared code path." + "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are separate problems." }, { "left": "issue:44448", "right": "issue:44743", "accept": false, - "reason": "Pegasus v4/v5 output differences are a model behavior regression; qwen3_5 recurrent states resetting in cached seq_len>1 is a different state-management bug." + "reason": "A v4/v5 output regression in Pegasus is unrelated to recurrent-state reset logic in qwen3_5." }, { - "left": "issue:43335", - "right": "issue:43653", + "left": "issue:45200", + "right": "issue:45588", "accept": false, - "reason": "Sparse-layer creation in SwitchTransformersConfig and BigBirdTokenizer special-token registration are unrelated; one is config construction, the other tokenizer metadata." + "reason": "Different default/None-handling bugs in separate multimodal and flash-attention code paths." }, { - "left": "issue:45200", - "right": "issue:45588", + "left": "issue:43335", + "right": "issue:43653", "accept": false, - "reason": "Gemma 4 mm_token_type_ids defaulting issue and flash_attention sink-less models crashing on s_aux=None are different multimodal/attention bugs with different APIs." + "reason": "Sparse-layer creation and BigBird tokenizer special-token registration are unrelated." }, { "left": "issue:45127", "right": "issue:45245", "accept": false, - "reason": "LoRA merging with extended vocabulary and the 2^24 category limit are unrelated model-editing vs categorical-encoding constraints." + "reason": "LoRA/tied-embeddings collapse and category-count overflow are unrelated failures." }, { - "left": "issue:44855", - "right": "issue:45083", + "left": "issue:37428", + "right": "issue:44351", "accept": false, - "reason": "Python 3.13 IndentationError in torch.jit.script parsing and qwen3_omni_moe feature-length helper behavior are unrelated; syntax/parsing vs helper logic." + "reason": "Both are import errors, but they involve different missing names and different modules." }, { - "left": "issue:37428", - "right": "issue:44351", + "left": "issue:44855", + "right": "issue:45083", "accept": false, - "reason": "Both are import errors, but one is a missing flash-attention helper symbol and the other is missing HybridCache. Different APIs and likely fixes." + "reason": "Python 3.13 torch.jit parsing error is unrelated to qwen3_omni_moe output-length helper behavior." }, { "left": "issue:43550", "right": "issue:44112", "accept": false, - "reason": "Bamba torch.compile/SDPA failure and GraniteSpeech stale device override test failure are unrelated; one is runtime compilation, the other CI test staleness." + "reason": "Different areas entirely: torch.compile/SDPA model bug vs GraniteSpeech CI test flake." } ] }, @@ -8461,20 +8480,20 @@ ], "soft_pairs": [ "issue:43819|issue:44493", - "issue:43335|issue:43504", "issue:44206|issue:45823", + "issue:43335|issue:43504", "issue:43278|issue:43825", "issue:29127|issue:32090", "issue:45588|issue:45859", - "issue:39692|issue:43526", "issue:45290|issue:45381", + "issue:39692|issue:43526", "issue:45310|issue:45478", "issue:43335|issue:45083", "issue:39692|issue:44493", "issue:43720|issue:44112", "issue:44448|issue:44779", - "issue:43335|issue:44112", "issue:45459|issue:46234", + "issue:43335|issue:44112", "issue:43335|issue:45593", "issue:34689|issue:35141", "issue:43606|issue:44857", @@ -8483,30 +8502,30 @@ "issue:38175|issue:44493", "issue:44857|issue:45072", "issue:45920|issue:46192", - "issue:28218|issue:32090", "issue:44448|issue:44977", + "issue:28218|issue:32090", "issue:40444|issue:43526", - "issue:43701|issue:45083", "issue:44977|issue:45229", + "issue:43701|issue:45083", "issue:43746|issue:46209", "issue:32090|issue:33357", "issue:29942|issue:35141", "issue:43550|issue:45083" ], - "prepared_review_unit_hash": "5ee6ab5296588e352b1fe604e58425f350779808819675e0030e5dccea71f968", + "prepared_review_unit_hash": "9c4814659d41d6967dc3b91c56261becada391f2adbd752a64ea2600be81be40", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 37332, + "serialized_chars": 37329, "estimated_input_tokens": 9333, "estimated_eval_tokens": 18922 }, @@ -8515,210 +8534,210 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a collection of unrelated issue reports across different models and subsystems. The only clear duplicate pair is the repeated Qwen3.5 MoE from_pretrained failure report (issue 45310 and 45478); the rest are only superficially similar and do not look mergeable as the same bug.", - "confidence": 0.94, - "canonical_issue_reason": "issue 45310 is the best canonical issue for the one true duplicate pair: it is the earlier, clearer report of the Qwen3.5 MoE from_pretrained failure and is matched by issue 45478 with the same title and underlying symptom.", + "summary": "The set is mostly heterogeneous and does not form a coherent duplicate cluster. Only one soft pair looks like an exact duplicate: the two Qwen3.5 MoE from_pretrained regression reports.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45310 is the best canonical issue: it is an exact duplicate of 45478 and the earlier report of the same Qwen3.5 MoE from_pretrained failure.", "canonical_pr_reason": null, - "best_issue_reason": "issue 45310 is the strongest representative issue in this set because it has a direct duplicate and a focused, concrete failure mode; the other issues are distinct bugs or feature regressions.", + "best_issue_reason": "45310 is the strongest representative for the only true duplicate pair in this set; the rest of the issues are unrelated and do not support a broader canonical cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "Different failures: DAC from_latents/STE mismatch vs model loading warnings about unexpected position_id keys." + "reason": "Different bugs: DAC.from_latents/STE mismatch vs position-id key warnings." }, { - "left": "issue:43335", - "right": "issue:43504", + "left": "issue:44206", + "right": "issue:45823", "accept": false, - "reason": "Unrelated model/config bugs: sparse-layer creation in SwitchTransformers vs BEiT legacy field loading." + "reason": "Both are device/dtype issues, but they affect different models and code paths." }, { - "left": "issue:44206", - "right": "issue:45823", + "left": "issue:43335", + "right": "issue:43504", "accept": false, - "reason": "Different code paths and symptoms: LASR feature extractor center arg crash vs Gemma4 device-map forward mismatch." + "reason": "Both involve Beit, but one is sparse-layer config creation and the other is legacy preset loading." }, { "left": "issue:43278", "right": "issue:43825", "accept": false, - "reason": "Training/eval dtype drift is unrelated to the pipeline translation-task error message issue." + "reason": "Training/eval dtype drift is unrelated to the pipeline translation error message." }, { "left": "issue:29127", "right": "issue:32090", "accept": false, - "reason": "LayoutLMv3 box-validation error and Trainer _gpu_broadcast_one NoneType error are distinct bugs." + "reason": "Unrelated problems: LayoutLMv3 error clarity vs Trainer _gpu_broadcast_one NoneType handling." }, { "left": "issue:45588", "right": "issue:45859", "accept": false, - "reason": "Both touch Qwen-related attention/rotary code, but one is a flash_attention.py s_aux=None crash and the other is CPU-offload incompatibility in rotary embedding; not the same bug." + "reason": "Both mention offload/flash-attention, but the failing code paths and symptoms differ." }, { - "left": "issue:39692", - "right": "issue:43526", + "left": "issue:45290", + "right": "issue:45381", "accept": false, - "reason": "Documentation/example errors are not the same underlying bug as BeitImageProcessorFast reduce_labels returning only one label." + "reason": "Chat template crash with tool calls is unrelated to Qwen2.5-VL vision_position_ids." }, { - "left": "issue:45290", - "right": "issue:45381", + "left": "issue:39692", + "right": "issue:43526", "accept": false, - "reason": "Different multimodal chat/template bugs: tool-call content handling vs Qwen2.5-VL video vision_position_ids." + "reason": "Documentation example errors vs BeitImageProcessorFast reduce_labels behavior are different issues." }, { "left": "issue:45310", "right": "issue:45478", "accept": true, - "reason": "Same bug reported twice with the same title and symptom: transformers>=5.4.0 Qwen3.5 Moe from_pretrained error." + "reason": "Same bug report: identical Qwen3.5 Moe from_pretrained regression on transformers>=5.4.0." }, { "left": "issue:43335", "right": "issue:45083", "accept": false, - "reason": "Both involve model internals, but sparse-layer config creation is unrelated to _get_feat_extract_output_lengths behavior." + "reason": "SwitchTransformers sparse-layer config bug is unrelated to qwen3_omni_moe output-length helper behavior." }, { "left": "issue:39692", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 docs/example problems are unrelated to unexpected position_id key warnings." + "reason": "SigLIP2 docs example issues are unrelated to generic position-id key warnings." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is a different failure mode than a stale GraniteSpeech device-override CI test." + "reason": "BitNet accelerate-loading failure and GraniteSpeech CI device-override flake are unrelated." }, { "left": "issue:44448", "right": "issue:44779", "accept": false, - "reason": "Both mention v4/v5 regressions, but Pegasus output drift and Deepseek tokenizer corruption are different concrete bugs." + "reason": "Different regressions: Pegasus output changes vs Deepseek tokenizer correctness." }, { - "left": "issue:43335", - "right": "issue:44112", + "left": "issue:45459", + "right": "issue:46234", "accept": false, - "reason": "SwitchTransformers sparse-layer config and GraniteSpeech CI device-override issues are unrelated." + "reason": "Tokenizer error masking via protobuf import is unrelated to a missing import-utils symbol." }, { - "left": "issue:45459", - "right": "issue:46234", + "left": "issue:43335", + "right": "issue:44112", "accept": false, - "reason": "Tokenizer error masking when protobuf is absent is unrelated to a missing import_utils availability helper." + "reason": "SwitchTransformers config bug is unrelated to GraniteSpeech CI flakiness." }, { "left": "issue:43335", "right": "issue:45593", "accept": false, - "reason": "Sparse-layer creation bug and D-FINE auxiliary-loss omission are different subsystems and fixes." + "reason": "Different model components and symptoms; not the same underlying defect." }, { "left": "issue:34689", "right": "issue:35141", "accept": false, - "reason": "Model-loading regression for Llama 3.2 vision is not the same as embedding reinitialization after resizing token embeddings." + "reason": "Both are loading/initialization issues, but they concern different models and causes." }, { "left": "issue:43606", "right": "issue:44857", "accept": false, - "reason": "CPU offload device mismatch in Bark-small is unrelated to float16 AMP crashes in LwDetrImageLoss." + "reason": "CPU offload device mismatch is unrelated to LwDetrImageLoss AMP crashes." }, { "left": "issue:42490", "right": "issue:43854", "accept": false, - "reason": "Trailing-slash import warning and GLM-4.7-Flash test loading failure are distinct issues." + "reason": "Local-path warning and GLM-4.7 model-load test failure are not the same bug." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer missing ReLU is unrelated to trust_remote_code breakage from removing is_torch_fx_available." + "reason": "DSA indexer/ReLU discussion is unrelated to removal of is_torch_fx_available." }, { "left": "issue:38175", "right": "issue:44493", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 do not match the unexpected position_id key regression." + "reason": "Unexpected zero probabilities in SigLIP2 is not the same as position-id key warnings." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Different dtype-related failures in different models/code paths; not the same concrete bug." + "reason": "Different failure modes: AMP crash vs bfloat16 dtype mismatch in inference." }, { "left": "issue:45920", "right": "issue:46192", "accept": false, - "reason": "Wrong token IDs in AutoTokenizer for many models is not the same as silently ignoring return_tensors when text_kwargs is passed." + "reason": "Wrong token IDs regression is unrelated to return_tensors being ignored with text_kwargs." }, { - "left": "issue:28218", - "right": "issue:32090", + "left": "issue:44448", + "right": "issue:44977", "accept": false, - "reason": "Tokenizer added-space behavior and Trainer broadcast NoneType error are unrelated." + "reason": "Pegasus v4/v5 output drift and Qwen3.5 flash-attention generation failure are different bugs." }, { - "left": "issue:44448", - "right": "issue:44977", + "left": "issue:28218", + "right": "issue:32090", "accept": false, - "reason": "Pegasus output divergence and Qwen3.5 flash-attention generation failure are different model-specific regressions." + "reason": "Tokenizer spacing after added token is unrelated to Trainer GPU broadcast NoneType errors." }, { "left": "issue:40444", "right": "issue:43526", "accept": false, - "reason": "IterableDataset multi-image finetuning failure is unrelated to BeitImageProcessorFast label reduction." + "reason": "Multi-image IterableDataset finetuning failure is unrelated to BeitImageProcessorFast label reduction." }, { - "left": "issue:43701", - "right": "issue:45083", + "left": "issue:44977", + "right": "issue:45229", "accept": false, - "reason": "resume_from_checkpoint key mismatch and Qwen3_omni_moe feature-length helper behavior are not the same bug." + "reason": "Flash-attention generation issue and multi-GPU CUDA OOM are different failure classes." }, { - "left": "issue:44977", - "right": "issue:45229", + "left": "issue:43701", + "right": "issue:45083", "accept": false, - "reason": "Flash-attention generation issues and Gemma4 multi-GPU CUDA OOM are distinct failures." + "reason": "resume_from_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 unrelated." + "reason": "Local checkpoint loading with PEFT adapters is unrelated to FP8 inference on B200." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer GPU broadcast NoneType error and MacOS bus error in CLIP loading are different issues." + "reason": "Trainer broadcast NoneType and MacOS bus error on CLIP are unrelated." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 test failures are unrelated to embedding reinitialization after resize." + "reason": "Flash Attention 2 test failures are unrelated to output embedding reinitialization in post_init." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "torch.compile SDPA failure in Bamba-9B-v2 is unrelated to the Qwen3_omni_moe helper-function bug." + "reason": "torch.compile/SDPA failure in Bamba is unrelated to qwen3_omni_moe helper behavior." } ] }, "evaluator_result": { "accept": true, - "feedback": "Grounded overall and conservative in the soft-edge calls. The duplicate pair 45310/45478 is well supported by the matching title and symptom, and the remaining pairs are appropriately treated as non-mergeable distinct bugs." + "feedback": "Grounded and conservative overall: the exact duplicate pair 45310/45478 is clearly supported by the packet (same title and same Qwen3.5 MoE from_pretrained regression), and the summary does not over-claim a broader cluster. The false soft-edge verdicts are phrased cautiously as different bugs/code paths, which fits the available evidence." }, "evaluator_used": true, "retried": false, @@ -8792,48 +8811,48 @@ "issue:44855|issue:46399", "issue:42490|issue:43575", "issue:10000|issue:15354", + "issue:37428|issue:44987", "issue:43927|issue:45698", "issue:44561|issue:44779", - "issue:37428|issue:44987", "issue:43278|issue:43329", "issue:27957|issue:34689", "issue:44857|issue:45588", "issue:43873|issue:45229", "issue:41720|issue:45229", - "issue:43278|issue:43381", "issue:44568|issue:45491", "issue:44977|issue:45563", + "issue:43278|issue:43381", "issue:45372|issue:45588", "issue:45588|issue:45823", - "issue:44448|issue:44466", "issue:43412|issue:45588", + "issue:44448|issue:44466", "issue:45372|issue:45406", "issue:43976|issue:45440", - "issue:25251|issue:32090", "issue:45198|issue:45803", + "issue:25251|issue:32090", "issue:43653|issue:44743", "issue:45245|issue:45563", "issue:32090|issue:33290", - "issue:42490|issue:43873", "issue:44186|issue:45198", + "issue:44206|issue:44479", "issue:44361|issue:45198", - "issue:33357|issue:39692", - "issue:44206|issue:44479" + "issue:42490|issue:43873", + "issue:33357|issue:39692" ], - "prepared_review_unit_hash": "34c80fdb7a626b22defe42dcde4273e3e8adaf45669e9f917eeb3e03588a0f2f", + "prepared_review_unit_hash": "e3a1bce9c6108e41be27b20946ee20e3b2c51ec913af3d5cde7cde568656f9eb", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36964, + "serialized_chars": 36962, "estimated_input_tokens": 9241, "estimated_eval_tokens": 18738 }, @@ -8842,192 +8861,192 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a grab bag of unrelated Transformers bugs; none of the soft pairs look like the same concrete underlying defect, so I would not merge any of them as duplicates.", - "confidence": 0.38, - "canonical_issue_reason": "Issue 45698 is the strongest anchor here: it is open, has substantial discussion/inbound attention, and describes a broadly useful save/load regression in `from_pretrained`/`save_pretrained` that is representative of a common bug class.", + "summary": "This set is a loose similarity cluster, not a true duplicate cluster: the issues span unrelated models, loaders, tokenizers, attention paths, and memory/runtime failures. No soft edge is strong enough to merge as the same underlying bug or change.", + "confidence": 0.9, + "canonical_issue_reason": "issue:43873 is the broadest and most representative standalone issue in the set, with the highest apparent discussion/inbound activity and a general loading/offloading theme that is closer to a central cluster anchor than the many highly specific bugs.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45698 is the best single triage candidate in this mixed set because it is active, well-scoped, and actionable, with clear user impact and enough context to drive follow-up work.", + "best_issue_reason": "issue:43873 is the best issue representative because it is broad, active, and ties to a common infrastructure problem class (quantization/offloading/memory behavior), even though the rest of the items are not true duplicates of it.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44855", "right": "issue:46399", "accept": false, - "reason": "Both mention decorator/parser-related failures, but they affect different code paths and models (`DebertaV2Model` import vs Qwen3.5 rotary kernel registration); not the same bug." + "reason": "Both mention decorator/script-related failures, but one is a Python 3.13 torch.jit parsing issue and the other is a missing decorator/registration bug in Qwen3.5; different code paths and fixes." }, { "left": "issue:42490", "right": "issue:43575", "accept": false, - "reason": "A trailing-slash import warning is unrelated to a Qwen2 tensor-parallel OOM; different symptoms and failure modes." + "reason": "Local-path import warnings and tensor-parallel OOM on Qwen2-57B are unrelated failure modes." }, { "left": "issue:10000", "right": "issue:15354", "accept": false, - "reason": "German DistilBertModel raising an issue and a `torch.jit.script` GeneratorExp export problem are distinct model/export failures." + "reason": "Generic DistilBertModel failure and torch.jit GeneratorExp export limitation on ViT are distinct issues." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "A missing flash-attention utility import and a model-loading failure for a specific repo are not the same underlying bug." }, { "left": "issue:43927", "right": "issue:45698", "accept": false, - "reason": "Both involve save/load behavior, but one is config token-ID persistence and the other is custom-module resolution after save; different underlying bugs." + "reason": "Both concern save/load behavior, but one loses custom token IDs and the other restores the wrong custom module; separate defects." }, { "left": "issue:44561", "right": "issue:44779", "accept": false, - "reason": "Removal of `is_torch_fx_available` breaking trust-remote-code models is unrelated to Deepseek tokenizer output changes." - }, - { - "left": "issue:37428", - "right": "issue:44987", - "accept": false, - "reason": "Both are loading failures, but the first is an import error for a missing flash-attention helper while the second is a model-specific load failure with no evidence they share the same code path." + "reason": "API removal breaking trust_remote_code models is unrelated to DeepSeek tokenizer output regressions." }, { "left": "issue:43278", "right": "issue:43329", "accept": false, - "reason": "Embedding dtype drift during eval and a multimodal token-counting NameError are unrelated defects." + "reason": "Embedding dtype drift in eval and an undefined-variable multimodal token bug are unrelated." }, { "left": "issue:27957", "right": "issue:34689", "accept": false, - "reason": "Flash Attention 2 support for XLM-Roberta and Llama 3.2 vision model loading are different model/regression issues." + "reason": "Both involve model loading and attention features, but they target different models and different breakages." }, { "left": "issue:44857", "right": "issue:45588", "accept": false, - "reason": "AMP/CUDA crash in `LwDetrImageLoss` and flash-attention sink-less model crash are separate runtime failures." + "reason": "AMP/CUDA loss crash and a flash-attention s_aux=None AttributeError are different runtime paths." }, { "left": "issue:43873", "right": "issue:45229", "accept": false, - "reason": "Quantization offloading behavior and Gemma4 multi-GPU OOM are both memory-related but not the same concrete bug or fix." + "reason": "Both mention memory pressure, but one is quantization/offloading behavior and the other is a specific multi-GPU inference OOM on Gemma4." }, { "left": "issue:41720", "right": "issue:45229", "accept": false, - "reason": "Auto device-mapping cuda assert and Gemma4 multi-GPU OOM are distinct inference failures." - }, - { - "left": "issue:43278", - "right": "issue:43381", - "accept": false, - "reason": "Eval-time dtype change and gradient-checkpointing-in-eval-mode are different configuration bugs." + "reason": "Auto device-map cudaErrorAssert and CUDA OOM are different symptoms with different likely causes." }, { "left": "issue:44568", "right": "issue:45491", "accept": false, - "reason": "Tokenizer special-token insertion and Gemma3 NaN embeddings from sliding-window attention are unrelated." + "reason": "Tokenizer special-token handling and Gemma3 NaN embeddings are unrelated bugs." }, { "left": "issue:44977", "right": "issue:45563", "accept": false, - "reason": "Flash-attention generation regression and a stale warning in paged generate are not the same bug." + "reason": "Flash-attention generation regression and a stale warning in paged generate() are not the same code-path problem." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Eval/training dtype mismatch and gradient-checkpointing-in-eval are different behaviors." }, { "left": "issue:45372", "right": "issue:45588", "accept": false, - "reason": "Gemma4 processor import failure and flash-attention `s_aux=None` crash affect different layers and code paths." + "reason": "Gemma4 processor import failure and flash_attention integration crash are separate failures." }, { "left": "issue:45588", "right": "issue:45823", "accept": false, - "reason": "Both touch Gemma/attention internals, but one is a flash-attention crash and the other is a device-mismatch in PLE forward; not mergeable as one fix." + "reason": "An AttributeError in flash_attention integration and a Gemma4 device mismatch during forward are not the same bug." }, { - "left": "issue:44448", - "right": "issue:44466", + "left": "issue:43412", + "right": "issue:45588", "accept": false, - "reason": "Different-output regression for Pegasus and inconsistent `lm_head.weight` serialization are separate v5 regressions." + "reason": "Executorch export problems and flash-attention integration errors are unrelated." }, { - "left": "issue:43412", - "right": "issue:45588", + "left": "issue:44448", + "right": "issue:44466", "accept": false, - "reason": "Executorch export failures and flash-attention sink handling are unrelated bugs." + "reason": "Different output behavior and inconsistent serialization are related to v5 behavior but are not the same concrete issue." }, { "left": "issue:45372", "right": "issue:45406", "accept": false, - "reason": "Both mention Gemma4, but one is processor-loading import failure and the other is `transformers serve` missing `_tokenizer`; different code paths." + "reason": "Both are Gemma4-related, but one is a missing dependency import and the other is a missing _tokenizer attribute in serve." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility and DeepSeekV3MoE parity divergence are unrelated." + "reason": "Python version incompatibility and DeepSeekV3MoE divergence are unrelated." }, { - "left": "issue:25251", - "right": "issue:32090", + "left": "issue:45198", + "right": "issue:45803", "accept": false, - "reason": "Pipeline output shape change and Trainer `_gpu_broadcast_one` type error are unrelated." + "reason": "Wav2Vec2 save/tokenization failure and a bare-except bug in Fuyu tensor conversion are different issues." }, { - "left": "issue:45198", - "right": "issue:45803", + "left": "issue:25251", + "right": "issue:32090", "accept": false, - "reason": "Wav2Vec2 save/tokenization failure and Fuyu's bare `except:` swallowing errors are different bugs." + "reason": "Pipeline top_k output-shape change and Trainer _gpu_broadcast_one TypeError are unrelated." }, { "left": "issue:43653", "right": "issue:44743", "accept": false, - "reason": "BigBird tokenizer special-token registration and Qwen3.5 recurrent-state reset are unrelated." + "reason": "BigBirdTokenizer special-token registration and Qwen3.5 recurrent-state reset are different model subsystems." }, { "left": "issue:45245", "right": "issue:45563", "accept": false, - "reason": "Category-count limit error and a stale warning in paged generate are not the same issue." + "reason": "Category cardinality limit and a stale warning in paged generate() have no common underlying defect." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "Trainer broadcast type error and Adafactor+DeepSpeed OOM are distinct training failures." + "reason": "Trainer broadcast TypeError and deepspeed Adafactor OOM are different runtime failures." }, { - "left": "issue:42490", - "right": "issue:43873", + "left": "issue:44186", + "right": "issue:45198", "accept": false, - "reason": "Trailing-slash path import warning and quantization offloading behavior are too different to be duplicates." + "reason": "Different tokenizers, different crash modes, and different affected tasks; not a duplicate." }, { - "left": "issue:44186", - "right": "issue:45198", + "left": "issue:44206", + "right": "issue:44479", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes on NER/padding and Wav2Vec2 save/tokenization failures are unrelated tokenizer bugs." + "reason": "Unsupported center argument in feature extraction and Qwen VL video regression are separate bugs." }, { "left": "issue:44361", "right": "issue:45198", "accept": false, - "reason": "MLuke tokenizer AttributeError and Wav2Vec2 save/tokenization failure do not share the same underlying defect." + "reason": "MLukeTokenizer AttributeError and Wav2Vec2 save/tokenization failure are unrelated." }, { - "left": "issue:33357", - "right": "issue:39692", + "left": "issue:42490", + "right": "issue:43873", "accept": false, - "reason": "MacOS bus error with a community CLIP model and SigLIP2 documentation errors are different problems; one is runtime, the other docs/example correctness." + "reason": "Import warning for local paths and quantization/offloading behavior are not the same issue." }, { - "left": "issue:44206", - "right": "issue:44479", + "left": "issue:33357", + "right": "issue:39692", "accept": false, - "reason": "Unsupported `center` argument in `LasrFeatureExtractor` and Qwen video-input regression are separate issues." + "reason": "MacOS bus error on a community CLIP model and SigLIP2 docs errors are different problems, one runtime and one documentation/example-related." } ] }, @@ -9104,8 +9123,8 @@ "issue:29127|issue:34567", "issue:43065|issue:45412", "issue:43611|issue:46121", - "issue:44871|issue:44977", "issue:42915|issue:42994", + "issue:44871|issue:44977", "issue:43065|issue:43994", "issue:43066|issue:44462", "issue:45459|issue:45712", @@ -9132,235 +9151,235 @@ "issue:44442|issue:45072", "issue:42490|issue:43381", "issue:43575|issue:44387", - "issue:42915|issue:43931", - "issue:43761|issue:45859" + "issue:43761|issue:45859", + "issue:42915|issue:43931" ], - "prepared_review_unit_hash": "eda2268056f6b6a1f8975ed01f1e7588bbadd6c8e55c680fb383f3efadc50298", + "prepared_review_unit_hash": "2ab49ca1bf75fcc75ac3ad1776cd013886862684377f84b3308b0797bc39ac3e", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37515, - "estimated_input_tokens": 9379, - "estimated_eval_tokens": 19014 + "serialized_chars": 37522, + "estimated_input_tokens": 9381, + "estimated_eval_tokens": 19018 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are broadly heterogeneous and do not form a duplicate cluster. They span tokenizer, model-loading, quantization, vision, training-state, and memory-regression topics with no shared underlying bug or change.", - "confidence": 0.97, - "canonical_issue_reason": "No canonical issue: the cluster is a grab-bag of unrelated bugs with no shared root cause or common fix path.", + "summary": "This looks like a heterogeneous Transformers issue bundle rather than a true duplicate cluster. The soft-similarity pairs share keywords or subsystems, but they do not appear to describe the same underlying bug or change.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue really fits this set because the reports span unrelated bugs; the closest thing to a representative item is a concrete, user-facing tokenizer-loading bug.", "canonical_pr_reason": null, - "best_issue_reason": "No single issue is a good global representative for deduping; none subsumes the others.", + "best_issue_reason": "Best overall issue representative is the most self-contained and actionable bug report; the rest are mostly model- or subsystem-specific and not duplicate-alike.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "Different subsystems: LayoutLMv3 error-message clarity vs TrainerState token-count tracking." + "reason": "Different areas: LayoutLMv3 error messaging vs TrainerState token counting." }, { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Sam3PixelDecoder dummy Conv2d placeholder vs RT-DETR memory not being released." + "reason": "One is a dummy Conv2d in Sam3PixelDecoder; the other is RT-DETR memory not being released." }, { "left": "issue:43611", "right": "issue:46121", "accept": false, - "reason": "base_model_prefix loading regression vs rope-parameter JSON list TypeError." + "reason": "Loading via base_model_prefix and rope-parameter JSON handling are unrelated bugs." }, { - "left": "issue:44871", - "right": "issue:44977", + "left": "issue:42915", + "right": "issue:42994", "accept": false, - "reason": "Gemma-3 EOS config mismatch vs Qwen3.5 flash-attention generation failure." + "reason": "Both involve quantization, but one is a Qwen3Moe FP8 failure and the other is quantized model saving." }, { - "left": "issue:42915", - "right": "issue:42994", + "left": "issue:44871", + "right": "issue:44977", "accept": false, - "reason": "Qwen3Moe FP8 config failure vs quantized-model saving failure." + "reason": "Gemma eos token config mismatch vs Qwen3.5 flash-attention generation failure." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Sam3PixelDecoder dummy layer issue vs SigLIP2 produces nonsensical outputs." + "reason": "Sam3PixelDecoder dummy layer issue is unrelated to SigLIP2 bad AutoModel/pipeline outputs." }, { "left": "issue:43066", "right": "issue:44462", "accept": false, - "reason": "Wrong tokenizer decoder type in v5 vs AutoTokenizer ignoring tokenizer.json." + "reason": "Both are tokenizer-loading issues, but the concrete failure modes differ and are not the same bug." }, { "left": "issue:45459", "right": "issue:45712", "accept": false, - "reason": "Tokenizer error masking due to missing protobuf vs leftover dummy classes in dummy_pt_objects.py." + "reason": "Tokenizer error masking with protobuf missing is unrelated to leftover dummy class cleanup." }, { "left": "issue:43278", "right": "issue:43575", "accept": false, - "reason": "Embedding dtype drift in eval vs Qwen2-57B TP loading OOM." + "reason": "Embedding dtype drift in eval mode is unrelated to Qwen2 loading OOM." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention 2 test failures vs void segmentation-map processing failure." + "reason": "Flash Attention 2 test failures and void segmentation map processing are different issues." }, { "left": "issue:45412", "right": "issue:46369", "accept": false, - "reason": "RT-DETR garbage-collection leak vs Sam3-litetext memory overflow." + "reason": "Both are memory-related, but RT-DETR GC leaks and Sam3-litetext overflow are different code paths." }, { "left": "issue:45372", "right": "issue:45459", "accept": false, - "reason": "Missing ReasoningEffort import for Gemma 4 processor vs protobuf-hidden tokenizer errors." + "reason": "Gemma 4 processor import error from mistral_common is unrelated to tokenizer protobuf error masking." }, { "left": "issue:43994", "right": "issue:45412", "accept": false, - "reason": "SigLIP2 inference corruption vs RT-DETR memory leak." + "reason": "SigLIP2 bad inference and RT-DETR memory release are unrelated bugs." }, { "left": "issue:44265", "right": "issue:45198", "accept": false, - "reason": "torch.export incompatibility with torch_compilable_check vs Wav2Vec2 save/tokenization failure." + "reason": "torch.export failures from torch_compilable_check are not the same as Wav2Vec2 save/tokenization failures." }, { "left": "issue:43381", "right": "issue:43701", "accept": false, - "reason": "Gradient checkpointing in eval mode vs resume_from_checkpoint key mismatch." + "reason": "Gradient-checkpointing eval-mode behavior and resume_from_checkpoint key mismatch are separate training issues." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBirdTokenizer mask-token registration bug vs FastSpeech2ConformerTokenizer load failure." + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer bugs." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Smollm3 RoPE-layer count bug vs DeepseekV3MoE implementation divergence." + "reason": "Smollm3 RoPE-layer dropout and DeepseekV3MoE divergence from remote implementation are not the same defect." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "TokenizerBackend post-processor update bug vs Gemma3 NaN embeddings on mixed-length batches." + "reason": "Runtime special-token post-processing and Gemma3 NaN embeddings are unrelated." }, { "left": "issue:27957", "right": "issue:29942", "accept": false, - "reason": "XLMRoberta FlashAttention2 support request vs failing FA2 tests; not the same bug." + "reason": "Both mention Flash Attention 2, but one is XLMRoberta support and the other is failing tests." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "LayoutLMv3 error clarity vs CLIP MacOS bus error." + "reason": "LayoutLMv3 error clarity and MacOS CLIP bus error are unrelated." }, { "left": "issue:43824", "right": "issue:44279", "accept": false, - "reason": "Qwen2_5_VL import error vs generic dependency issue." + "reason": "Missing Qwen2_5_VL import and a generic dependency issue are not the same bug." }, { "left": "issue:45440", "right": "issue:45593", "accept": false, - "reason": "DeepseekV3MoE divergence vs D-FINE auxiliary-loss logic." + "reason": "DeepseekV3MoE implementation drift is unrelated to D-FINE auxiliary-loss behavior." }, { "left": "issue:43994", "right": "issue:46369", "accept": false, - "reason": "SigLIP2 wrong outputs vs Sam3-litetext memory overflow." + "reason": "SigLIP2 wrong outputs and Sam3-litetext memory overflow are different symptoms and code paths." }, { "left": "issue:43526", "right": "issue:45834", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug vs Kosmos2.5 OCR index error." + "reason": "BeitImageProcessorFast label reduction and Kosmos2.5 OCR index error are unrelated image-processing bugs." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "BeitImageProcessorFast reduce_labels bug vs LayoutLMv2Tokenizer NER/padding crash." + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to LayoutLMv2Tokenizer NER padding/truncation crashes." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "Unclear LayoutLMv3 error message vs embedding reinitialization after resize_token_embeddings." + "reason": "Error-message clarity in LayoutLMv3 and embedding reinit after resize are unrelated." }, { "left": "issue:44625", "right": "issue:45584", "accept": false, - "reason": "Qwen3.5 num_labels propagation vs Whisper empty-transcription failure after align_special_tokens." + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are unrelated." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "Qwen2-57B TP OOM vs checkpoint key mismatch." + "reason": "Qwen2 loading OOM and checkpoint resume key mismatch are different problems." }, { "left": "issue:44442", "right": "issue:45072", "accept": false, - "reason": "AutoTokenizer failing FastSpeech2ConformerTokenizer vs bfloat16 dtype mismatches." + "reason": "Tokenizer loading failure and dtype mismatch in inference are unrelated." }, { "left": "issue:42490", "right": "issue:43381", "accept": false, - "reason": "Trailing-slash import warning vs gradient-checkpointing eval-mode limitation." + "reason": "Trailing-slash local model import warning and gradient-checkpointing eval-mode behavior are unrelated." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Qwen2-57B TP OOM vs increased reserved memory under int4 quantization." + "reason": "Both are OOM-related, but one is model/tp loading and the other is int4 quantization memory regression." }, { - "left": "issue:42915", - "right": "issue:43931", + "left": "issue:43761", + "right": "issue:45859", "accept": false, - "reason": "Qwen3Moe FP8 config failure vs Qwen3-VL weight-shape mismatch." + "reason": "CLIPVisionModel hidden_states regression and CPU-offload incompatibility in Qwen3_5MoeTextRotaryEmbedding are unrelated." }, { - "left": "issue:43761", - "right": "issue:45859", + "left": "issue:42915", + "right": "issue:43931", "accept": false, - "reason": "CLIPVisionModel hidden_states regression vs Qwen3_5Moe CPU-offload incompatibility." + "reason": "Qwen3Moe FP8 failure and Qwen3-VL weight-shape mismatch are different model-loading bugs." } ] }, @@ -9435,19 +9454,19 @@ ], "soft_pairs": [ "issue:43526|issue:44857", - "issue:46121|issue:46399", "issue:43334|issue:43531", + "issue:46121|issue:46399", "issue:44977|issue:45245", - "issue:44464|issue:45491", "issue:43262|issue:44206", + "issue:44464|issue:45491", "issue:27957|issue:29127", "issue:44743|issue:44811", "issue:43452|issue:45588", "issue:27957|issue:33290", "issue:43646|issue:44077", "issue:33290|issue:34689", - "issue:33290|issue:35141", "issue:44062|issue:45907", + "issue:33290|issue:35141", "issue:33290|issue:34567", "issue:43381|issue:43873", "issue:44442|issue:44898", @@ -9456,209 +9475,209 @@ "issue:42490|issue:44743", "issue:43316|issue:43525", "issue:43761|issue:44079", - "issue:34689|issue:43611", "issue:42175|issue:46234", + "issue:34689|issue:43611", "issue:43720|issue:44442", "issue:43232|issue:43388", - "issue:44077|issue:44479", - "issue:43931|issue:45216" + "issue:43931|issue:45216", + "issue:44077|issue:44479" ], - "prepared_review_unit_hash": "7cbf9ab5e4c729d27b506114811c13673f22e7f93aa4af925217f09d1a9bc383", + "prepared_review_unit_hash": "ba1c197cf768e52f08100cb975353f8dd727d59be017668eee78412226f7e39b", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36286, - "estimated_input_tokens": 9072, - "estimated_eval_tokens": 18400 + "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": "This is a noisy cluster of mostly unrelated issues; none of the soft pairs look like the same underlying bug or change, and there are no PRs to consider.", - "confidence": 0.98, - "canonical_issue_reason": "If a single representative is needed, issue 33290 is the most connected item in the set, but it is not a true duplicate hub because the surrounding issues cover unrelated bugs.", + "summary": "The set is highly heterogeneous: most items are distinct bugs/regressions in different subsystems or model families, and none of the soft pairs look like true duplicates of the same concrete underlying issue.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the cluster spans unrelated loading, generation, preprocessing, optimizer, and config bugs.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 33290 is the least-bad cluster representative by soft-link degree, but its Deepspeed/Adafactor OOM report does not match the other issues closely enough to serve as a real canonical duplicate target.", + "best_issue_reason": "issue:43646 is the broadest representative regression in the set, affecting custom model initialization across models.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43526", "right": "issue:44857", "accept": false, - "reason": "Different failures: BeitImageProcessorFast label reduction vs LwDetrImageLoss AMP/CUDA crash." + "reason": "Different components and failures: BEiT image processor label reduction vs LwDETR loss crashing under float16/CUDA." }, { - "left": "issue:46121", - "right": "issue:46399", + "left": "issue:43334", + "right": "issue:43531", "accept": false, - "reason": "Different code paths: rope validation TypeError vs Qwen3.5 kernel registration crash." + "reason": "Different Qwen variants and bugs: missing pad_token_id on a text config vs sliding-window recurrence reset." }, { - "left": "issue:43334", - "right": "issue:43531", + "left": "issue:46121", + "right": "issue:46399", "accept": false, - "reason": "Qwen3-VL pad_token_id loading error is unrelated to Qwen3-MoE sliding_window behavior." + "reason": "Unrelated code paths: rope-validation TypeError vs rotary-pos-embedding registration decorator failure." }, { "left": "issue:44977", "right": "issue:45245", "accept": false, - "reason": "Flash-attention generation regression is unrelated to the 2^24 category limit error." + "reason": "Different problems entirely: flash-attention generation regression vs category-count limit error." }, { - "left": "issue:44464", - "right": "issue:45491", + "left": "issue:43262", + "right": "issue:44206", "accept": false, - "reason": "Compiled forward chunked generation inconsistency is not the same bug as Gemma3 NaNs from sliding-window padding." + "reason": "Audio chat-template sampling-rate default bug vs Laser/feature-extractor unsupported-center crash." }, { - "left": "issue:43262", - "right": "issue:44206", + "left": "issue:44464", + "right": "issue:45491", "accept": false, - "reason": "Audio chat-template sample-rate defaulting is unrelated to the unsupported center-arg crash in the feature extractor." + "reason": "Different generation/attention failures: compiled-forward chunked generation inconsistency vs Gemma3 NaN embeddings with sliding-window all-padding cases." }, { "left": "issue:27957", "right": "issue:29127", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta and LayoutLMv3 box error messaging are different issues." + "reason": "Different subsystems: XLM-RoBERTa Flash Attention 2 support vs LayoutLMv3 error-message clarity." }, { "left": "issue:44743", "right": "issue:44811", "accept": false, - "reason": "Recurrent-state reset in caching is unrelated to Whisper batch_decode skipping special tokens." + "reason": "Different model paths and bugs: Qwen3_5 recurrent state reset vs Whisper batch_decode skipping special tokens." }, { "left": "issue:43452", "right": "issue:45588", "accept": false, - "reason": "GGUF loading for tokenizer/model APIs is a different problem from the flash_attention s_aux=None crash." + "reason": "Tokenizer/model loading with gguf_file vs flash-attention sink-less-model AttributeError are unrelated." }, { "left": "issue:27957", "right": "issue:33290", "accept": false, - "reason": "XLMRoberta flash-attention support and Adafactor/DeepSpeed OOM are not the same bug." + "reason": "One is Flash Attention model support, the other is Adafactor OOM in DeepSpeed; no shared underlying bug." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Custom model init breakage in v5 is broader and distinct from patchtsmixer's post_init policy issue." + "reason": "Both touch initialization, but one is a broad v5 custom-model-init regression and the other is a specific patchtsmixer post_init API rule change." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "OOM during Adafactor training is unrelated to Llama 3.2 Vision model-loading regression." + "reason": "Optimizer OOM in DeepSpeed vs a model-loading regression for Llama 3.2 Vision; unrelated failures." }, { - "left": "issue:33290", - "right": "issue:35141", + "left": "issue:44062", + "right": "issue:45907", "accept": false, - "reason": "Training OOM does not match the token-embedding reinitialization bug after resize." + "reason": "AddedToken keyword duplication vs list_repo_files revision forwarding are unrelated utility bugs." }, { - "left": "issue:44062", - "right": "issue:45907", + "left": "issue:33290", + "right": "issue:35141", "accept": false, - "reason": "AddedToken special-argument duplication and list_repo_files revision forwarding are unrelated." + "reason": "DeepSpeed optimizer OOM and embedding reinitialization after resize are distinct issues." }, { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "OOM with DeepSpeed/Adafactor is not the same as TrainerState num_input_tokens_seen not updating." + "reason": "Training memory usage vs TrainerState token-count tracking are unrelated." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing in eval mode and quantization offloading are different failure modes." + "reason": "Gradient checkpointing eval-mode restriction vs quantization offloading behavior are different code paths." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "AutoTokenizer loading a tokenizer and Perceiver non-default-resolution inference are unrelated." + "reason": "Tokenizer loading failure for FastSpeech2ConformerTokenizer vs Perceiver non-default-resolution failure are unrelated." }, { "left": "issue:45003", "right": "issue:45440", "accept": false, - "reason": "Unsafe sys.modules access in modeling_utils is not the same as DeepseekV3MoE divergence from remote implementation." + "reason": "Unsafe sys.modules access in modeling_utils vs DeepseekV3MoE divergence are different bugs." }, { "left": "issue:43723", "right": "issue:45479", "accept": false, - "reason": "AutoTokenizer loading in v5 and degenerate single-label classification loss are unrelated." + "reason": "AutoTokenizer loading regression in v5 vs zero-loss classification with num_labels=1 are not the same issue." }, { "left": "issue:42490", "right": "issue:44743", "accept": false, - "reason": "Local-path trailing-slash import warning has no relation to recurrent-state reset in generation." + "reason": "Local-path trailing-slash import warning vs Qwen3_5 recurrent-state reset; different areas and symptoms." }, { "left": "issue:43316", "right": "issue:43525", "accept": false, - "reason": "Gemma3TextConfig API discrepancy and missing Llama4Config pad_token_id are similar in theme but not the same concrete bug." + "reason": "Both concern config APIs, but the described discrepancies are not clearly the same missing attribute or failure mode." }, { "left": "issue:43761", "right": "issue:44079", "accept": false, - "reason": "CLIPVision hidden_states output regression is distinct from ModelOutput key assignment behavior." + "reason": "CLIPVisionModel hidden_states regression vs ModelOutput key assignment semantics are different underlying problems." }, { - "left": "issue:34689", - "right": "issue:43611", + "left": "issue:42175", + "right": "issue:46234", "accept": false, - "reason": "Llama 3.2 Vision loading regression is different from the base_model_prefix loading issue." + "reason": "Packaging/backend dependency issue vs missing import_utils symbol; no common bug." }, { - "left": "issue:42175", - "right": "issue:46234", + "left": "issue:34689", + "right": "issue:43611", "accept": false, - "reason": "Packaging extra omission is unrelated to the missing import_utils helper." + "reason": "Both are version regressions, but they affect different loading paths and failure modes." }, { "left": "issue:43720", "right": "issue:44442", "accept": false, - "reason": "BitNet accelerate-loading unpacking bug is unrelated to FastSpeech2Conformer tokenizer loading." + "reason": "BitNet packed-weight loading bug vs tokenizer loading failure; unrelated." }, { "left": "issue:43232", "right": "issue:43388", "accept": false, - "reason": "Generation kwargs update after sync_gpus is unrelated to gather_for_metrics dropping tuple labels." + "reason": "Generation kwargs update after sync_gpus vs metrics label dropping are unrelated." }, { - "left": "issue:44077", - "right": "issue:44479", + "left": "issue:43931", + "right": "issue:45216", "accept": false, - "reason": "patchtsmixer post_init policy and Qwen3.5 video-input regression do not share the same code-path problem." + "reason": "Weight-shape mismatch on model load vs incorrect save_pretrained checkpoint for Qwen3.5 are different bugs." }, { - "left": "issue:43931", - "right": "issue:45216", + "left": "issue:44077", + "right": "issue:44479", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch on load is different from an incorrect checkpoint saved by save_pretrained." + "reason": "PatchTSMixer post_init policy vs Qwen video-input regression; different models and code paths." } ] }, @@ -9749,29 +9768,29 @@ "issue:25251|issue:28282", "issue:43550|issue:43701", "issue:43606|issue:43701", - "issue:42617|issue:43575", "issue:45632|issue:46424", + "issue:42617|issue:43575", "issue:43575|issue:43742", "issue:27957|issue:32090", "issue:16998|issue:22355", "issue:43316|issue:43716", - "issue:43295|issue:44871", - "issue:43065|issue:44617" + "issue:43065|issue:44617", + "issue:43295|issue:44871" ], - "prepared_review_unit_hash": "94365233add8bb15a1e3fe227f12ca46866ac5b4f588eab88a8e0d383f346d84", + "prepared_review_unit_hash": "a0f67fb85e284ad8f7435ec60c139355f2478f6dc334e9428c274c2c565a457a", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 26, - "serialized_chars": 34687, + "serialized_chars": 34686, "estimated_input_tokens": 8672, "estimated_eval_tokens": 17600 }, @@ -9780,168 +9799,168 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a set of semantically similar but mostly unrelated Transformers issues. The soft pairs share broad themes (loading, caching, configs, multimodal regressions), but none describe the same concrete bug or change, so no duplicates should be collapsed.", - "confidence": 0.93, - "canonical_issue_reason": "No clear canonical issue: the items span many different bugs/regressions across distinct code paths, so there is no safe duplicate target.", + "summary": "The set is heterogeneous and the soft-similarity pairs are mostly false positives: they share broad themes like loading, caching, offload, or v5 regressions, but not the same concrete bug or fix path.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No strong representative issue: the cluster is too heterogeneous for one issue to stand in for the others.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43761", "right": "issue:44188", "accept": false, - "reason": "Hidden-states regression in CLIPVisionModel is unrelated to torch.compile attention-kernel divergence." + "reason": "Both involve torch.compile/regression behavior, but one is CLIPVisionModel hidden_states handling and the other is attention-kernel branching; different code paths and symptoms." }, { "left": "issue:44987", "right": "issue:46178", "accept": false, - "reason": "Model load failure for a specific repo is different from the backend-mapping error for tensorflow_text." + "reason": "Both are loading-time failures, but one is a specific model/package regression and the other is a backend mapping error for tensorflow_text; not the same bug." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 implementation mismatch is unrelated to docs still referencing pipeline() after v5 removals." + "reason": "SigLIP2 implementation discrepancy vs stale docs referencing pipeline() after v5 removals; one is model correctness, the other is documentation drift." }, { "left": "issue:43873", "right": "issue:44871", "accept": false, - "reason": "Quantized offloading/cache behavior is a different bug from Gemma eos_token_id inconsistency." + "reason": "Quantization/offloading behavior vs Gemma eos_token_id config inconsistency; unrelated issues." }, { "left": "issue:44442", "right": "issue:45803", "accept": false, - "reason": "Tokenizer loading failure is unrelated to Fuyu's broad exception handling in convert_to_tensors()." + "reason": "Tokenizer loading failure for FastSpeech2ConformerTokenizer vs a bare-except bug in Fuyu tensor conversion; different components and failure modes." }, { "left": "issue:42907", "right": "issue:45698", "accept": false, - "reason": "Saving dequantized models is not the same as loading the wrong custom module after save_pretrained()." + "reason": "Both touch save/load flows for custom models, but one is dequantized model saving and the other is wrong custom-module selection after save_pretrained; not the same underlying defect." }, { "left": "issue:44464", "right": "issue:46421", "accept": false, - "reason": "Chunked-generation output inconsistency is different from cache lazy-init causing recompilation." + "reason": "Both mention chunked/compiled execution, but one is inconsistent generation outputs and the other is cache lazy-init causing recompilation; different root causes." }, { "left": "issue:43749", "right": "issue:43931", "accept": false, - "reason": "FSDP CPU RAM efficient loading is a different failure mode from Qwen3-VL weight-shape mismatch." + "reason": "FSDP CPU RAM efficient loading broken vs Qwen3-VL weight-shape mismatch; unrelated loading problems." }, { "left": "issue:43643", "right": "issue:44913", "accept": false, - "reason": "Missing fields with trust_remote_code is not the same as GPTNeoX rotary_pct resetting on reload." + "reason": "Both are config persistence issues, but one loses fields with trust_remote_code and the other resets rotary_pct on reload; different model/config paths." }, { "left": "issue:43316", "right": "issue:45759", "accept": false, - "reason": "Both involve config APIs, but Gemma3TextConfig discrepancy and Qwen text_config unwrapping are distinct bugs." + "reason": "Both concern multimodal config APIs, but one is Gemma3TextConfig inconsistency and the other is AutoModelForCausalLM.from_config not unwrapping text_config for Qwen; not mergeable as one fix." }, { "left": "issue:44479", "right": "issue:45290", "accept": false, - "reason": "Video-input regression in Qwen VL models is unrelated to chat template crashes on tool-call assistant messages." + "reason": "Video-input regression in Qwen family vs chat-template crash on tool-call assistant messages; unrelated features and code paths." }, { "left": "issue:44855", "right": "issue:45137", "accept": false, - "reason": "Python 3.13 torch.jit parsing failure is unrelated to the DeepSpeed ZeRO3 deque IndexError." + "reason": "Python 3.13 import/parsing failure in DebertaV2Model vs DeepSpeed ZeRO3 deque underflow; completely different failures." }, { "left": "issue:44190", "right": "issue:44496", "accept": false, - "reason": "Local dataset loading in an example script is not the same as an unrecognized custom model config." + "reason": "Local dataset loading in an image-classification script vs unrecognized-model config loading; different subsystems." }, { "left": "issue:43650", "right": "issue:46202", "accept": false, - "reason": "A placeholder/data issue is unrelated to Gemma pipeline roundtrip tokenization failure." + "reason": "A vague request and a Gemma pipeline roundtrip tokenization bug are not the same underlying issue." }, { "left": "issue:43295", "right": "issue:43825", "accept": false, - "reason": "Processor/tokenizer regression is different from an incorrect pipeline support error message." + "reason": "Both are v5 regressions around pipeline/processor behavior, but one is processor.tokenizer/images handling and the other is an error message about translation support; different fixes." }, { "left": "issue:25251", "right": "issue:28282", "accept": false, - "reason": "top_k nesting behavior in pipeline has nothing to do with missing PyTorch installation." + "reason": "Nested-list output change in top_k pipelines vs missing PyTorch import error; unrelated." }, { "left": "issue:43550", "right": "issue:43701", "accept": false, - "reason": "torch.compile + SDPA failure is unrelated to resume_from_checkpoint key mismatch." + "reason": "torch.compile/SDPA failure in Bamba vs resume_from_checkpoint key mismatch; no shared code path." }, { "left": "issue:43606", "right": "issue:43701", "accept": false, - "reason": "CPU offload device mismatch is a different bug from checkpoint key handling." + "reason": "CPU offload device mismatch vs checkpoint key mismatch; different bugs." }, { - "left": "issue:42617", - "right": "issue:43575", + "left": "issue:45632", + "right": "issue:46424", "accept": false, - "reason": "3d_parallel.py execution problems are not the same as Qwen2 tensor-parallel OOM." + "reason": "Both mention cache behavior, but one is trust_remote_code cache-path collisions and the other is max_cache_length handling for quantized cache; not the same defect." }, { - "left": "issue:45632", - "right": "issue:46424", + "left": "issue:42617", + "right": "issue:43575", "accept": false, - "reason": "Cache path collisions for trust_remote_code are unrelated to max_cache_length being honored only for quantized cache." + "reason": "3d_parallel.py execution problem vs tensor-parallel OOM loading Qwen2-57B; different failure modes." }, { "left": "issue:43575", "right": "issue:43742", "accept": false, - "reason": "Tensor-parallel OOM for Qwen2-57B is unrelated to a MobileLLM config/key loading error." + "reason": "OOM during tensor-parallel load vs key error loading MobileLLM-125M; not the same bug." }, { "left": "issue:27957", "right": "issue:32090", "accept": false, - "reason": "XLMRoberta FlashAttention support is unrelated to Trainer's NoneType GPU broadcast error." + "reason": "Flash Attention 2 support in XLMRoberta vs NoneType broadcast error in Trainer; unrelated." }, { "left": "issue:16998", "right": "issue:22355", "accept": false, - "reason": "model_max_length confusion for DeBERTa-V3 is unrelated to a missing transformers.onnx module." + "reason": "model_max_length question for DeBERTa-V3 vs missing transformers.onnx module; unrelated issues." }, { "left": "issue:43316", "right": "issue:43716", "accept": false, - "reason": "Gemma3TextConfig API mismatch is unrelated to a dtype mismatch between image preprocessor and model." + "reason": "Gemma3TextConfig API discrepancy vs Mistral-3 dtype mismatch between preprocessor and model; different problems." }, { - "left": "issue:43295", - "right": "issue:44871", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "processor.tokenizer regression is different from the Gemma eos_token_id configuration mismatch." + "reason": "Dummy Conv2d in Sam3PixelDecoder vs CUDA OOM in Sam3Video; same family, but not the same underlying bug." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:43295", + "right": "issue:44871", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder is not the same as Sam3Video CUDA OOM." + "reason": "v5 processor/tokenizer regression vs Gemma eos_token_id mismatch; unrelated." } ] }, @@ -10020,9 +10039,9 @@ "issue:44610|issue:46369", "issue:44162|issue:44792", "issue:44871|issue:45584", + "issue:43404|issue:43441", "issue:44077|issue:44534", "issue:45803|issue:46365", - "issue:43404|issue:43441", "issue:45803|issue:46170", "issue:44871|issue:46050", "issue:41950|issue:43475", @@ -10044,202 +10063,202 @@ "issue:43761|issue:44242", "issue:45563|issue:45584" ], - "prepared_review_unit_hash": "a69223c7fefbfc80bebface4901243b4c9b6e3c6ba2f23a63bd64ce705898f08", + "prepared_review_unit_hash": "e53da95d9e968250749f71fb9ab1cd5c575da96f005bd602ead2865a792aaf9e", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 36161, - "estimated_input_tokens": 9041, - "estimated_eval_tokens": 18338 + "serialized_chars": 36169, + "estimated_input_tokens": 9043, + "estimated_eval_tokens": 18342 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The items are mostly distinct Transformers regressions across different model families and code paths. None of the soft-paired issues look like true duplicates; they share only broad themes like loading, generation, or cache handling.", - "confidence": 0.89, - "canonical_issue_reason": "`issue:45698` is the strongest representative: it is open, broadly impacts custom model serialization/loading after `save_pretrained`/`from_pretrained`, and has the highest discussion activity among the open candidates.", + "summary": "This cluster is mostly a false-positive mix of unrelated Transformers bugs: model-specific regressions, config/serialization issues, generation problems, backend-import failures, and a few video/speech cases. A couple of pairs share broad subsystem vocabulary (cache/chunked prefill, save/load, config fields), but they still describe different concrete failures.", + "confidence": 0.96, + "canonical_issue_reason": "issue:45698 is the broadest, most representative open report: it covers broken custom-module resolution after save_pretrained/from_pretrained, has the strongest discussion activity, and overlaps loosely with the other load/save-related item in the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:45698` is the best global issue to anchor this cluster because it describes a high-impact, reusable failure mode affecting custom modules across many models rather than a narrow model-specific regression.", + "best_issue_reason": "issue:45698 is the best overall issue candidate because it is open, broadly applicable to custom model loading, and has the most active discussion/inbound references among the issues here.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43540", "right": "issue:44514", "accept": false, - "reason": "Both are video-related, but they affect different model stacks and different failure modes: Qwen3OmniMoe video processing vs Qwen2.5-VL batched chat template padding behavior." + "reason": "Both involve multimodal/video processing, but they are different model families and different failures: Qwen3OmniMoe video-input handling vs Qwen2.5-VL batched chat-template padding." }, { "left": "issue:43316", "right": "issue:44877", "accept": false, - "reason": "Both concern config handling, but one is a Gemma3TextConfig API mismatch and the other is a Granite Speech loading/strict-config issue; not the same bug." + "reason": "Both are config/API complaints, but the concrete problems differ: Gemma3TextConfig API inconsistency vs granite_speech being blocked by strict config loading." }, { "left": "issue:44610", "right": "issue:46369", "accept": false, - "reason": "Different models and symptoms: OmDet-Turbo input-size mismatch vs Sam3 LiteText memory overflow." + "reason": "Different models and different bug classes: OmDet processor image-size mismatch vs sam3-litetext memory overflow." }, { "left": "issue:44162", "right": "issue:44792", "accept": false, - "reason": "Unrelated model areas and failures; ESM2 workflow breakage is not the same as a Janus image-generation test failure." + "reason": "No shared underlying bug; ESM2 is a model breakage report, while janus test failure is a separate image-generation regression." }, { "left": "issue:44871", "right": "issue:45584", "accept": false, - "reason": "Both involve generation edge cases, but Gemma-3 EOS config inconsistency and Whisper empty-transcription failure are distinct code paths and model families." + "reason": "Both mention generation/token IDs, but Gemma-3 eos_token_id config inconsistency and Whisper empty-transcription failure are distinct code paths." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Same family name is not enough: one is a weight-tying bug in Mistral3ForConditionalGeneration, the other is a FlashAttention failure in Ministral-3." }, { "left": "issue:44077", "right": "issue:44534", "accept": false, - "reason": "One is about optional `post_init` in patchtsmixer, the other about non-persistent buffer initialization in v5; no shared underlying bug." + "reason": "Unrelated issues: patchtsmixer post_init policy vs a v5 non-persistent-buffer initialization bug." }, { "left": "issue:45803", "right": "issue:46365", "accept": false, - "reason": "Bare `except` in Fuyu tensor conversion is unrelated to Trainer import failure involving `datasets`." - }, - { - "left": "issue:43404", - "right": "issue:43441", - "accept": false, - "reason": "Mistral3 weight tying and Ministral-3 FlashAttention failures are separate problems, not one fixable code-path." + "reason": "Different failure modes and surfaces: Fuyu tensor conversion exception handling vs Trainer import failure tied to datasets." }, { "left": "issue:45803", "right": "issue:46170", "accept": false, - "reason": "A tensor conversion exception bug is unrelated to Windows/ROCm import errors from `torch.distributed` dependency handling." + "reason": "Bare-except tensor conversion bug is unrelated to the Windows/ROCm torch.distributed import problem." }, { "left": "issue:44871", "right": "issue:46050", "accept": false, - "reason": "Gemma-3 EOS token config mismatch and sinq quantized model save/load corruption are different serialization/generation bugs." + "reason": "Both concern model state/config after changes, but Gemma eos_token_id mismatch and sinq quantized save/load corruption are different bugs." }, { "left": "issue:41950", "right": "issue:43475", "accept": false, - "reason": "Different subsystems: video-classification pipeline processor lookup vs SAM 3 Video missing output attribute." + "reason": "Different multimodal pipelines and symptoms: video-classification processor lookup vs SAM3 vision encoder missing attribute." }, { "left": "issue:43452", "right": "issue:44496", "accept": false, - "reason": "Both are model-loading issues, but gguf filename handling and missing `model_type` in config.json are different root causes." + "reason": "Both are loading failures, but one is a gguf/from_pretrained path issue and the other is an unrecognized model_type/config.json problem." }, { "left": "issue:43646", "right": "issue:44913", "accept": false, - "reason": "Broad v5 initialization breakage is not the same as GPTNeoX rotary_pct not persisting on reload." + "reason": "Generic v5 initialization breakage vs a specific GPTNeoX rotary_pct persistence bug; related area, not the same issue." }, { "left": "issue:43761", "right": "issue:44479", "accept": false, - "reason": "Different regressions: CLIPVision hidden_states handling vs Qwen video input regression." + "reason": "Both are regressions, but they affect different models and different behavior: CLIPVision hidden_states vs Qwen VL video-input handling." }, { "left": "issue:45127", "right": "issue:45200", "accept": false, - "reason": "LoRA merge collapse with tied embeddings is unrelated to Gemma 4 text-only mm_token_type_ids defaults." + "reason": "Unrelated topics: LoRA merge with extended vocab causing collapse vs Gemma 4 mm_token_type_ids defaulting for text-only fine-tuning." }, { "left": "issue:43388", "right": "issue:43688", "accept": false, - "reason": "One is a last-batch label truncation bug, the other is auxiliary-loss normalization; same training area but different defects." + "reason": "Different bugs in different training losses: metric gathering drops tuple labels vs auxiliary-loss normalization in MoE models." }, { "left": "issue:43572", "right": "issue:46178", "accept": false, - "reason": "StableLm pad token config regression is unrelated to a missing backend entry for `tensorflow_text`." + "reason": "Both involve config/backend compatibility, but StableLm pad_token_idx regression and missing tensorflow_text backend registration are separate issues." }, { "left": "issue:43673", "right": "issue:44743", "accept": false, - "reason": "Both mention cache/state behavior, but one is missing cache in chunked prefill and the other is recurrent-state reset in modular Qwen3.5; not the same bug." + "reason": "Both touch cache/state machinery, but one is missing cache during chunked prefill and the other is recurrent states resetting in modular_qwen3_5." }, { "left": "issue:44617", "right": "issue:45584", "accept": false, - "reason": "Sam3Video OOM is unrelated to Whisper empty-transcription generation failure." + "reason": "Different products and symptoms: Sam3Video OOM vs Whisper empty-transcription generation failure." }, { "left": "issue:43504", "right": "issue:45698", "accept": false, - "reason": "Both are load/save related, but Beit legacy-field preset loading and wrong custom-module restoration are distinct issues." + "reason": "Both are load/save issues, but BEiT legacy preset loading and wrong custom-module resolution after save_pretrained are not the same concrete bug." }, { "left": "issue:44734", "right": "issue:45406", "accept": false, - "reason": "Both involve `transformers serve`, but KV-cache continuation indexing and Gemma4Processor missing `_tokenizer` are different crash causes." + "reason": "Different transformers serve crashes: KV-cache continuation indexing vs Gemma4Processor missing _tokenizer." }, { "left": "issue:45542", "right": "issue:46234", "accept": false, - "reason": "TensorBoard/TensorFlow backend detection is unrelated to the missing `is_falcon_mamba_ssm_available` import utility symbol." + "reason": "Unrelated backend/import issues: undefined tf backend from tensorboard-only install vs missing falcon_mamba_ssm availability helper." }, { "left": "issue:43673", "right": "issue:46421", "accept": false, - "reason": "Both concern chunked prefill/cache initialization, but one reports missing cache in generation and the other reports spurious recompilation; not clearly the same underlying defect." + "reason": "Both mention chunked prefill and cache, but one reports missing cache in v5.0.0 while the other reports lazy-init recompilation; different observable failures." }, { "left": "issue:43525", "right": "issue:43531", "accept": false, - "reason": "Llama4 missing `pad_token_id` and Qwen3-MoE sliding-window behavior are different config/runtime problems." + "reason": "Both are config-field regressions, but they affect different models and different missing settings: Llama4 pad_token_id vs Qwen3-MoE sliding_window." }, { "left": "issue:44792", "right": "issue:44977", "accept": false, - "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation issue are unrelated." + "reason": "Different models and paths: janus image-generation test failure vs Qwen3.5 flash-attention generation bug." }, { "left": "issue:45381", "right": "issue:46369", "accept": false, - "reason": "Qwen2.5-VL video position-id regression and Sam3 LiteText memory overflow are different model-specific issues." + "reason": "Different domains and failure modes: Qwen2.5-VL video vision_position_ids bug vs sam3-litetext memory overflow." }, { "left": "issue:43761", "right": "issue:44242", "accept": false, - "reason": "CLIPVision hidden-state regression is unrelated to auxiliary-loss gating on `output_router_logits`." + "reason": "Same general MoE/regression theme is too broad; CLIPVision hidden_states and router load-balancing loss are unrelated concrete issues." }, { "left": "issue:45563", "right": "issue:45584", "accept": false, - "reason": "A stale warning in paged generate is not the same as Whisper failing on empty transcription after special-token alignment." + "reason": "Both are generation-related, but a stale warning for num_return_sequences is not the same bug as Whisper failing on empty transcription after special-token alignment." } ] }, @@ -10314,25 +10333,25 @@ "soft_pairs": [ "issue:44792|issue:45600", "issue:44534|issue:44913", - "issue:44190|issue:44964", "issue:43931|issue:44060", + "issue:44190|issue:44964", "issue:35141|issue:42490", "issue:44855|issue:45588", - "issue:43827|issue:43994", "issue:44479|issue:44734", + "issue:43827|issue:43994", "issue:41720|issue:45245", "issue:44617|issue:46369", "issue:43065|issue:44610", - "issue:44464|issue:45563", "issue:44521|issue:45381", + "issue:44464|issue:45563", "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:25251|issue:27957", "issue:44462|issue:45907", + "issue:25251|issue:27957", "issue:43540|issue:44734", "issue:44112|issue:45600", "issue:45632|issue:46234", @@ -10340,208 +10359,196 @@ "issue:44737|issue:45803", "issue:43116|issue:44190" ], - "prepared_review_unit_hash": "ea773e17c7a2abfff7529e0b9d88761cd5e8a226c87d41f6be9b74cb06b08dad", + "prepared_review_unit_hash": "e5b90b5d9db4a8c720234c94cd2380a32680764873ea3873f0b4ef17248b9ddb", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 27, - "serialized_chars": 35175, - "estimated_input_tokens": 8794, - "estimated_eval_tokens": 17844 + "serialized_chars": 35181, + "estimated_input_tokens": 8796, + "estimated_eval_tokens": 17848 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous and only contains two true duplicate pairs: the non-persistent-buffer junk regression (43644 \u2194 44534) and the Qwen video-input regression (44479 \u2194 45381). All other soft edges are distinct issues.", - "confidence": 0.94, - "canonical_issue_reason": "No single canonical issue for the whole set: the cluster is heterogeneous. The only clear duplicate anchors are 43644\u219444534 and 44479\u219445381.", + "summary": "This cluster is mostly heterogeneous: the candidate pairs span unrelated bugs in model loading, config handling, generation, docs, CI, and example scripts. None of the soft edges are safe to collapse into one duplicate group.", + "confidence": 0.96, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No single global best issue exists for this cluster. If forced to pick a representative for one sub-bug, 44479 is the broadest accepted duplicate pair, but it only covers the Qwen video-regression slice.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44792", "right": "issue:45600", "accept": false, - "reason": "Different failures: Janus image-generation test regression vs. Sam3LiteText CI break from stale auto_mappings entries." + "reason": "Different failures: Janus image-generation test vs Sam3LiteText auto-mapping/CI break." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Both are config-related, but they describe different bugs: non-persistent-buffer junk vs. GPTNeoX rotary_pct not persisting on reload." + "reason": "One is a v5 non-persistent-buffer initialization bug; the other is GPTNeoXConfig reload losing rotary_pct." }, { - "left": "issue:44190", - "right": "issue:44964", + "left": "issue:43931", + "right": "issue:44060", "accept": false, - "reason": "Unrelated loading problems: local dataset loading in an example script vs. loading Phi-4-multimodal-instruct." + "reason": "Different models and code paths: Qwen3-VL shape mismatch vs Qwen3-Next tied-weights warning." }, { - "left": "issue:43931", - "right": "issue:44060", + "left": "issue:44190", + "right": "issue:44964", "accept": false, - "reason": "Different Qwen3 issues: weight-shape mismatch on model load vs. an incorrect tied-weights warning." + "reason": "Different loaders and symptoms: example-script local dataset issue vs multimodal model load failure." }, { "left": "issue:35141", "right": "issue:42490", "accept": false, - "reason": "Different subsystems and symptoms: embedding reinitialization in post_init vs. trailing-slash local-model import warning." + "reason": "Embedding reinit on resize is unrelated to the local-path trailing-slash import warning." }, { "left": "issue:44855", "right": "issue:45588", "accept": false, - "reason": "Different code paths: Python 3.13 torch.jit parsing error vs. flash_attention.py AttributeError on s_aux=None." + "reason": "Python 3.13 import parse error in DebertaV2Model is unrelated to flash-attention crashing on s_aux=None." }, { - "left": "issue:43827", - "right": "issue:43994", + "left": "issue:44479", + "right": "issue:44734", "accept": false, - "reason": "Docs cleanup about removed pipeline() references is not the same bug as incorrect SigLIP pipeline outputs." + "reason": "Different subsystems: video-input regression in Qwen VL models vs serve KV-cache continuation indexing." }, { - "left": "issue:44479", - "right": "issue:44734", + "left": "issue:43827", + "right": "issue:43994", "accept": false, - "reason": "Different bugs: Qwen video-input regression vs. transformers serve KV-cache continuation tensor indexing crash." + "reason": "Docs still referencing pipeline() is not the same as SigLIP2 producing bad outputs with AutoModel/pipeline." }, { "left": "issue:41720", "right": "issue:45245", "accept": false, - "reason": "Completely different failures: Qwen3 auto device mapping cuda assert vs. category-count limit error." + "reason": "Qwen3 auto device-mapping CUDA assert is unrelated to the 2^24 category-count runtime limit." }, { "left": "issue:44617", "right": "issue:46369", "accept": false, - "reason": "Both mention memory pressure in Sam3 variants, but the affected models and failure modes differ; not the same concrete bug." + "reason": "Same family name, but different model variants and likely different memory causes; too broad to merge." }, { "left": "issue:43065", "right": "issue:44610", "accept": false, - "reason": "Different problems: dummy Conv2d in Sam3PixelDecoder vs. processor/model input-size mismatch for OmDet-Turbo." + "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to the OmDet-Turbo processor/model resolution mismatch." }, { - "left": "issue:44464", - "right": "issue:45563", + "left": "issue:44521", + "right": "issue:45381", "accept": false, - "reason": "Different generation issues: inconsistent chunked generation under compiled forward vs. a stale num_return_sequences warning." + "reason": "Both are multimodal, but one is assistant_mask construction and the other is vision_position_ids for qwen2.5-vl video input." }, { - "left": "issue:44521", - "right": "issue:45381", + "left": "issue:44464", + "right": "issue:45563", "accept": false, - "reason": "Both are multimodal/Qwen-related, but assistant mask generation and video vision_position_ids are different bugs." + "reason": "Chunked-generation inconsistency under compile is unrelated to the stale num_return_sequences warning." }, { "left": "issue:44933", "right": "issue:45362", "accept": false, - "reason": "Unrelated: missing image_utils import vs. Qwen3.5-35B crashing with transformers chat." + "reason": "Missing image_utils import is unrelated to Qwen3.5 chat crashes." }, { "left": "issue:43644", "right": "issue:44913", "accept": false, - "reason": "43744 is about junk-filled non-persistent buffers; 44913 is about rotary_pct resetting on config reload." + "reason": "Non-persistent-buffer junk filling is a different bug from rotary_pct resetting on reload." }, { "left": "issue:43976", "right": "issue:44855", "accept": false, - "reason": "Different platform/version issues: Python 3.9+ compatibility claim vs. Python 3.13 IndentationError from torch.jit.script." + "reason": "Python-version support breakage is unrelated to the DebertaV2Model IndentationError." }, { "left": "issue:44077", "right": "issue:44877", "accept": false, - "reason": "Both concern config handling, but one is patchtsmixer post_init policy and the other is granite_speech strict-config loading." + "reason": "PatchTSMixer post_init policy change is unrelated to strict granite_speech config loading." }, { "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "Different areas entirely: Bamba torch.compile + SDPA failure vs. example script loading a local dataset." + "reason": "Torch.compile+SDPA failure in Bamba is unrelated to the image-classification example local-dataset issue." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "Different model-management bugs: LoRA merge with extended vocab collapse vs. save_pretrained writing wrong visual encoder keys." + "reason": "LoRA merge collapse with extended vocab is unrelated to incorrect visual encoder keys in save_pretrained." }, { - "left": "issue:25251", - "right": "issue:27957", + "left": "issue:44462", + "right": "issue:45907", "accept": false, - "reason": "Different feature areas: pipeline top_k nesting behavior vs. XLMRoberta Flash Attention 2 support." + "reason": "Tokenizer loading ignoring tokenizer.json is unrelated to list_repo_files failing to forward revision." }, { - "left": "issue:44462", - "right": "issue:45907", + "left": "issue:25251", + "right": "issue:27957", "accept": false, - "reason": "Different repository I/O bugs: AutoTokenizer ignoring tokenizer.json vs. list_repo_files not forwarding revision." + "reason": "Pipeline top_k nesting change and XLMRoberta Flash Attention 2 support are unrelated." }, { "left": "issue:43540", "right": "issue:44734", "accept": false, - "reason": "Different video/chat serving failures: Qwen3OmniMoe video processing ValueError vs. serve KV-cache indexing crash." + "reason": "Qwen3OmniMoe video-input ValueError is unrelated to serve-side KV-cache continuation indexing." }, { "left": "issue:44112", "right": "issue:45600", "accept": false, - "reason": "CI failures in different places: stale GraniteSpeech device override test vs. removed Sam3LiteText config references." + "reason": "Both touch CI, but they concern different stale references in different models/modules." }, { "left": "issue:45632", "right": "issue:46234", "accept": false, - "reason": "Different import/cache issues: trust_remote_code cache path collision vs. missing is_falcon_mamba_ssm_available helper." + "reason": "Cache-path collision for trust_remote_code is unrelated to the missing import_utils helper." }, { "left": "issue:43335", "right": "issue:44190", "accept": false, - "reason": "Different bugs: SwitchTransformers sparse-layer creation at zero layers vs. local dataset loading failure." + "reason": "SwitchTransformers sparse-layer bug is unrelated to the example-script dataset loading failure." }, { "left": "issue:44737", "right": "issue:45803", "accept": false, - "reason": "Different runtime problems: XLNet device placement bug vs. Fuyu bare-except error handling bug." + "reason": "XLNet CPU tensor allocation bug is unrelated to the bare-except KeyboardInterrupt swallowing issue." }, { "left": "issue:43116", "right": "issue:44190", "accept": false, - "reason": "Both are example-script failures, but they affect different scripts and failure mechanisms; not the same bug." - }, - { - "left": "issue:43644", - "right": "issue:44534", - "accept": true, - "reason": "Same underlying regression: Transformers v5 filling non-persistent buffers with junk; 44534 is a reworded duplicate of 43644." - }, - { - "left": "issue:44479", - "right": "issue:45381", - "accept": true, - "reason": "Same Qwen video-input regression in transformers 5.3.0; 45381 is a narrower symptom report of the bug described in 44479." + "reason": "Different example scripts and different bugs: empty multi-label results vs local dataset loading." } ] }, @@ -10572,8 +10579,10 @@ "issue:41628", "issue:41950", "issue:42175", + "issue:42907", "issue:43295", "issue:43404", + "issue:43441", "issue:43479", "issue:43526", "issue:43531", @@ -10599,7 +10608,6 @@ "issue:45216", "issue:45245", "issue:45276", - "issue:45335", "issue:45375", "issue:45446", "issue:45542", @@ -10614,14 +10622,14 @@ "issue:46424" ], "soft_pairs": [ - "issue:45276|issue:45584", "issue:43295|issue:45276", + "issue:45276|issue:45584", "issue:44938|issue:45542", "issue:46082|issue:46121", "issue:36246|issue:41628", "issue:45632|issue:46421", - "issue:45245|issue:45584", "issue:42175|issue:44279", + "issue:45245|issue:45584", "issue:43404|issue:45216", "issue:43650|issue:44220", "issue:44246|issue:45003", @@ -10634,211 +10642,211 @@ "issue:43673|issue:46424", "issue:44964|issue:45600", "issue:43854|issue:44964", - "issue:44112|issue:44964", "issue:44016|issue:44220", + "issue:44112|issue:44964", "issue:27957|issue:35141", "issue:46123|issue:46369", "issue:43531|issue:45600", "issue:43976|issue:44987", "issue:43526|issue:43901", - "issue:43295|issue:45335" + "issue:42907|issue:43441" ], - "prepared_review_unit_hash": "19978f91ce5d2b22f1807588ededf3b0e3103d7a7793cc1379a1b67f222bcce5", + "prepared_review_unit_hash": "855bbc891532d9a82cdbce666a4c4710bc341858367f15ec3463f8ff823d6f8a", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 28, - "serialized_chars": 35530, - "estimated_input_tokens": 8883, - "estimated_eval_tokens": 18022 + "serialized_chars": 36072, + "estimated_input_tokens": 9018, + "estimated_eval_tokens": 18292 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the soft pairs mostly connect superficially similar but distinct regressions across model loading, config handling, embeddings, cache behavior, and pipelines. I found no true duplicate pair to merge, and there are no PRs in the cluster.", - "confidence": 0.86, - "canonical_issue_reason": "issue:44964 is the clearest representative user-facing model-loading regression in the set, though the overall cluster is not a true duplicate group.", + "summary": "This cluster is only loosely related by Transformers API/regression themes; the paired issues mostly describe different models, components, or failure modes rather than the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 44246 is the broadest and most representative of the set\u2019s general import/regression theme, but it is still not a true duplicate anchor for the other issues.", "canonical_pr_reason": null, - "best_issue_reason": "It is concrete, reproducible, and broadly relevant to current multimodal model loading, making it the best single issue representative here.", + "best_issue_reason": "Issue 44246 has the widest scope and highest discussion activity among the candidates, making it the best single representative issue for the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45276", - "right": "issue:45584", + "left": "issue:43295", + "right": "issue:45276", "accept": false, - "reason": "Different bugs: token embedding resize propagation in Gemma4 vs Whisper empty-transcription generation failure after token alignment." + "reason": "Both involve embedding/resize behavior, but they are different bugs in different code paths and models." }, { - "left": "issue:43295", - "right": "issue:45276", + "left": "issue:45276", + "right": "issue:45584", "accept": false, - "reason": "Different regressions: processor/tokenizer API breakage vs resize_token_embeddings not updating Gemma4 embeddings." + "reason": "Unrelated failures: token embedding resize vs Whisper empty-transcription handling after special-token alignment." }, { "left": "issue:44938", "right": "issue:45542", "accept": false, - "reason": "Unrelated symptoms: Python 3.14 import/load failure vs TensorBoard-only install causing missing TF backend handling." + "reason": "Python version load failure and missing TensorFlow backend are unrelated dependency problems." }, { "left": "issue:46082", "right": "issue:46121", "accept": false, - "reason": "Different code paths: LlamaConfig head_dim validation vs rope-parameter list handling in convert_rope_params_to_dict." + "reason": "Different config validation paths: Llama head_dim rules vs rope-parameter JSON list handling." }, { "left": "issue:36246", "right": "issue:41628", "accept": false, - "reason": "Different import errors for different symbols and model families; not the same underlying bug." + "reason": "Both are import errors, but they concern different symbols and different modules." }, { "left": "issue:45632", "right": "issue:46421", "accept": false, - "reason": "Local trust-remote-code cache path collision is unrelated to cache lazy initialization causing recompilation." + "reason": "One is a trust_remote_code cache-path collision; the other is cache lazy init causing recompilation." }, { - "left": "issue:45245", - "right": "issue:45584", + "left": "issue:42175", + "right": "issue:44279", "accept": false, - "reason": "Category-cardinality runtime error is unrelated to Whisper empty transcription generation behavior." + "reason": "One is about TensorFlow not being included in a torch install extra; the other is a generic dependency issue." }, { - "left": "issue:42175", - "right": "issue:44279", + "left": "issue:45245", + "right": "issue:45584", "accept": false, - "reason": "Different dependency issues: missing TensorFlow in torch install extras vs a general transformers dependency complaint." + "reason": "Category-count overflow and empty-transcription generation are completely different failures." }, { "left": "issue:43404", "right": "issue:45216", "accept": false, - "reason": "Different regressions: Mistral3 lm_head tying/save-pretrained issue vs Qwen3.5 checkpoint save corruption." + "reason": "Both mention tied weights, but one is model-specific lm_head tying and the other is a save_pretrained regression." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "These are unrelated; the titles/body hints do not point to the same audio feature extraction bug." + "reason": "No shared bug surface; this looks like a stray/nontechnical issue versus an audio feature extraction bug." }, { "left": "issue:44246", "right": "issue:45003", "accept": false, - "reason": "Import slowness from sys.modules access is a different problem than unsafe sys.modules access in modeling_utils." + "reason": "Both touch import behavior, but one is slowness and the other is unsafe sys.modules access." }, { "left": "issue:36246", "right": "issue:38617", "accept": false, - "reason": "Different missing imports from different modules; no evidence of one shared fix." + "reason": "Different missing imports with no common underlying code-path." }, { "left": "issue:41950", "right": "issue:44016", "accept": false, - "reason": "Video-classification pipeline selecting image processors is unrelated to a notebook syntax error." + "reason": "Video-classification pipeline bug and notebook syntax error are unrelated." }, { "left": "issue:44186", "right": "issue:44734", "accept": false, - "reason": "Different failure modes: LayoutLMv2 tokenizer padding/NER crash vs serve KV-cache continuation tensor indexing bug." + "reason": "Tokenizer padding/NER crash and KV-cache continuation indexing are distinct code paths." }, { "left": "issue:43479", "right": "issue:45375", "accept": false, - "reason": "Different config initialization bug vs silent field drop by @strict in a separate config class." + "reason": "Both involve config defaults/fields, but one is default initialization and the other is strict-field dropping." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Different flex_attention issues: compiled flex_attention on torch>=2.9 vs incorrect version check for AuxRequest import." + "reason": "Compiled flex_attention breakage on torch>=2.9 is not the same as a version-check bug for AuxRequest import." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "AutoConfig missing fields with trust_remote_code is unrelated to unrecognized model/config lacking model_type." + "reason": "Missing fields with trust_remote_code and unrecognized model_type are different config-loading problems." }, { "left": "issue:43673", "right": "issue:46424", "accept": false, - "reason": "Chunked prefill cache-missing regression is different from max_cache_length only applying to Quantized Cache." + "reason": "Chunked-prefill cache absence and max_cache_length being ignored outside quantized cache are different cache bugs." }, { "left": "issue:44964", "right": "issue:45600", "accept": false, - "reason": "Phi-4 multimodal loading failure and removed Sam3LiteText auto_mappings references are different model/config breakages." + "reason": "Model loading failure and auto_mappings stale references/CI breakage do not describe the same issue." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "Both are model loading failures, but for different models and different root causes; no shared code-path evidence." + "reason": "Different models and symptoms; one is a unit-test loading problem, the other is a general latest-transformers load failure." }, { - "left": "issue:44112", - "right": "issue:44964", + "left": "issue:44016", + "right": "issue:44220", "accept": false, - "reason": "CI device override test flake in GraniteSpeech is unrelated to loading Phi-4 multimodal." + "reason": "Notebook syntax error is unrelated to the fbank feature extraction issue." }, { - "left": "issue:44016", - "right": "issue:44220", + "left": "issue:44112", + "right": "issue:44964", "accept": false, - "reason": "Notebook syntax error and audio feature extraction bug are unrelated." + "reason": "CI stale device override and model loading are unrelated." }, { "left": "issue:27957", "right": "issue:35141", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta is unrelated to token embedding resize reinitialization." + "reason": "Flash Attention support for XLMRoberta and embedding reinitialization on resize are different bugs." }, { "left": "issue:46123", "right": "issue:46369", "accept": false, - "reason": "Partial-batch result dropping in mask generation is unrelated to sam3-litetext memory overflow." + "reason": "Partial-batch result dropping and model memory overflow are unrelated." }, { "left": "issue:43531", "right": "issue:45600", "accept": false, - "reason": "Sliding-window issue in Qwen3-MoE does not match removed-config CI breakage." + "reason": "Qwen3-MoE sliding_window behavior and stale auto_mappings references are different issues." }, { "left": "issue:43976", "right": "issue:44987", "accept": false, - "reason": "Python-version support complaint is unrelated to a specific model load failure for physical-intelligence/fast." + "reason": "Python-version compatibility and a specific model-load failure are not duplicates." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to outdated TextClassificationPipeline docs." + "reason": "ImageProcessor label reduction and pipeline docs behavior mismatch are unrelated." }, { - "left": "issue:43295", - "right": "issue:45335", + "left": "issue:42907", + "right": "issue:43441", "accept": false, - "reason": "Different architectures and failures: processor/tokenizer regression vs t5gemma decoder embedding resize propagation." + "reason": "Dequantized save failure and FlashAttention failure in Ministral are distinct model-path problems." } ] }, @@ -10867,7 +10875,6 @@ "issue:38175", "issue:39692", "issue:41950", - "issue:42907", "issue:43295", "issue:43299", "issue:43381", @@ -10884,6 +10891,7 @@ "issue:44242", "issue:44279", "issue:44336", + "issue:44351", "issue:44368", "issue:44496", "issue:44617", @@ -10895,6 +10903,7 @@ "issue:45103", "issue:45200", "issue:45216", + "issue:45278", "issue:45335", "issue:45357", "issue:45372", @@ -10911,18 +10920,18 @@ "issue:46424" ], "soft_pairs": [ - "issue:42907|issue:43441", + "issue:43295|issue:45335", "issue:28282|issue:35141", "issue:44496|issue:45042", "issue:43441|issue:45588", "issue:44016|issue:44062", "issue:45372|issue:46234", - "issue:43901|issue:44496", "issue:43299|issue:45216", - "issue:43673|issue:44336", - "issue:43381|issue:45593", "issue:44792|issue:46424", + "issue:43901|issue:44496", + "issue:43381|issue:45593", "issue:43824|issue:44933", + "issue:43673|issue:44336", "issue:43976|issue:44279", "issue:44496|issue:45759", "issue:45335|issue:45584", @@ -10930,212 +10939,219 @@ "issue:46082|issue:46399", "issue:43650|issue:46209", "issue:43577|issue:45216", + "issue:38175|issue:44617", "issue:45103|issue:45372", "issue:44242|issue:45200", - "issue:38175|issue:44617", "issue:44279|issue:45405", + "issue:41950|issue:45020", "issue:43295|issue:44779", "issue:43299|issue:45357", - "issue:41950|issue:45020", "issue:44779|issue:45812", + "issue:44351|issue:45278", "issue:39692|issue:43901" ], - "prepared_review_unit_hash": "bc4f59310fb52d22d401a7b6c42f879a91a72c45706414568a7e12eb8a426188", + "prepared_review_unit_hash": "5739e3405151239b93f628e0c356bceb5f6b65cb1add861dc5c8bf0645d96a26", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35627, - "estimated_input_tokens": 8907, - "estimated_eval_tokens": 18070 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36467, + "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 a loose collection of unrelated Transformers issues rather than a duplicate set. The items share only a broad theme of model/version regressions, but the underlying bugs span loading, tokenization, caching, docs, quantization, and model-specific failures. No soft-edge pair looks safe to merge.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue fits the cluster well; issue 45020 is the broadest umbrella, but it is still too generic and does not represent the other distinct bugs as true duplicates.", + "summary": "This cluster is a loose collection of mostly unrelated Transformers issues (imports, model-loading regressions, tokenizer bugs, docs, and a few training/inference edge cases). There are no PRs here, and the soft pairs do not appear to be true duplicates of the same underlying bug/change.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45020 is the broadest umbrella item (recent-version regressions affecting remote_code models), but the cluster is heterogeneous and does not form a clean duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the closest thing to a representative issue because it is the broadest regression report, but it is still not a good canonical duplicate target for this heterogeneous cluster.", + "best_issue_reason": "issue:45020 is the best single representative of the cluster\u2019s general regression theme, but it should not be treated as a true canonical duplicate for the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42907", - "right": "issue:43441", + "left": "issue:43295", + "right": "issue:45335", "accept": false, - "reason": "Both mention Ministral/FlashAttention, but one is about saving dequantized models and the other is an inference crash; different code paths and fixes." + "reason": "Different failures: processor/tokenizer regression with images vs resize_token_embeddings not updating decoder.embed_tokens." }, { "left": "issue:28282", "right": "issue:35141", "accept": false, - "reason": "Completely different problems: missing PyTorch import vs embedding reinitialization during resize." + "reason": "Missing PyTorch import is unrelated to embedding reinitialization after resize." }, { "left": "issue:44496", "right": "issue:45042", "accept": false, - "reason": "One is an unrecognized-model/config issue, the other is a PIL/torchvision dependency regression in image processors." + "reason": "Unrecognized model/config loading error is unrelated to PIL image processors requiring torchvision." }, { "left": "issue:43441", "right": "issue:45588", "accept": false, - "reason": "Both touch FlashAttention, but the failures are model-/path-specific and not the same underlying bug." + "reason": "Both mention flash attention, but one is a Ministral-specific crash and the other is an s_aux=None AttributeError in sink-less models." }, { "left": "issue:44016", "right": "issue:44062", "accept": false, - "reason": "Notebook syntax error and tokenizer argument error are unrelated." + "reason": "Notebook syntax error and AddedToken special-argument TypeError are distinct bugs." }, { "left": "issue:45372", "right": "issue:46234", "accept": false, - "reason": "Different missing-import regressions affecting different subsystems and models." + "reason": "Different missing imports in different subsystems: Gemma processor loading vs falcon mamba availability helper." }, { - "left": "issue:43901", - "right": "issue:44496", + "left": "issue:43299", + "right": "issue:45216", "accept": false, - "reason": "Docs mismatch for TextClassificationPipeline vs model config/import failure; unrelated." + "reason": "Qwen3VL model loading regression is unrelated to Qwen3.5 save_pretrained checkpoint corruption." }, { - "left": "issue:43299", - "right": "issue:45216", + "left": "issue:44792", + "right": "issue:46424", "accept": false, - "reason": "Both are Qwen3.5 regressions, but one is model loading and the other is save_pretrained/checkpoint serialization." + "reason": "A janus test failure is not the same as cache max length being ignored outside quantized cache." }, { - "left": "issue:43673", - "right": "issue:44336", + "left": "issue:43901", + "right": "issue:44496", "accept": false, - "reason": "Generation cache issue vs ANSI formatting in loading_report; unrelated." + "reason": "Documentation mismatch for return_all_scores is unrelated to an unrecognized model config error." }, { "left": "issue:43381", "right": "issue:45593", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction is unrelated to D-FINE auxiliary-loss behavior." + "reason": "Gradient checkpointing eval-mode behavior is unrelated to D-FINE auxiliary-loss handling." }, { - "left": "issue:44792", - "right": "issue:46424", + "left": "issue:43824", + "right": "issue:44933", "accept": false, - "reason": "A failing Janus image-generation test and a cache-length setting bug are not the same issue." + "reason": "Both are import errors, but they concern different symbols and different failing modules." }, { - "left": "issue:43824", - "right": "issue:44933", + "left": "issue:43673", + "right": "issue:44336", "accept": false, - "reason": "Different import errors in different modules." + "reason": "Cache-missing failure during chunked prefill is unrelated to ANSI codes in loading_report." }, { "left": "issue:43976", "right": "issue:44279", "accept": false, - "reason": "Python-version support bug vs generic dependency issue; not the same defect." + "reason": "Python-version compatibility and a generic dependency issue are not the same underlying bug." }, { "left": "issue:44496", "right": "issue:45759", "accept": false, - "reason": "Both are model/config loading issues, but one is missing model_type and the other is incorrect text_config unwrapping for composite multimodal configs." + "reason": "Different model-loading paths: unrecognized config vs missing text_config unwrap in composite Qwen configs." }, { "left": "issue:45335", "right": "issue:45584", "accept": false, - "reason": "Token-embedding resize bug and Whisper empty-transcription generation bug are unrelated." + "reason": "Token embedding resize bug is unrelated to Whisper empty-transcription generation failure." }, { "left": "issue:44075", "right": "issue:44368", "accept": false, - "reason": "SGD argument handling vs tie_word_embeddings warning; different code paths." + "reason": "Optimizer argument handling and tie_word_embeddings warnings are distinct issues." }, { "left": "issue:46082", "right": "issue:46399", "accept": false, - "reason": "LlamaConfig validation and Qwen3.5 kernel registration are different bugs." + "reason": "Llama head_dim validation and Qwen3.5 kernel registration are unrelated." }, { "left": "issue:43650", "right": "issue:46209", "accept": false, - "reason": "Placeholder/data issue vs FP8 inference failure; no relation." + "reason": "Placeholder data request is unrelated to FP8 inference failure on B200." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "BLIP2 dtype propagation bug vs Qwen3.5 save_pretrained regression; different models and behavior." + "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained checkpoint issues do not match." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Zero-probabilities in SigLIP2 is unrelated to Sam3Video CUDA OOM." }, { "left": "issue:45103", "right": "issue:45372", "accept": false, - "reason": "Docstring annotation crash and Gemma 4 processor import failure are unrelated." + "reason": "auto_docstring annotation parsing crash is unrelated to Gemma processor import failure." }, { "left": "issue:44242", "right": "issue:45200", "accept": false, - "reason": "Load-balancing loss omission and mm_token_type_ids defaulting are different training-time bugs." + "reason": "Mixture-of-experts load-balancing loss and Gemma 4 mm_token_type_ids are different model-training bugs." }, { - "left": "issue:38175", - "right": "issue:44617", + "left": "issue:44279", + "right": "issue:45405", "accept": false, - "reason": "SigLIP2 zero probabilities and Sam3Video OOM are unrelated model failures." + "reason": "A transformers dependency problem is unrelated to a PEFT minimum-version bump." }, { - "left": "issue:44279", - "right": "issue:45405", + "left": "issue:41950", + "right": "issue:45020", "accept": false, - "reason": "Generic dependency issue and a specific MIN_PEFT_VERSION bump are not the same underlying problem." + "reason": "Video-classification processor lookup bug is unrelated to remote_code model breakage." }, { "left": "issue:43295", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer/processor regressions, but they affect different model families and different APIs." + "reason": "Processor/tokenizer regression with images is unrelated to Deepseek tokenizer output changes." }, { "left": "issue:43299", "right": "issue:45357", "accept": false, - "reason": "Both are Qwen3.5 regressions, but one is loading failure and the other is incorrect visual-encoder keys on save_pretrained." + "reason": "Qwen3VL loading and Qwen3.5 visual-encoder key saving are different code paths." }, { - "left": "issue:41950", - "right": "issue:45020", + "left": "issue:44779", + "right": "issue:45812", "accept": false, - "reason": "Video-classification processor lookup bug is unrelated to remote_code model breakages." + "reason": "Both are tokenizer regressions, but they affect different model families and likely different root causes." }, { - "left": "issue:44779", - "right": "issue:45812", + "left": "issue:44351", + "right": "issue:45278", "accept": false, - "reason": "Different tokenizer regressions on different model families; not a single fix." + "reason": "A specific missing HybridCache import is not enough to treat this as the same bug as a broad import-error report." }, { "left": "issue:39692", "right": "issue:43901", "accept": false, - "reason": "SigLIP2 doc/example errors and TextClassificationPipeline docs mismatch are unrelated documentation issues." + "reason": "A SigLIP2 example with multiple errors is unrelated to a docs mismatch about return_all_scores." } ] }, @@ -11161,6 +11177,7 @@ "nodes": [ "issue:33290", "issue:41950", + "issue:42175", "issue:42994", "issue:43299", "issue:43317", @@ -11182,7 +11199,6 @@ "issue:44279", "issue:44291", "issue:44297", - "issue:44351", "issue:44361", "issue:44462", "issue:44466", @@ -11194,7 +11210,6 @@ "issue:44779", "issue:45081", "issue:45276", - "issue:45278", "issue:45325", "issue:45397", "issue:45405", @@ -11209,7 +11224,6 @@ "issue:46132" ], "soft_pairs": [ - "issue:44351|issue:45278", "issue:44521|issue:45325", "issue:41950|issue:45823", "issue:43479|issue:43525", @@ -11220,8 +11234,8 @@ "issue:43299|issue:43854", "issue:43450|issue:44279", "issue:43650|issue:46132", - "issue:44466|issue:45276", "issue:45081|issue:45907", + "issue:44466|issue:45276", "issue:44016|issue:44291", "issue:43441|issue:45834", "issue:42994|issue:43317", @@ -11235,198 +11249,206 @@ "issue:44671|issue:44779", "issue:44016|issue:45823", "issue:43723|issue:44162", - "issue:43901|issue:43994" + "issue:42175|issue:43976", + "issue:43901|issue:43994", + "issue:43441|issue:44683" ], - "prepared_review_unit_hash": "72c57b9b0d0999b4695adeede64c850dfe8d7ab859e223f1b3e92a000f064c6a", + "prepared_review_unit_hash": "f2d99915d9072d61d264d7676da37e5ddf552ba50d39762c70333bbeacccaf45", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 27, - "serialized_chars": 35623, - "estimated_input_tokens": 8906, - "estimated_eval_tokens": 18068 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35414, + "estimated_input_tokens": 8854, + "estimated_eval_tokens": 17964 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is heterogeneous: the items span unrelated bugs in tokenizers, multimodal models, deepspeed, docs, CI, and config handling. The soft pairs are mostly broad thematic matches from v5 regressions, not the same concrete underlying issue, so none should be merged as duplicates.", - "confidence": 0.98, - "canonical_issue_reason": null, + "summary": "This cluster is heterogeneous: it mixes unrelated Transformer v5 regressions, tokenizer/model-loading bugs, processor/config issues, deepspeed training behavior, packaging, and docs. The soft-similarity links are mostly superficial and do not look like true duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue: the cluster mixes unrelated bugs across tokenizers, multimodal processing, config, deepspeed, packaging, and docs.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:44521 is the most active/referenced node, but it is not a true duplicate hub and should not be treated as a canonical merge target.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44351", - "right": "issue:45278", - "accept": false, - "reason": "44351 is a specific missing HybridCache import; 45278 is a broad post-upgrade import-error report. Related area, but not the same bug." - }, { "left": "issue:44521", "right": "issue:45325", "accept": false, - "reason": "Both involve multimodal/video behavior, but one is assistant mask generation and the other is Qwen2.5-VL rope index scaling. Different code paths." + "reason": "Both are multimodal-related, but one is about all-zero assistant masks and the other about temporal position_id scaling; different code paths and fixes." }, { "left": "issue:41950", "right": "issue:45823", "accept": false, - "reason": "Video-classification processor lookup and Gemma4 device-mismatch on forward are unrelated problems." + "reason": "Video-classification processor selection and Gemma4 device-mismatch during forward are unrelated bugs." }, { "left": "issue:43479", "right": "issue:43525", "accept": false, - "reason": "One is default multimodal config initialization; the other is a missing pad_token_id attribute on Llama4Config. Not the same failure mode." + "reason": "Different config problems: multimodal defaults vs missing pad_token_id on Llama4Config." }, { "left": "issue:44683", "right": "issue:45542", "accept": false, - "reason": "Torch flex_attention compilation failure and tensorboard/TensorFlow backend error are unrelated." + "reason": "Torch flex_attention compilation on 2.9 and TensorFlow packaging/backend availability are unrelated." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python version incompatibility and a bumped PEFT dependency version are different issues." + "reason": "Python version support regression and a bumped PEFT dependency are separate release issues." }, { "left": "issue:44462", "right": "issue:45081", "accept": false, - "reason": "Tokenizer.json precedence vs. a Mistral regex patch crash are both tokenizer-related, but they are distinct bugs." + "reason": "AutoTokenizer ignoring repository tokenizer.json and Mistral regex patch crashing are different loading paths." }, { "left": "issue:44297", "right": "issue:44462", "accept": false, - "reason": "Save-pretrained tokenizer_class mismatch and AutoTokenizer ignoring tokenizer.json are separate serialization/loading problems." + "reason": "Tokenizer save metadata mismatch is not the same bug as AutoTokenizer ignoring tokenizer.json." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "Both are model-loading regressions, but they affect different models and different loading paths." + "reason": "Different model-loading failures for different models; no shared concrete code-path." }, { "left": "issue:43450", "right": "issue:44279", "accept": false, - "reason": "Incorrect batched video-processor shape and a dependency issue are unrelated." + "reason": "Video processor batched-shape bug and a vague dependency issue are not the same defect." }, { "left": "issue:43650", "right": "issue:46132", "accept": false, - "reason": "A vague data request is not the same as an AttentionInterface.register behavior change." + "reason": "A placeholder 'ADD THE DATA' issue and AttentionInterface.register behavior change are unrelated." }, { - "left": "issue:44466", - "right": "issue:45276", + "left": "issue:45081", + "right": "issue:45907", "accept": false, - "reason": "Inconsistent lm_head serialization and resize_token_embeddings not updating Gemma4 embeddings are different model-state bugs." + "reason": "Mistral regex patch crash and list_repo_files missing revision forwarding are separate repository-loading bugs." }, { - "left": "issue:45081", - "right": "issue:45907", + "left": "issue:44466", + "right": "issue:45276", "accept": false, - "reason": "Mistral tokenizer regex crash and list_repo_files revision forwarding are unrelated." + "reason": "Inconsistent lm_head serialization and resize_token_embeddings not propagating are different weight-handling problems." }, { "left": "issue:44016", "right": "issue:44291", "accept": false, - "reason": "Notebook syntax error and init_empty_weights argument incompatibility do not share a code path." + "reason": "Notebook syntax error and init_empty_weights argument mismatch are unrelated." }, { "left": "issue:43441", "right": "issue:45834", "accept": false, - "reason": "FlashAttention failure for Ministral-3 and OCR index errors in Kosmos2.5 are unrelated." + "reason": "FlashAttention failure in Ministral and a long-OCR index error in Kosmos2.5 are different model-specific bugs." }, { "left": "issue:42994", "right": "issue:43317", "accept": false, - "reason": "Quantized model saving failure and dequantized-model loading with device_map=auto are adjacent topics but not the same concrete bug." + "reason": "Quantized saving failure and dequantized loading with device_map=auto are distinct quantization/offload issues." }, { "left": "issue:43653", "right": "issue:45593", "accept": false, - "reason": "BigBirdTokenizer special-token registration and D-FINE auxiliary-loss behavior are unrelated." + "reason": "Tokenizer special-token registration and D-FINE auxiliary-loss behavior are unrelated." }, { "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Missing processor.save_pretrained files and Gemma4 zero3 from_pretrained are different bugs." + "reason": "processor.save_pretrained missing files and gemma-4 zero3 from_pretrained are different save/load issues." }, { "left": "issue:44623", "right": "issue:44749", "accept": false, - "reason": "Processor save/pretrained file issues and a Chinese performance regression are unrelated." + "reason": "Processor save-file omission and a Chinese report about filtering slowdown are not the same bug." }, { "left": "issue:45600", "right": "issue:45712", "accept": false, - "reason": "Removed auto-mapping configs breaking CI and leftover dummy classes failing repo checks are separate maintenance issues." + "reason": "Both are cleanup/CI-related, but they target different stale-reference problems in different files." }, { "left": "issue:33290", "right": "issue:45656", "accept": false, - "reason": "OOM with Adafactor under DeepSpeed and optimizer.step being called twice are both DeepSpeed-related, but not the same underlying bug." + "reason": "OOM with Adafactor under deepspeed and optimizer.step being called twice are not clearly the same underlying bug." }, { "left": "issue:44361", "right": "issue:45593", "accept": false, - "reason": "MLukeTokenizer task AttributeError and D-FINE auxiliary-loss behavior are unrelated." + "reason": "MLukeTokenizer AttributeError and D-FINE auxiliary-loss behavior are unrelated." }, { "left": "issue:44261", "right": "issue:44263", "accept": false, - "reason": "Missing rms_norm_eps in MLA q_a_layernorm and torch.split return values in GlmMoeDsaIndexer are different issues." + "reason": "A q_a_layernorm precision/config issue and torch.split return-value handling are different bugs." }, { "left": "issue:44671", "right": "issue:44779", "accept": false, - "reason": "CamemBERT masked-LM regression and Deepseek tokenizer regression are both v5-related but affect different components and fixes." + "reason": "CamemBERT MLM prediction regression and Deepseek tokenizer output regression are different components and fixes." }, { "left": "issue:44016", "right": "issue:45823", "accept": false, - "reason": "Notebook syntax error and Gemma4 device mismatch have no common bug surface." + "reason": "Notebook syntax error and Gemma4 PLE device mismatch are unrelated." }, { "left": "issue:43723", "right": "issue:44162", "accept": false, - "reason": "AutoTokenizer loading issue and ESM2 breakage are unrelated model/tokenizer problems." + "reason": "AutoTokenizer v5 loading issues and ESM2 breakage are different model/loading problems." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "TensorFlow backend packaging and Python version support are separate dependency/release issues." }, { "left": "issue:43901", "right": "issue:43994", "accept": false, - "reason": "Docs mismatch for return_all_scores and nonsensical SigLIP2 outputs are not the same defect." + "reason": "Documentation mismatch for return_all_scores and incorrect SigLIP2 predictions are not the same bug." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "FlashAttention failure and compiled flex_attention failure are different attention-kernel problems, not one duplicate." } ] }, @@ -11453,13 +11475,11 @@ "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", @@ -11484,11 +11504,13 @@ "issue:44496", "issue:44521", "issue:44589", - "issue:44683", "issue:44734", + "issue:44779", + "issue:44857", "issue:44871", "issue:45003", "issue:45103", + "issue:45292", "issue:45362", "issue:45397", "issue:45542", @@ -11500,8 +11522,6 @@ "issue:46424" ], "soft_pairs": [ - "issue:42175|issue:43976", - "issue:43441|issue:44683", "issue:43673|issue:43761", "issue:43262|issue:44521", "issue:44297|issue:45397", @@ -11516,8 +11536,8 @@ "issue:45103|issue:46399", "issue:45561|issue:45563", "issue:44246|issue:44496", - "issue:44336|issue:44373", "issue:44295|issue:45397", + "issue:44336|issue:44373", "issue:44242|issue:45593", "issue:43526|issue:45003", "issue:42491|issue:43299", @@ -11525,220 +11545,229 @@ "issue:41950|issue:44220", "issue:44246|issue:44871", "issue:42886|issue:46132", - "issue:44336|issue:44485", "issue:36246|issue:42757", "issue:36246|issue:44351", + "issue:44336|issue:44485", "issue:45003|issue:45561", - "issue:44734|issue:46424" + "issue:44734|issue:46424", + "issue:44779|issue:45292", + "issue:44336|issue:46424", + "issue:44242|issue:44857" ], - "prepared_review_unit_hash": "a60f010e3ace6e511d50b3b4aacedffe47f0f2d2f0de95d72a56fc8171b2bc6e", + "prepared_review_unit_hash": "319e88a167032d5f793c72a1a35bc5f1334825790f0d3fa814c04ed3db60d7d8", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36597, - "estimated_input_tokens": 9150, - "estimated_eval_tokens": 18556 + "soft_pair_count": 31, + "serialized_chars": 36900, + "estimated_input_tokens": 9225, + "estimated_eval_tokens": 18706 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of Transformers bug reports around imports, model loading, cache behavior, attention kernels, and docs; there is no strong duplicate core. The closest anchor is the Qwen3VL MoE regression report, but the soft pairs mostly describe distinct failures and should not be merged.", - "confidence": 0.38, - "canonical_issue_reason": "issue:43299 is the best anchor because it is a concrete, high-signal v5 regression report with a specific model family and failure mode, and it sits near the largest concentration of model-loading-related issues in the set.", + "summary": "This set is heterogeneous: most items are unrelated bug reports spanning imports, model loading, cache handling, pipelines, docs, and testing. None of the soft pairs look like true duplicates with the same underlying bug/change, so I rejected them all.", + "confidence": 0.46, + "canonical_issue_reason": "issue:44521 is the most concrete and actionable standalone bug report here: it is open, has multiple inbound references, and describes a specific multimodal apply_chat_template failure with clear behavior to fix. There is no strong duplicate cluster, so it is the best practical canonical issue for this mixed set.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43299 is the strongest standalone issue here: it is clearly described, reproducible in a specific version, and actionable without being as narrow as the one-off import/config/doc complaints.", + "best_issue_reason": "issue:44521 is the strongest representative issue overall because it is current, well-scoped, and tied to a core user-facing API with observable broken behavior.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:42175", - "right": "issue:43976", - "accept": false, - "reason": "Different problems: missing TensorFlow backend from an install extra vs a Python version compatibility regression." - }, - { - "left": "issue:43441", - "right": "issue:44683", - "accept": false, - "reason": "Both mention attention kernels, but they affect different models and different code paths (FlashAttention vs compiled flex_attention)." - }, { "left": "issue:43673", "right": "issue:43761", "accept": false, - "reason": "Unrelated regressions: chunked prefill cache handling vs CLIP vision hidden state forwarding." + "reason": "Different subsystems and failure modes: generation cache during chunked prefill vs CLIP hidden_states output." }, { "left": "issue:43262", "right": "issue:44521", "accept": false, - "reason": "Both involve chat templates, but one is audio sampling-rate defaulting and the other is multimodal assistant mask generation." + "reason": "Both mention apply_chat_template, but one is an audio sampling-rate default bug and the other is a multimodal assistant-mask bug." }, { "left": "issue:44297", "right": "issue:45397", "accept": false, - "reason": "Tokenizer save metadata mismatch is unrelated to the gemma-4 zero3 from_pretrained loading failure." + "reason": "Tokenizer save/load metadata mismatch is unrelated to gemma-4 Zero3 from_pretrained loading." }, { "left": "issue:43824", "right": "issue:43976", "accept": false, - "reason": "A Qwen2.5-VL import/export issue is not the same as a Python version support regression." + "reason": "Missing Qwen2.5-VL export and Python 3.9/3.10 compatibility are different problems." }, { "left": "issue:43901", "right": "issue:44246", "accept": false, - "reason": "Docs mismatch for return_all_scores is unrelated to sporadic slow imports." + "reason": "Docs mismatch for return_all_scores is unrelated to intermittent import slowness." }, { "left": "issue:43525", "right": "issue:43881", "accept": false, - "reason": "Different model-specific loading failures: missing pad_token_id on Llama4Config vs GLM-4V-9B load failure." + "reason": "Different models and code paths: Llama4Config pad_token_id vs glm-4v-9b loading." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "One is a missing config attribute on a specific model; the other is a generic trust_remote_code config field omission." + "reason": "Llama4 config field absence and AutoConfig trust_remote_code missing fields are not the same bug." }, { "left": "issue:44589", "right": "issue:45362", "accept": false, - "reason": "Float8 storage type resolution and Qwen3.5 chat crashes are unrelated." + "reason": "Float8 storage lookup failure is unrelated to Qwen3.5 chat crashes." }, { "left": "issue:43784", "right": "issue:45542", "accept": false, - "reason": "Import-time NameError in sentence-transformers integration is unrelated to TensorFlow backend detection." + "reason": "Both are import-time issues, but one is sentence-transformers missing nn and the other is TensorFlow backend handling." }, { "left": "issue:41628", "right": "issue:43976", "accept": false, - "reason": "Missing AutoImageProcessor import and Python version incompatibility are different failure classes." + "reason": "AutoImageProcessor import failure is unrelated to the Python version support issue." }, { "left": "issue:44188", "right": "issue:44336", "accept": false, - "reason": "Attention-kernel divergence is unrelated to ANSI code emission in loading_report." + "reason": "Attention kernel divergence under torch.compile is unrelated to ANSI color output in loading_report." }, { "left": "issue:45103", "right": "issue:46399", "accept": false, - "reason": "Auto-docstring annotation handling and Qwen3.5 kernel registration errors are different subsystems." + "reason": "Auto-docstring annotation handling and Qwen3.5 kernel registration both involve decorators, but not the same underlying bug." }, { "left": "issue:45561", "right": "issue:45563", "accept": false, - "reason": "A xdist file race in testing utils is unrelated to a stale generate warning." + "reason": "Testing-utils xdist race and stale num_return_sequences warning are different concerns." }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Import slowness is not the same as model auto-detection failing for missing model_type." + "reason": "Import latency is unrelated to an unrecognized model/config.json model_type error." }, { - "left": "issue:44336", - "right": "issue:44373", + "left": "issue:44295", + "right": "issue:45397", "accept": false, - "reason": "ANSI escape leakage and a wrong docstring are unrelated documentation/utilities issues." + "reason": "Reading position_ids after buffer registration is unrelated to gemma-4 Zero3 loading." }, { - "left": "issue:44295", - "right": "issue:45397", + "left": "issue:44336", + "right": "issue:44373", "accept": false, - "reason": "Buffer access bug for position_ids is unrelated to gemma-4 zero3 loading." + "reason": "ANSI terminal handling and a wrong position_ids docstring are unrelated." }, { "left": "issue:44242", "right": "issue:45593", "accept": false, - "reason": "These are both about auxiliary losses, but they affect different models and different training branches." + "reason": "Load-balancing loss omission in one model family is unrelated to D-FINE auxiliary loss behavior." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BEiTImageProcessorFast label reduction and sys.modules access in modeling_utils are unrelated." + "reason": "BeitImageProcessorFast reduce_labels bug and sys.modules access safety are different code paths." }, { "left": "issue:42491", "right": "issue:43299", "accept": false, - "reason": "Both involve Qwen MoE models, but one is a LoRA compatibility issue and the other is a base-model loading regression; not the same concrete bug." + "reason": "Both involve Qwen3 MoE, but one is LoRA compatibility across versions and the other is a Qwen3VL-MoE loading regression; not the same concrete fix." }, { "left": "issue:43531", "right": "issue:45362", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE is distinct from a Qwen3.5 chat crash." + "reason": "Sliding_window behavior in Qwen3-MoE is unrelated to Qwen3.5 chat crashes." }, { "left": "issue:41950", "right": "issue:44220", "accept": false, - "reason": "Video-classification pipeline processor selection and fbank feature extraction are separate audio/video pipeline bugs." + "reason": "Video-classification pipeline image-processor lookup and fbank feature extraction are unrelated." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import performance and Gemma-3 eos_token_id inconsistency are unrelated." + "reason": "Import slowness and Gemma-3 eos_token_id inconsistency do not share a concrete bug." }, { "left": "issue:42886", "right": "issue:46132", "accept": false, - "reason": "Offline tokenizer cache loading and AttentionInterface.register behavior are different issues." - }, - { - "left": "issue:44336", - "right": "issue:44485", - "accept": false, - "reason": "Terminal ANSI output leakage is unrelated to GLM-5 RoPE implementation." + "reason": "Offline cache loading and AttentionInterface.register behavior are unrelated." }, { "left": "issue:36246", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different packages and different missing symbols." + "reason": "Both are import errors, but for different symbols and different packages." }, { "left": "issue:36246", "right": "issue:44351", "accept": false, - "reason": "Different missing exports: Qwen2.5-VLImageProcessor vs HybridCache." + "reason": "Both are missing exports from transformers, but they are different symbols with no evidence of a shared root cause." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI loading-report output and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:45003", "right": "issue:45561", "accept": false, - "reason": "Unsafe sys.modules access and a testing-file race have no shared code path." + "reason": "Unsafe sys.modules access is unrelated to the xdist captured_info.txt race." }, { "left": "issue:44734", "right": "issue:46424", "accept": false, - "reason": "KV-cache continuation indexing in serve is different from max_cache_length being ignored outside quantized cache." + "reason": "KV-cache continuation tensor indexing and max_cache_length handling are different cache bugs." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Deepseek tokenizer regression and resize_token_embeddings not updating output_embeddings are unrelated." + }, + { + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "Terminal color output and cache-length configuration have no shared underlying bug." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "One is a missing auxiliary-loss path; the other is an AMP/CUDA crash in LwDetrImageLoss." } ] }, @@ -11769,20 +11798,20 @@ "issue:43475", "issue:43540", "issue:43596", + "issue:43653", "issue:43673", "issue:43901", + "issue:44060", + "issue:44077", "issue:44162", - "issue:44242", "issue:44246", "issue:44295", - "issue:44336", "issue:44466", "issue:44479", + "issue:44483", "issue:44496", "issue:44561", "issue:44749", - "issue:44779", - "issue:44857", "issue:44877", "issue:44987", "issue:44995", @@ -11805,21 +11834,18 @@ "issue:45859", "issue:46123", "issue:46170", + "issue:46202", "issue:46291", "issue:46369", "issue:46399", - "issue:46424", "issue:46431" ], "soft_pairs": [ - "issue:44779|issue:45292", - "issue:44336|issue:46424", - "issue:44242|issue:44857", "issue:44295|issue:44749", "issue:43475|issue:46369", "issue:45859|issue:46399", - "issue:46291|issue:46399", "issue:36246|issue:43596", + "issue:46291|issue:46399", "issue:43901|issue:45003", "issue:43352|issue:45250", "issue:43673|issue:44561", @@ -11830,8 +11856,8 @@ "issue:45292|issue:45584", "issue:45292|issue:45701", "issue:45412|issue:46123", - "issue:44561|issue:46399", "issue:45375|issue:45759", + "issue:44561|issue:46399", "issue:44877|issue:45375", "issue:45685|issue:46431", "issue:44987|issue:45459", @@ -11840,216 +11866,213 @@ "issue:45397|issue:45656", "issue:45137|issue:45600", "issue:44479|issue:46170", - "issue:43475|issue:43540" + "issue:43475|issue:43540", + "issue:43653|issue:44162", + "issue:44483|issue:46202", + "issue:44060|issue:44077" ], - "prepared_review_unit_hash": "b60dc616637b443cc4824c75e95e8cc55208dfd7d58c4554269ab08bb4d51a83", + "prepared_review_unit_hash": "c9623ebb04864d6bb0c7f1059ccd9613c1847212c58c63c9c5fbfab4f0e6e3a3", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36456, - "estimated_input_tokens": 9114, - "estimated_eval_tokens": 18484 + "serialized_chars": 36525, + "estimated_input_tokens": 9132, + "estimated_eval_tokens": 18520 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft pair appears to be the same underlying bug or change. The set spans unrelated model-specific regressions, config issues, docs bugs, and infrastructure problems, so all candidate merges should be rejected.", - "confidence": 0.87, - "canonical_issue_reason": "issue:46399 is the most actionable and specific report in the set: it has a clear crash mode, a narrow code path (`use_kernels=True` / `apply_rotary_pos_emb` registration), and strong discussion activity, making it a good representative issue.", + "summary": "This set is mostly a grab-bag of distinct Transformers bugs/regressions across unrelated subsystems. None of the soft pairs look like the same underlying change or defect, so all soft edges should be rejected.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 44479 is the clearest representative bug report: a concrete v5.3 video-input regression affecting several related Qwen model variants, with a well-defined scope and failure mode.", "canonical_pr_reason": null, - "best_issue_reason": "issue:46399 is the strongest standalone issue here because it is concrete, reproducible, currently active, and likely to get an immediate code fix.", + "best_issue_reason": "Issue 44479 is the strongest standalone issue to anchor on because it is broad, specific, and directly describes a concrete regression in a shared code path.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:44779", - "right": "issue:45292", - "accept": false, - "reason": "Different failures: DeepSeek tokenization regression vs embedding resize propagation bug." - }, - { - "left": "issue:44336", - "right": "issue:46424", - "accept": false, - "reason": "Unrelated subsystems: ANSI output formatting vs cache-length configuration handling." - }, - { - "left": "issue:44242", - "right": "issue:44857", - "accept": false, - "reason": "Different model/training bugs: MoE load-balancing loss vs RT-DETR AMP crash." - }, { "left": "issue:44295", "right": "issue:44749", "accept": false, - "reason": "Buffer access error and post-upgrade performance slowdown are not the same bug." + "reason": "Unrelated problems: position_ids buffer handling vs. a slowdown after upgrading Transformers." }, { "left": "issue:43475", "right": "issue:46369", "accept": false, - "reason": "SAM 3 Video attribute error and SAM3 LiteText memory overflow are distinct issues." + "reason": "Both mention SAM 3, but one is a missing output attribute and the other is a memory overflow in a different model variant." }, { "left": "issue:45859", "right": "issue:46399", "accept": false, - "reason": "Both are Qwen3.5-related but on different code paths: CPU offload vs kernel registration crash." + "reason": "Same general Qwen family, but different failures: CPU offload incompatibility vs. kernel registration crash." }, { - "left": "issue:46291", - "right": "issue:46399", + "left": "issue:36246", + "right": "issue:43596", "accept": false, - "reason": "Import-time dependency/version issue is unrelated to the `apply_rotary_pos_emb` registration crash." + "reason": "Completely different areas: missing Qwen2.5 VL import vs. DeepSpeed/BERT zero3 index error." }, { - "left": "issue:36246", - "right": "issue:43596", + "left": "issue:46291", + "right": "issue:46399", "accept": false, - "reason": "Completely different areas: missing image processor export vs ZeRO-3/BertModel indexing failure." + "reason": "Both touch kernels, but one is an external kernels dependency change and the other is a Qwen3.5 rotary-pos-emb registration bug." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch for a pipeline option is unrelated to unsafe `sys.modules` access." + "reason": "Docs mismatch for TextClassificationPipeline vs. unsafe sys.modules access in modeling_utils are unrelated." }, { "left": "issue:43352", "right": "issue:45250", "accept": false, - "reason": "A specific model lacking Flash Attention 2 support is not the same as a generic Flash Attention 2 issue." + "reason": "Specific Nemotron Flash Attention unsupported error vs. a generic Flash Attention 2.0 issue; not the same bug." }, { "left": "issue:43673", "right": "issue:44561", "accept": false, - "reason": "Cache handling in generation during chunked prefill is unrelated to removal of `is_torch_fx_available`." + "reason": "Chunked prefill cache bug and removal of is_torch_fx_available breakage are different code paths." }, { "left": "issue:44162", "right": "issue:45200", "accept": false, - "reason": "ESM2 breakage and Gemma 4 token-type defaults affect different models and mechanisms." + "reason": "ESM2 breakage is unrelated to Gemma 4 mm_token_type_ids defaulting to zeros." }, { "left": "issue:44496", "right": "issue:45003", "accept": false, - "reason": "Model config loading error is unrelated to `sys.modules` access in `modeling_utils`." + "reason": "Unrecognized model/config loading vs. sys.modules access safety are distinct failures." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MoE weight conversion and empty-deque ZeRO-3 crash are different DeepSpeed problems." + "reason": "Both involve DeepSpeed/ZeRO3, but the reported symptoms and likely root causes differ." }, { "left": "issue:43262", "right": "issue:44246", "accept": false, - "reason": "Audio chat-template sample-rate default bug is unrelated to slow `import transformers` behavior." + "reason": "Audio processor sampling-rate default bug vs. slow import issue; no shared defect." }, { "left": "issue:45292", "right": "issue:45584", "accept": false, - "reason": "Embedding resize propagation is unrelated to Whisper empty-transcription generation failure." + "reason": "Resize-token-embeddings propagation bug is unrelated to Whisper empty-transcription generation failure." }, { "left": "issue:45292", "right": "issue:45701", "accept": false, - "reason": "Resize-embedding bug and tokenizer-version differences are not the same issue." + "reason": "Embedding resize behavior and version-dependent tokenization changes are different issues." }, { "left": "issue:45412", "right": "issue:46123", "accept": false, - "reason": "RT-DETR memory cleanup and mask pipeline batching are unrelated." + "reason": "RT-DETR memory cleanup and batch-finalization logic are unrelated." }, { - "left": "issue:44561", - "right": "issue:46399", + "left": "issue:45375", + "right": "issue:45759", "accept": false, - "reason": "Trust-remote-code breakage from a removed helper is unrelated to Qwen3.5 kernel registration." + "reason": "Both are Qwen multimodal config issues, but one is a missing vision config field and the other is from_config unwrapping behavior." }, { - "left": "issue:45375", - "right": "issue:45759", + "left": "issue:44561", + "right": "issue:46399", "accept": false, - "reason": "Missing config field under strict validation and `from_config` text-config unwrapping are different composite-config bugs." + "reason": "Breaking trust_remote_code from a removed helper is not the same as a Qwen3.5 kernel registration crash." }, { "left": "issue:44877", "right": "issue:45375", "accept": false, - "reason": "Different model families and failure modes: Granite config strictness vs Qwen3.5 MoE vision config fields." + "reason": "Strict granite_speech config loading and Qwen config field dropping are unrelated config problems." }, { "left": "issue:45685", "right": "issue:46431", "accept": false, - "reason": "MPS histogram kernel limitation and missing chat template are unrelated." + "reason": "MOE histogram MPS dtype error is unrelated to missing chat template support in a processor." }, { "left": "issue:44987", "right": "issue:45459", "accept": false, - "reason": "Loading a model on v5.1.0 and protobuf error masking are separate import/tokenizer issues." + "reason": "Loading failure for a specific model/package and protobuf-decoding error handling are different concerns." }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Device-dependent tied-weight serialization is unrelated to `resize_token_embeddings` not updating decoder embeddings." + "reason": "Both involve weights/embeddings, but one is tied-weight serialization and the other is resize_token_embeddings propagation." }, { "left": "issue:44995", "right": "issue:45200", "accept": false, - "reason": "Second-forward MoE cache bug and Gemma 4 token-type default bug are unrelated." + "reason": "GlmMoeDsa stale cache crash is unrelated to Gemma 4 multimodal token defaults." }, { "left": "issue:45397", "right": "issue:45656", "accept": false, - "reason": "Gemma-4 ZeRO-3 loading and double optimizer stepping are different DeepSpeed concerns." + "reason": "Both mention DeepSpeed, but one is a gemma-4 zero3 loading bug and the other is a double optimizer-step issue." }, { "left": "issue:45137", "right": "issue:45600", "accept": false, - "reason": "Empty-deque ZeRO-3 crash is unrelated to stale auto-mappings references breaking CI." + "reason": "ZeRO3 deque pop error and auto_mappings CI breakage are unrelated." }, { "left": "issue:44479", "right": "issue:46170", "accept": false, - "reason": "Qwen video regression and Windows/ROCm distributed import failure are unrelated." + "reason": "Qwen video-input regression vs. Windows ROCm distributed import/type error are unrelated." }, { - "left": "issue:45250", - "right": "issue:43352", + "left": "issue:43475", + "right": "issue:43540", "accept": false, - "reason": "Same broad feature area, but one is a generic Flash Attention 2 topic and the other is a model-specific unsupported-path error; not the same bug." + "reason": "Different Qwen/SAM model families and different video-processing failures." }, { - "left": "issue:43475", - "right": "issue:43540", + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer special-token registration bug vs. ESM2 breakage are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "Chat completions request validation and Gemma roundtrip tokenization failure are different bugs." + }, + { + "left": "issue:44060", + "right": "issue:44077", "accept": false, - "reason": "Both involve Qwen/SAM video, but the concrete failures are different: missing attribute vs video-input value error." + "reason": "Qwen3-Next tied-weights warning and patchtsmixer post_init allowance are unrelated." } ] }, @@ -12075,6 +12098,8 @@ "nodes": [ "issue:28282", "issue:30990", + "issue:41628", + "issue:42371", "issue:42886", "issue:42994", "issue:43065", @@ -12084,26 +12109,24 @@ "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:45092", "issue:45229", "issue:45230", "issue:45356", @@ -12122,9 +12145,6 @@ "issue:46421" ], "soft_pairs": [ - "issue:43653|issue:44162", - "issue:44483|issue:46202", - "issue:44060|issue:44077", "issue:42886|issue:45229", "issue:44995|issue:45588", "issue:44368|issue:44977", @@ -12136,12 +12156,12 @@ "issue:43931|issue:44368", "issue:45584|issue:46170", "issue:44188|issue:46421", - "issue:44297|issue:44623", "issue:43065|issue:46123", + "issue:44297|issue:44623", "issue:44361|issue:44749", "issue:44623|issue:45656", - "issue:44749|issue:45656", "issue:44869|issue:45356", + "issue:44749|issue:45656", "issue:43646|issue:44162", "issue:43701|issue:44749", "issue:43232|issue:43673", @@ -12151,216 +12171,212 @@ "issue:44038|issue:44623", "issue:43475|issue:43643", "issue:43475|issue:45381", - "issue:28282|issue:30990" + "issue:28282|issue:30990", + "issue:41628|issue:44933", + "issue:42371|issue:45092" ], - "prepared_review_unit_hash": "56199e0943d23896c14187971d87dc70b86d641cd6484c9ec7cef5148a28beae", + "prepared_review_unit_hash": "57216a432441ce494149f742da12c41f497938cafcdb1f2bfa4764ee8b58bcdf", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 35968, - "estimated_input_tokens": 8992, - "estimated_eval_tokens": 18240 + "soft_pair_count": 29, + "serialized_chars": 35575, + "estimated_input_tokens": 8894, + "estimated_eval_tokens": 18044 }, "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 tokenizer, generation, loading, config, and performance issues. I don't see a true duplicate pair or a single clear canonical item to anchor the cluster.", - "confidence": 0.94, - "canonical_issue_reason": null, + "summary": "This cluster looks like a broad false-positive similarity bucket rather than a duplicate set: it spans unrelated issues in loading, tokenization, generation, config, cache, model-specific regressions, and environment/import failures. No soft edge should be accepted.", + "confidence": 0.9, + "canonical_issue_reason": "Issue:43646 is the broadest v5-related regression in the set and the closest thing to a hub, but the surrounding issues are mostly unrelated, so it should not be treated as a true duplicate anchor for the cluster.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "If one issue must represent the cluster, 43646 is the most general and central; however, it only covers a narrow slice of the items and is not a good duplicate target for the rest.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43653", - "right": "issue:44162", - "accept": false, - "reason": "BigBirdTokenizer special-token registration bug vs ESM2 model breakage; different components and failure modes." - }, - { - "left": "issue:44483", - "right": "issue:46202", - "accept": false, - "reason": "Chat completions API rejection vs Gemma pipeline roundtrip failure; unrelated behavior." - }, - { - "left": "issue:44060", - "right": "issue:44077", - "accept": false, - "reason": "Tied-weights warning bug vs patchtsmixer post_init policy change; not the same underlying problem." - }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline tokenizer cache load issue vs multi-GPU CUDA OOM; completely different symptoms." + "reason": "Offline tokenizer-cache loading failure and multi-GPU CUDA OOM are different bugs with different code paths." }, { "left": "issue:44995", "right": "issue:45588", "accept": false, - "reason": "Stale indexer cache on second forward vs flash-attention crash on s_aux=None; different code paths." + "reason": "Stale indexer-cache crash and flash-attention s_aux=None AttributeError are unrelated failures." }, { "left": "issue:44368", "right": "issue:44977", "accept": false, - "reason": "Tie-word-embeddings warning vs flash-attention generation bug; not a duplicate." + "reason": "A warning about tie_word_embeddings during LoRA fine-tuning is not the same as a flash-attention generation bug." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder vs model-type mismatch warning; same family, but different bugs." + "reason": "Dummy Conv2d in Sam3PixelDecoder and sam3_video/sam3_tracker config mismatch are distinct SAM 3 issues." }, { "left": "issue:44368", "right": "issue:45563", "accept": false, - "reason": "Embedding-tie warning vs stale num_return_sequences warning; different warnings and fixes." + "reason": "These are different warnings in different subsystems; one is config advice, the other is a paged generation warning bug." }, { "left": "issue:44188", "right": "issue:46202", "accept": false, - "reason": "torch.compile attention divergence vs Gemma pipeline roundtrip issue; unrelated." + "reason": "torch.compile attention-kernel divergence is unrelated to a Gemma pipeline roundtrip tokenization failure." }, { "left": "issue:44877", "right": "issue:46170", "accept": false, - "reason": "Strict config loading for granite_speech vs Windows ROCm torch.distributed import failure; different areas." + "reason": "Granite speech config strictness and Windows/ROCm torch.distributed import problems are separate environment/config issues." }, { "left": "issue:45987", "right": "issue:46082", "accept": false, - "reason": "StaticCache return-type contract bug vs LlamaConfig head_dim validation; distinct configuration/runtime issues." + "reason": "StaticCache returning a tensor and LlamaConfig rejecting head_dim are different API/validation problems." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch vs embedding-tying warning; not the same bug." + "reason": "Qwen3-VL weight-shape mismatch is not the same bug as a LoRA tie_word_embeddings warning." }, { "left": "issue:45584", "right": "issue:46170", "accept": false, - "reason": "Whisper empty-transcription generation failure vs Windows ROCm import/type error; unrelated." + "reason": "Whisper empty-transcription generation failure is unrelated to the Windows/ROCm torch.distributed import issue." }, { "left": "issue:44188", "right": "issue:46421", "accept": false, - "reason": "Compile-time branching divergence vs cache lazy-init recompilation; both mention compile/cache, but not the same concrete defect." + "reason": "Diverging attention kernels under torch.compile and cache lazy-init recompilation are different cache/compilation regressions." }, { - "left": "issue:44297", - "right": "issue:44623", + "left": "issue:43065", + "right": "issue:46123", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch vs processor.save_pretrained missing files; both serialization-related, but different artifacts and failures." + "reason": "A dummy Conv2d in Sam3PixelDecoder and MaskGenerationPipeline dropping the final partial batch are unrelated." }, { - "left": "issue:43065", - "right": "issue:46123", + "left": "issue:44297", + "right": "issue:44623", "accept": false, - "reason": "SAM3 pixel-decoder placeholder layer vs MaskGenerationPipeline partial-batch handling; unrelated." + "reason": "Tokenizer save_pretrained metadata mismatch and processor.save_pretrained missing files are similar area-wise but not the same concrete failure." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer AttributeError vs data-filtering slowdown regression; different problem class." + "reason": "MLukeTokenizer AttributeError and a performance regression after upgrading Transformers are different issues." }, { "left": "issue:44623", "right": "issue:45656", "accept": false, - "reason": "Missing processor save files vs deepspeed optimizer step called twice; unrelated." + "reason": "Missing files from processor.save_pretrained and a double optimizer.step under deepspeed are unrelated." }, { - "left": "issue:44749", - "right": "issue:45656", + "left": "issue:44869", + "right": "issue:45356", "accept": false, - "reason": "Performance regression after upgrade vs double optimizer-step bug; no shared root cause." + "reason": "Whisper timestamp decode crash on trailing replacement character and Kimi tokenizer codec handling are different model/tokenizer bugs." }, { - "left": "issue:44869", - "right": "issue:45356", + "left": "issue:44749", + "right": "issue:45656", "accept": false, - "reason": "Whisper timestamp decode crash on trailing replacement char vs Kimi tokenizer codec regression; different underlying bugs." + "reason": "Tokenizer slowdown after upgrade and deepspeed calling optimizer.step twice do not share a root cause." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Transformers 5 custom model initialization breakage vs ESM2 broken; different model-loading failures." + "reason": "Generic custom model initialization breakage and ESM2 being broken are separate model-specific regressions." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "resume_from_checkpoint key mismatch vs throughput regression; unrelated." + "reason": "resume_from_checkpoint key mismatch is not the same as a tokenizer performance regression." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "Generation kwargs sync_gpus bug vs missing cache during chunked_prefill; both generation-related, but not the same path." + "reason": "sync_gpus generation-kwargs update and chunked-prefill cache-missing are both generation bugs, but not the same concrete defect." }, { "left": "issue:44871", "right": "issue:46132", "accept": false, - "reason": "EOS token config inconsistency vs AttentionInterface.register behavior change; different issues." + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior change are unrelated API issues." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "Generic bug report placeholder vs gemma-4 zero3 from_pretrained error; no evidence they are the same issue." + "reason": "A generic bug report and a Gemma-4 ZeRO-3 from_pretrained failure are not enough evidence to be duplicates." }, { "left": "issue:42994", "right": "issue:43482", "accept": false, - "reason": "Quantized model saving failure vs Qwen2.5-GGUF loading failure; opposite directions and different bugs." + "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are opposite operations and different bugs." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe v5 regression vs processor.save_pretrained missing files; unrelated." + "reason": "Qwen3-VL-Moe v5 bug and processor.save_pretrained missing files are unrelated model/subsystem issues." }, { "left": "issue:43475", "right": "issue:43643", "accept": false, - "reason": "Missing fpn_position_embeddings on SAM3 vision output vs trust_remote_code config field loss; both mention missing fields, but different objects and causes." + "reason": "Sam3VisionEncoderOutput missing attribute and trust_remote_code AutoConfig missing fields are different remote-code failures." }, { "left": "issue:43475", "right": "issue:45381", "accept": false, - "reason": "SAM3 vision encoder output attribute error vs Qwen2.5-VL video position-id bug; different models and paths." + "reason": "Sam3 vision encoder attribute loss and Qwen2.5-VL video vision_position_ids are unrelated model-specific bugs." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch not found ImportError and SentenceTransformers loading hang have different root causes and symptoms." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "AutoImageProcessor import failure and image_utils nonexistent import are both import issues, but not the same missing symbol/module." }, { - "left": "issue:28282", - "right": "issue:30990", + "left": "issue:42371", + "right": "issue:45092", "accept": false, - "reason": "PyTorch import error vs Sentence Transformers hanging on load; not the same underlying bug." + "reason": "TF32 API warning and old InternVL2 remote-code/meta-init incompatibility are unrelated." } ] }, @@ -12384,10 +12400,11 @@ "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-10000-451", "nodes": [ + "issue:30990", + "issue:35141", "issue:36010", "issue:41628", "issue:42175", - "issue:42371", "issue:42673", "issue:42831", "issue:42898", @@ -12411,12 +12428,12 @@ "issue:44351", "issue:44496", "issue:44568", + "issue:44623", "issue:44792", "issue:44933", "issue:44945", "issue:44964", "issue:45081", - "issue:45092", "issue:45103", "issue:45137", "issue:45310", @@ -12433,8 +12450,6 @@ "issue:46399" ], "soft_pairs": [ - "issue:41628|issue:44933", - "issue:42371|issue:45092", "issue:45375|issue:45478", "issue:45406|issue:46202", "issue:42994|issue:43525", @@ -12462,232 +12477,234 @@ "issue:42175|issue:44933", "issue:43377|issue:44242", "issue:42898|issue:44568", + "issue:43976|issue:44933", "issue:43278|issue:45137", - "issue:43976|issue:44933" + "issue:30990|issue:35141", + "issue:44623|issue:45310" ], - "prepared_review_unit_hash": "d4b9adbcdad78462ce3347b11c53ad9e971cb6b0fb46beeaa7c789346ab89123", + "prepared_review_unit_hash": "ba06df8f9d4d8f145c409346184b0a8c6eb34f2a6358a2c5ab22608a1f03bdfa", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 36365, - "estimated_input_tokens": 9092, - "estimated_eval_tokens": 18440 + "serialized_chars": 36912, + "estimated_input_tokens": 9228, + "estimated_eval_tokens": 18712 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous. The strongest likely duplicate signal is the Qwen3ForCausalLM VRAM leak pair (42673/43704), which shares the same title and symptom. Separately, 45375/45478 also have identical titles and should be treated as a likely duplicate pair, but the packet still doesn\u2019t justify collapsing the whole cluster into one issue.", - "confidence": 0.84, - "canonical_issue_reason": "If one issue must anchor the cluster, 42673 is the best representative because it has the clearest likely duplicate counterpart (43704) and the most concrete bug description. Even so, the surrounding issues are too mixed to treat the cluster as one unified duplicate set.", + "summary": "This cluster is mostly a loose mix of unrelated Transformers issues. There are a couple of exact duplicate reports, but most soft pairs are only superficially similar and should stay separate.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 45310 is the best canonical anchor because it is a concrete, reproducible regression report and has an exact duplicate counterpart (45478) with the same title and symptom.", "canonical_pr_reason": null, - "best_issue_reason": "42673 is the strongest issue anchor in this packet because the 42673/43704 pair is the most compelling likely duplicate match. It should be used as the representative issue, not as proof that the entire cluster is a single duplicate group.", + "best_issue_reason": "Issue 45310 is the strongest representative issue in this set: specific scope, clear failure mode, and an exact duplicate exist nearby, making it the most suitable cluster anchor.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:41628", - "right": "issue:44933", - "accept": false, - "reason": "Both are import failures, but they involve different missing symbols and different code paths, so they do not look like the same bug." - }, - { - "left": "issue:42371", - "right": "issue:45092", - "accept": false, - "reason": "TF32 settings behavior and InternVL2 meta-initialization compatibility are unrelated problems." - }, { "left": "issue:45375", "right": "issue:45478", "accept": false, - "reason": "The titles are identical, but this pair is not in the soft-edge list and the packet does not provide enough extra evidence here to merge them as the same underlying bug." + "reason": "Same general model family, but different bugs: a missing config field vs a from_pretrained load regression." }, { "left": "issue:45406", "right": "issue:46202", "accept": false, - "reason": "Both mention Gemma, but the failure modes differ: tokenizer state/_tokenizer missing versus pipeline roundtrip/tokenization behavior." + "reason": "Both involve Gemma4, but one is a processor attribute error and the other is a pipeline/tokenization roundtrip failure." }, { "left": "issue:42994", "right": "issue:43525", "accept": false, - "reason": "Quantized model saving failure and a missing config attribute are different defects." + "reason": "Quantized model saving failure and Llama4Config pad_token_id absence are unrelated failure modes." }, { "left": "issue:43723", "right": "issue:46326", "accept": false, - "reason": "Both are tokenizer-related, but they target different mechanisms and distinct failures." + "reason": "Tokenizer loading in v5 and mistral_common continue_final_message are different code paths and symptoms." }, { "left": "issue:43784", "right": "issue:46399", "accept": false, - "reason": "Sentence-transformers import NameError and Qwen3.5 rotary-kernel registration are unrelated." + "reason": "Import-time NameError in sentence-transformers is unrelated to the Qwen3.5 kernel registration crash." }, { "left": "issue:45310", "right": "issue:45375", "accept": false, - "reason": "Both involve Qwen3.5/Moe, but one is a generic from_pretrained error and the other is a specific config-field omission; not enough to call them the same bug." + "reason": "Both mention Qwen3.5, but one is a from_pretrained regression and the other is a config-field omission." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Tokenizer regex patching and auto_docstring annotation handling are clearly different code paths." + "reason": "Both are AttributeErrors, but they affect different components and root causes." }, { "left": "issue:43873", "right": "issue:44945", "accept": false, - "reason": "Quantization offloading behavior and pipeline-parallel output corruption are different problems." + "reason": "Quantization/offloading behavior and pipeline-parallel output correctness are separate bugs." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "These look like the same Qwen3ForCausalLM VRAM leak report: identical title, same symptom, and no contrary evidence in the packet. This is a likely duplicate pair." + "reason": "These are the same reported bug with the same title: Qwen3ForCausalLM VRAM leak across multiple dataloader threads." }, { "left": "issue:44964", "right": "issue:45137", "accept": false, - "reason": "Phi-4 multimodal loading failure and a DeepSpeed ZeRO3 deque error are unrelated." + "reason": "A model-load failure for Phi-4 multimodal is unrelated to a DeepSpeed ZeRO3 deque underflow." }, { "left": "issue:41628", "right": "issue:44351", "accept": false, - "reason": "Both are import errors, but they concern different missing exports and likely different regressions." + "reason": "Both are import errors, but for different missing symbols and different regressions." }, { "left": "issue:43232", "right": "issue:44188", "accept": false, - "reason": "Generation kwargs handling after sync_gpus and attention-kernel divergence under torch.compile are unrelated." + "reason": "Generation kwargs handling after sync_gpus is unrelated to torch.compile attention-kernel branching." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "Both touch FP8/Moe/Qwen3.5 broadly, but one is save_pretrained behavior and the other is a from_pretrained load error." + "reason": "Both involve MoE/FP8-adjacent areas, but one is save_pretrained on FP8 and the other is a from_pretrained load regression." }, { "left": "issue:44945", "right": "issue:45753", "accept": false, - "reason": "Pipeline-parallel output issues and a Qwen3_5 image infinite-loop bug are different problems." + "reason": "Incorrect pipeline-parallel output and an infinite-loop image bug are different issues." }, { "left": "issue:42994", "right": "issue:43824", "accept": false, - "reason": "Quantized saving failure and a missing Qwen2_5_VL import are unrelated." + "reason": "Quantized model saving and missing Qwen2_5_VLForConditionalGeneration import are unrelated." }, { "left": "issue:46077", "right": "issue:46123", "accept": false, - "reason": "Gemma4 bidirectional attention masking and MaskGenerationPipeline batching are different defects." + "reason": "Bidirectional attention masking and final partial-batch result dropping are different pipeline/model bugs." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Both are Qwen/Moe-related, but they affect different models and different failure modes." + "reason": "Both mention Qwen and versioning, but the model variants and failure modes do not match." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Both are import failures, but for different symbols and likely different export regressions." + "reason": "Different import regressions: GenerationMixin versus AutoImageProcessor." }, { "left": "issue:43232", "right": "issue:44792", "accept": false, - "reason": "Generation kwarg logic and a janus model test failure are not the same issue." + "reason": "Generation sync kwargs bug is unrelated to the janus test failure." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "Both mention FP8, but one is an accuracy issue and the other is a save_pretrained issue." + "reason": "Both touch FP8, but one is an accuracy issue and the other is a save_pretrained failure." }, { "left": "issue:42175", "right": "issue:45712", "accept": false, - "reason": "Packaging backend omission and leftover dummy-class cleanup are unrelated." + "reason": "TensorFlow backend packaging and leftover dummy-class cleanup are unrelated." }, { "left": "issue:44496", "right": "issue:45561", "accept": false, - "reason": "Model config recognition failure and a pytest-xdist file race are unrelated." + "reason": "Model config recognition failure and a pytest-xdist file race are not the same bug." }, { "left": "issue:43526", "right": "issue:45561", "accept": false, - "reason": "BEiT label reduction and an xdist race condition are unrelated." + "reason": "BeitImageProcessorFast label handling and xdist captured_info.txt races are unrelated." }, { "left": "issue:43504", "right": "issue:45561", "accept": false, - "reason": "Legacy-field model loading failure and an xdist race are unrelated." + "reason": "Legacy preset loading failure and xdist file contention are different problems." }, { "left": "issue:42175", "right": "issue:44933", "accept": false, - "reason": "Packaging backend omission and an image_utils import error are unrelated." + "reason": "TensorFlow backend install issue is unrelated to a missing image_utils import." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "MIMI batching/padding-mask behavior and MoE load-balancing loss handling are different bugs." + "reason": "MIMI padding-mask behavior and MoE load-balancing loss gating are separate model bugs." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer-related v5 regressions, but they affect different behaviors: cleanup spaces versus special-token insertion." + "reason": "Tokenizer cleanup behavior in v5 and add_special_tokens BOS/EOS behavior are distinct changes." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility and missing image_utils import are unrelated." }, { "left": "issue:43278", "right": "issue:45137", "accept": false, - "reason": "Embedding dtype drift and a ZeRO3 deque underflow are unrelated." + "reason": "Embedding dtype drift in eval and a ZeRO3 deque pop error are unrelated." }, { - "left": "issue:43976", - "right": "issue:44933", + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence Transformers loading stall and embedding resize reinitialization are different issues." + }, + { + "left": "issue:44623", + "right": "issue:45310", "accept": false, - "reason": "Python-version compatibility and an image_utils import error are unrelated." + "reason": "Processor save_pretrained missing files is unrelated to the Qwen3.5 MoE from_pretrained regression." } ] }, "evaluator_result": { - "accept": false, - "feedback": "The overall cluster-level conclusion is reasonable, but the wording overstates the evidence: the packet only shows a same-title, high-similarity pair for 42673/43704, not a confirmed duplicate. Please soften claims like \"confirmed duplicate pair\" / \"exact duplicate\" to \"likely duplicate\" unless there is explicit linkage in the packet. Also note that at least one other pair (45375/45478) has identical titles and should be acknowledged carefully if you want the summary to be maximally precise." + "accept": true, + "feedback": "Grounded overall. The anchor choice is supported by the packet, the duplicate claims are conservative, and the soft-edge rejections mostly distinguish different failure modes rather than over-merging them." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -12709,7 +12726,6 @@ "issue:27957", "issue:28282", "issue:30333", - "issue:30990", "issue:33290", "issue:34689", "issue:35141", @@ -12755,8 +12771,6 @@ "issue:46431" ], "soft_pairs": [ - "issue:30990|issue:35141", - "issue:44623|issue:45310", "issue:30333|issue:33290", "issue:42222|issue:45200", "issue:43479|issue:44869", @@ -12778,227 +12792,245 @@ "issue:44933|issue:45712", "issue:44297|issue:45656", "issue:27957|issue:30333", - "issue:30333|issue:34689", "issue:44222|issue:44623", + "issue:30333|issue:34689", "issue:44448|issue:46132", "issue:42831|issue:44623", "issue:46209|issue:46431", - "issue:30333|issue:36331", "issue:43479|issue:43644", + "issue:30333|issue:36331", "issue:30333|issue:35141", "issue:22355|issue:28282", "issue:45245|issue:46132", "issue:44297|issue:45310", "issue:43122|issue:46209" ], - "prepared_review_unit_hash": "b27881a7b81e4fbe25f2c45da7873d2558e0eddb4ef98be2d22ada902d502572", + "prepared_review_unit_hash": "b2ee7b8c1c0552c676d649818d6d1fec0b31d9e34a60c001ac0aa83258615c6e", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 35, - "serialized_chars": 38213, - "estimated_input_tokens": 9554, - "estimated_eval_tokens": 19364 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "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 cluster is heterogeneous: most items are unrelated bugs, with only a loose tokenizer/regression theme among a few v5 issues. None of the soft-edge pairs look like the same underlying bug/change, so they should stay separate.", - "confidence": 0.86, - "canonical_issue_reason": "issue:43122 is the closest broad representative of the tokenizer-regression subgroup, but the cluster overall is not a true duplicate set.", + "summary": "The set is highly heterogeneous; the soft edges mostly connect superficially similar transformer/tokenizer/FP8/training bugs, but they do not look like the same underlying defect. No soft pair should be merged.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 45812 is the clearest representative bug report in this mixed set: it describes a concrete, high-impact v4\u2192v5 tokenizer regression affecting all Granite models, with a specific observable failure mode.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43122 is the most representative issue for the shared v5 tokenization regression theme, even though it does not cover the unrelated items in the cluster.", + "best_issue_reason": "Issue 45812 is the best issue for triage because it is precise, broadly impactful, and regression-oriented; the other issues are either narrower, orthogonal, or only loosely related by subsystem.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:30990", - "right": "issue:35141", - "accept": false, - "reason": "Sentence-transformers loading hang and token-embedding reinit are different bugs in different code paths." - }, - { - "left": "issue:44623", - "right": "issue:45310", - "accept": false, - "reason": "Missing processor save files and Qwen3.5 MoE from_pretrained failure are distinct serialization/loading problems." - }, { "left": "issue:30333", "right": "issue:33290", "accept": false, - "reason": "Failed-MLflow reporting and deepspeed Adafactor OOM are unrelated training issues." + "reason": "MLflow reporting on failed training jobs is unrelated to Deepspeed OOMs in Adafactor." }, { "left": "issue:42222", "right": "issue:45200", "accept": false, - "reason": "VitPose breakage and Gemma 4 mm_token_type_ids defaults affect different model families and mechanisms." + "reason": "VitPose breakage and Gemma 4 multimodal token defaults are different model-specific bugs." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Phi4MultimodalConfig default-init and Whisper timestamp decode crash are unrelated." + "reason": "Phi4MultimodalConfig default-init bug is unrelated to Whisper timestamp decode crashing." }, { "left": "issue:42898", "right": "issue:45812", "accept": false, - "reason": "Both are tokenizer regressions, but one is clean_up_tokenization_spaces behavior and the other is Granite token-id generation; not the same fix." + "reason": "Both are v5 tokenizer regressions, but one is about clean_up_tokenization_spaces behavior and the other about Granite token IDs; different code paths and symptoms." }, { "left": "issue:43479", "right": "issue:44079", "accept": false, - "reason": "Config defaulting for multimodal inputs and ModelOutput key assignment are separate code paths." + "reason": "Config default initialization and ModelOutput key assignment are distinct internal issues." }, { "left": "issue:43784", "right": "issue:43824", "accept": false, - "reason": "Missing nn import in sentence-transformers and missing Qwen2_5_VL export are different import errors." + "reason": "Sentence-transformers import NameError and Qwen2_5_VL missing export are unrelated import failures." }, { "left": "issue:43723", "right": "issue:44568", "accept": false, - "reason": "Tokenizer loading failure and add_special_tokens BOS/EOS behavior are related to tokenization but are different bugs." + "reason": "Both involve tokenizers, but one is about AutoTokenizer loading in v5 and the other is a model-specific add_special_tokens regression." }, { "left": "issue:42898", "right": "issue:44779", "accept": false, - "reason": "These are both tokenizer regressions, but they describe different model-specific output issues and different underlying causes." + "reason": "General tokenization behavior change and DeepSeek tokenizer regression are different bugs." }, { "left": "issue:44016", "right": "issue:46231", "accept": false, - "reason": "Notebook syntax error and documentation parameter-name mismatch are unrelated." + "reason": "Notebook syntax error and a documentation parameter-name mismatch are not the same issue." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "KV-cache continuation crash in serve and auto_docstring annotation handling are unrelated failures." + "reason": "Serve KV-cache continuation crash and auto_docstring annotation handling are unrelated code paths." }, { "left": "issue:42898", "right": "issue:45356", "accept": false, - "reason": "v5 clean-up-tokenization behavior changes and Kimi-K2.5 codec/warning regressions are different bugs." + "reason": "Both are tokenizer regressions, but they concern different models and different failure modes." }, { "left": "issue:43874", "right": "issue:45092", "accept": false, - "reason": "GLM46V image-patch method missing and InternVL2 meta-init incompatibility are different multimodal model issues." + "reason": "Missing image-patch method in GLM46V processor and InternVL2 meta-init incompatibility are separate multimodal bugs." }, { "left": "issue:44295", "right": "issue:45656", "accept": false, - "reason": "Buffer read error and double optimizer-step behavior under deepspeed are unrelated." + "reason": "Buffer registration/readback bug and Deepspeed double optimizer step are unrelated." }, { "left": "issue:44075", "right": "issue:46231", "accept": false, - "reason": "Unused SGD args and a docs parameter-name typo are not the same issue." + "reason": "SGD arguments being ignored is a runtime optimizer bug, not a documentation typo." }, { "left": "issue:42898", "right": "issue:44448", "accept": false, - "reason": "Tokenizer clean-up regression and Pegasus v4/v5 output difference are distinct behavior changes." + "reason": "Different tokenizer behavior regression vs Pegasus model output regression; not the same defect." }, { "left": "issue:44060", "right": "issue:45125", "accept": false, - "reason": "Incorrect tied-weights warning in Qwen3-Next and missing _tp_plan in Qwen3.5 MoE are different model-internal bugs." + "reason": "Incorrect tied-weights warning and missing tensor-parallel plan are different Qwen3.5-MoE issues." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "FP8 save_pretrained behavior and tokenizer_config class mismatch are separate serialization problems." + "reason": "FP8 MoE save_pretrained and tokenizer save_pretrained metadata mismatch are distinct serialization problems." }, { "left": "issue:45405", "right": "issue:45712", "accept": false, - "reason": "PEFT version bump and leftover dummy classes are unrelated maintenance issues." + "reason": "PEFT version pinning and leftover dummy classes are unrelated maintenance issues." }, { "left": "issue:44933", "right": "issue:45712", "accept": false, - "reason": "Missing image_utils import and leftover dummy classes are unrelated repository hygiene bugs." + "reason": "Missing image_utils import and dummy class cleanup are different repository hygiene problems." }, { "left": "issue:44297", "right": "issue:45656", "accept": false, - "reason": "Tokenizer metadata saving and deepspeed optimizer stepping are unrelated." + "reason": "Tokenizer config save metadata and Deepspeed double-step behavior are unrelated." }, { "left": "issue:27957", "right": "issue:30333", "accept": false, - "reason": "Flash Attention 2 support for XLMRoberta and MLflow reporting on failed jobs are unrelated." + "reason": "Flash Attention 2 on XLM-Roberta and MLflow failure reporting are unrelated." }, { - "left": "issue:30333", - "right": "issue:34689", + "left": "issue:44222", + "right": "issue:44623", "accept": false, - "reason": "MLflow failure reporting and Llama 3.2 Vision loading regression are unrelated." + "reason": "FP8 MoE save_pretrained and processor.save_pretrained missing files are both save-related, but not the same object or code path." }, { - "left": "issue:44222", - "right": "issue:44623", + "left": "issue:30333", + "right": "issue:34689", "accept": false, - "reason": "Both mention save_pretrained, but one is FP8 MoE model saving and the other is processor file export; not the same underlying bug." + "reason": "Failed-job MLflow handling is unrelated to Llama 3.2 Vision model loading regression." }, { "left": "issue:44448", "right": "issue:46132", "accept": false, - "reason": "Pegasus output regression and AttentionInterface.register behavior change are different code paths." + "reason": "Pegasus output regression and AttentionInterface.register behavior change are separate issues." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FP8 accuracy and missing processor save files are unrelated." + "reason": "FineGrainedFP8 accuracy issue and processor save_pretrained file omission do not share the same underlying bug." }, { "left": "issue:46209", "right": "issue:46431", "accept": false, - "reason": "FP8/B200 inference failure and missing chat-template processor support are unrelated." + "reason": "FP8 inference on B200 and missing chat template in a processor are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Incorrect multimodal config defaults and junk-filled non-persistent buffers are distinct initialization bugs." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow failure reporting and CustomTrainer compute_loss signature mismatch are unrelated trainer issues." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "Failed-job reporting and embedding reinitialization after resize are different bugs." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx module and missing PyTorch dependency are separate import errors." }, { "left": "issue:45245", "right": "issue:46132", "accept": false, - "reason": "Category-limit error and attention-registration behavior change are unrelated." + "reason": "Category-count runtime limit and AttentionInterface.register behavior change are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch and Qwen3.5-MoE from_pretrained error are different regressions." }, { "left": "issue:43122", "right": "issue:46209", "accept": false, - "reason": "Tokenizer version regression and FP8 inference on B200 are unrelated." + "reason": "Tokenizer behavior change across versions and FP8 inference on B200 are unrelated." } ] }, @@ -13029,8 +13061,8 @@ "issue:42831", "issue:42907", "issue:43232", + "issue:43408", "issue:43475", - "issue:43479", "issue:43644", "issue:43723", "issue:43824", @@ -13051,7 +13083,6 @@ "issue:44734", "issue:44829", "issue:44871", - "issue:44913", "issue:44977", "issue:44991", "issue:45071", @@ -13067,6 +13098,7 @@ "issue:46082", "issue:46305", "issue:46328", + "issue:46402", "issue:46421", "issue:46424" ], @@ -13075,8 +13107,8 @@ "issue:41628|issue:42757", "issue:43723|issue:44977", "issue:45071|issue:46082", - "issue:44206|issue:45103", "issue:45381|issue:46305", + "issue:44206|issue:45103", "issue:43232|issue:46424", "issue:44190|issue:46082", "issue:44222|issue:44484", @@ -13087,8 +13119,8 @@ "issue:44373|issue:44485", "issue:43723|issue:44871", "issue:43723|issue:44991", - "issue:44279|issue:46328", "issue:44483|issue:45406", + "issue:44279|issue:46328", "issue:44263|issue:44360", "issue:42907|issue:46050", "issue:30990|issue:38175", @@ -13099,217 +13131,220 @@ "issue:42831|issue:45310", "issue:41628|issue:43824", "issue:42831|issue:44038", - "issue:43479|issue:44913" + "issue:43408|issue:46402" ], - "prepared_review_unit_hash": "945bcf993fa8127c1d875a7b4dbf88347a163f4b2d09de044c2f8202b205fef3", + "prepared_review_unit_hash": "7a9ffaf4477106e910be7c97e7162d1c4885ac431d5b1e56ce08e84bbb8ad724", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35549, - "estimated_input_tokens": 8888, - "estimated_eval_tokens": 18032 + "serialized_chars": 35554, + "estimated_input_tokens": 8889, + "estimated_eval_tokens": 18034 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a mix of unrelated Transformers v5/model-specific bugs. The soft links are largely superficial API-area similarity rather than the same underlying defect, so I would not merge any of the candidate pairs.", - "confidence": 0.89, - "canonical_issue_reason": "issue:43723 is the broadest, most generic user-facing regression in the set (tokenizer loading in v5) and is the closest thing here to an umbrella issue, though the cluster is too heterogeneous for a strong canonical match.", + "summary": "This cluster is mostly heterogeneous and does not represent one coherent duplicate set. The only plausible duplicate pair is the v5 tokenizer-loading regression (43723/44991); all other soft edges look like distinct bugs in different subsystems or models.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 43723 is the best umbrella candidate because it states the broad v5 tokenizer-loading failure, which is the closest match to the only plausible duplicate pair in the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43723 is the best available anchor because it is a clear, high-level loading regression and more general than the more model-specific reports, but it still does not represent the whole cluster well.", + "best_issue_reason": "43723 is the most representative issue here: it is broad enough to subsume a specific tokenizer-loading instance like 44991, while the rest of the cluster is unrelated.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43475", "right": "issue:46305", "accept": false, - "reason": "Different subsystems and failure modes: SAM 3 vision encoder attribute access vs Moonshine streaming mask off-by-one." + "reason": "Different models and failure modes: SAM3 video encoder attribute error vs MoonshineStreaming sliding-window off-by-one." }, { "left": "issue:41628", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but for different symbols and likely different root causes; not the same bug." + "reason": "Both are import errors, but they involve different missing symbols in different packages; not the same bug." }, { "left": "issue:43723", "right": "issue:44977", "accept": false, - "reason": "Tokenizer loading regression vs flash-attention generation behavior; not the same code-path problem." + "reason": "Tokenizer-loading regression vs flash-attention generation quality issue; unrelated code paths." }, { "left": "issue:45071", "right": "issue:46082", "accept": false, - "reason": "Both concern config validation, but one is type checking and the other is Llama head_dim divisibility; different bugs." + "reason": "Both concern config validation, but they are different checks on different configs and are not the same underlying defect." }, { - "left": "issue:44206", - "right": "issue:45103", + "left": "issue:45381", + "right": "issue:46305", "accept": false, - "reason": "Feature extractor argument incompatibility vs auto-docstring crash under future annotations; unrelated." + "reason": "Both are off-by-one style reports, but they affect different models/components and do not look like one shared bug." }, { - "left": "issue:45381", - "right": "issue:46305", + "left": "issue:44206", + "right": "issue:45103", "accept": false, - "reason": "Different models and logic: Qwen2.5-VL vision_position_ids vs Moonshine streaming window masking." + "reason": "Feature extractor argument crash and auto-docstring AttributeError are unrelated." }, { "left": "issue:43232", "right": "issue:46424", "accept": false, - "reason": "Both are cache/generation related, but one is sync_gpus model kwargs handling and the other is cache length config behavior; not mergeable as one fix." + "reason": "Both are generation/cache related, but one is a sync_gpus kwargs bug and the other is cache-length handling for quantized cache." }, { "left": "issue:44190", "right": "issue:46082", "accept": false, - "reason": "Local dataset loading script issue vs LlamaConfig head_dim validation; unrelated." + "reason": "Local dataset loading in an example script vs LlamaConfig head_dim validation; unrelated." }, { "left": "issue:44222", "right": "issue:44484", "accept": false, - "reason": "Quantized save_pretrained bug vs max_shard_size question; same method name, different underlying issue." + "reason": "Saving FP8 MoE weights and max_shard_size behavior are different save_pretrained concerns." }, { "left": "issue:44734", "right": "issue:46421", "accept": false, - "reason": "Both involve cache, but one is KV continuation tensor indexing and the other is lazy init causing recompilation; different bugs." + "reason": "Serve KV-cache indexing crash vs cache lazy-init recompilation; different concrete problems." }, { "left": "issue:44295", "right": "issue:44623", "accept": false, - "reason": "Registering position_ids as a buffer vs processor.save_pretrained missing files; unrelated." + "reason": "Position_ids buffer access bug and missing processor save files are unrelated." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "DSA indexer torch.split behavior vs remote-code checkpoint/meta-init incompatibility; not the same defect." + "reason": "DSA indexer torch.split behavior vs InternVL2 meta-init incompatibility; not the same bug." }, { "left": "issue:43232", "right": "issue:45834", "accept": false, - "reason": "Generation kwargs/cache handling vs Kosmos2.5 OCR long-input index error; different paths." + "reason": "Generation kwargs after sync_gpus and Kosmos2.5 OCR index error are unrelated." }, { "left": "issue:44373", "right": "issue:44485", "accept": false, - "reason": "Docstring text issue vs GLM-5 RoPE implementation bug; not the same change." + "reason": "A wrong docstring is not the same as a RoPE implementation bug." }, { "left": "issue:43723", "right": "issue:44871", "accept": false, - "reason": "Tokenizer loading failure vs eos_token_id config mismatch; related area but different concrete bug." + "reason": "Tokenizer-loading regression vs eos_token_id configuration inconsistency; different symptoms and fixes." }, { "left": "issue:43723", "right": "issue:44991", - "accept": false, - "reason": "Both are tokenizer-load failures in v5, but on different models with no evidence of the same root cause." + "accept": true, + "reason": "Both describe v5 tokenizer-loading failures; 44991 looks like a model-specific instance of the broader regression in 43723." }, { - "left": "issue:44279", - "right": "issue:46328", + "left": "issue:44483", + "right": "issue:45406", "accept": false, - "reason": "Generic dependency complaint vs SequenceFeatureExtractor.pad performance issue; unrelated." + "reason": "Chat/completions request rejection and Gemma4Processor missing _tokenizer are unrelated server/processor failures." }, { - "left": "issue:44483", - "right": "issue:45406", + "left": "issue:44279", + "right": "issue:46328", "accept": false, - "reason": "Chat/completions request rejection vs Gemma4Processor missing _tokenizer; different serving failures." + "reason": "Dependency trouble vs a pad() performance issue; no shared underlying bug." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "Same DSA indexer area, but one is torch.split return values and the other is missing ReLU; distinct bugs." + "reason": "Same broad DSA area, but one is about split outputs and the other about a missing ReLU; distinct fixes." }, { "left": "issue:42907", "right": "issue:46050", "accept": false, - "reason": "Both involve quantization/save/load, but dequantized Ministrals/Devstrals saving is not the same as sinq save/load breakage." + "reason": "Both involve save/load of quantized models, but different quantization methods and failure points." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence Transformers loading hang vs zero probabilities in SigLIP2; unrelated symptoms and code paths." + "reason": "Completely different models and symptoms: loading hang vs zero probabilities." }, { "left": "issue:45092", "right": "issue:45357", "accept": false, - "reason": "Both touch Qwen/InternVL multimodal save/load regressions, but one is meta-init compatibility and the other is incorrect visual encoder keys on save." + "reason": "Remote-code meta-init incompatibility vs incorrect visual encoder keys on save_pretrained; different bugs." }, { "left": "issue:43644", "right": "issue:46305", "accept": false, - "reason": "Non-persistent buffer initialization vs Moonshine streaming mask off-by-one; unrelated." + "reason": "Non-persistent buffer junk in v5.0 vs a streaming window off-by-one; unrelated." }, { "left": "issue:44829", "right": "issue:46002", "accept": false, - "reason": "Training degradation with flash_attention_3 vs NaNs after checkout/init; similar symptoms only, not the same bug." + "reason": "Degenerate training with flash_attention_3 and NaNs on checkout/init are different classification-head issues." }, { "left": "issue:44261", "right": "issue:45092", "accept": false, - "reason": "Layernorm epsilon precision issue vs remote-code checkpoint/meta-init incompatibility; unrelated." + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and InternVL2 meta-init incompatibility are unrelated." }, { "left": "issue:42831", "right": "issue:45310", "accept": false, - "reason": "FineGrainedFP8 accuracy issue vs Qwen3.5 MoE from_pretrained error; different bug classes." + "reason": "FineGrainedFP8 accuracy issue vs Qwen3.5 MoE from_pretrained error; different code paths." }, { "left": "issue:41628", "right": "issue:43824", "accept": false, - "reason": "Missing import symbols, but different APIs/models and no shared concrete root cause." + "reason": "Both are import errors, but they concern different classes/symbols and do not look like one shared defect." }, { "left": "issue:42831", "right": "issue:44038", "accept": false, - "reason": "FP8 accuracy regression vs Qwen3-VL-Moe bug; not the same issue." + "reason": "Accuracy regression in FP8 vs Qwen3-VL-Moe bug; not the same problem." }, { - "left": "issue:43479", - "right": "issue:44913", + "left": "issue:43408", + "right": "issue:46402", "accept": false, - "reason": "Config default-init bug vs rotary_pct not persisting on reload; both config-related but different code paths and models." + "reason": "SAM3 model-type warning and SAM3 text-encoder weight loading are different issues." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet and is conservative. The only accepted soft edge, 43723/44991, is a credible same-bug/tokenizer-loading match, and the other rejections are appropriately cautious rather than overstated." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -13343,13 +13378,13 @@ "issue:43122", "issue:43232", "issue:43408", + "issue:43479", "issue:43550", "issue:43576", "issue:43646", "issue:43756", "issue:43761", "issue:43784", - "issue:43937", "issue:44060", "issue:44117", "issue:44188", @@ -13363,6 +13398,7 @@ "issue:44521", "issue:44568", "issue:44704", + "issue:44913", "issue:44945", "issue:45278", "issue:45593", @@ -13372,11 +13408,10 @@ "issue:46132", "issue:46305", "issue:46326", - "issue:46369", - "issue:46402" + "issue:46369" ], "soft_pairs": [ - "issue:43408|issue:46402", + "issue:43479|issue:44913", "issue:43761|issue:46082", "issue:44263|issue:44373", "issue:39401|issue:44568", @@ -13384,237 +13419,230 @@ "issue:42757|issue:45278", "issue:27957|issue:30990", "issue:42886|issue:43066", - "issue:22355|issue:28218", "issue:42886|issue:42994", - "issue:42831|issue:45753", + "issue:22355|issue:28218", "issue:44368|issue:46305", + "issue:42831|issue:45753", "issue:43408|issue:46369", "issue:43232|issue:45593", "issue:43576|issue:43784", "issue:44295|issue:44297", - "issue:43646|issue:46002", "issue:43550|issue:46132", + "issue:43646|issue:46002", "issue:44117|issue:44704", "issue:42371|issue:44263", "issue:43756|issue:44060", "issue:30990|issue:33357", "issue:44188|issue:44373", "issue:42907|issue:44945", - "issue:44261|issue:44373", "issue:43122|issue:44521", "issue:39401|issue:44488", - "issue:43122|issue:46326", "issue:41628|issue:45278", - "issue:42907|issue:43937" + "issue:43122|issue:46326", + "issue:44261|issue:44373" ], - "prepared_review_unit_hash": "b31397f29f897a00617ec1ec54b90d8305c0e83b22ab4f077f74f03d061b9f7e", + "prepared_review_unit_hash": "cb70cbd768e6da7372972bab4c5885772eac0cef6e22d7dad304a88aa5de77d8", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 36232, - "estimated_input_tokens": 9058, - "estimated_eval_tokens": 18372 + "soft_pair_count": 29, + "serialized_chars": 35977, + "estimated_input_tokens": 8995, + "estimated_eval_tokens": 18246 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These issues are largely heterogeneous and do not form a real duplicate cluster: they span unrelated tokenizer, import, generation, quantization, multimodal, and model-specific regressions. No soft pair looks like the same underlying bug/change with enough confidence to merge.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 45278 is the broadest umbrella-style report in the set (\u201cmany import errors after update from 4.57.0 to 5.5.0\u201d), so it is the best single issue to anchor the cluster, even though the surrounding items are not true duplicates.", + "summary": "The items are mostly unrelated standalone issues across tokenization, config, import, quantization, generation, and model-specific regressions. None of the proposed soft pairs look like the same underlying bug or change, so there is no strong duplicate cluster here.", + "confidence": 0.96, + "canonical_issue_reason": "issue:42886 is the clearest concrete regression to anchor triage because it has a specific failure mode, version, and repro; however, the overall set is heterogeneous rather than one duplicate family.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45278 is the most general and likely to attract related follow-up reports; it is the strongest representative issue among otherwise unrelated bugs.", + "best_issue_reason": "issue:42886 is the best single representative issue because it is precise and reproducible (offline tokenizer cache loading regression), but it should not be treated as a canonical duplicate for the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43408", - "right": "issue:46402", + "left": "issue:43479", + "right": "issue:44913", "accept": false, - "reason": "Both mention sam3, but one is a model-type warning and the other is missing text-encoder weights; different symptoms and code paths." + "reason": "Different bugs: multimodal config defaulting vs GPTNeoX config reload losing rotary_pct." }, { "left": "issue:43761", "right": "issue:46082", "accept": false, - "reason": "Unrelated failures: CLIPVisionModel hidden_states regression vs LlamaConfig head_dim validation." + "reason": "Different code paths: CLIPVision hidden_states regression vs LlamaConfig head_dim validation." }, { "left": "issue:44263", "right": "issue:44373", "accept": false, - "reason": "A torch.split indexing bug and a docstring issue are not the same bug." + "reason": "Runtime indexing/splitting bug vs a docstring wording issue; not the same bug." }, { "left": "issue:39401", "right": "issue:44568", "accept": false, - "reason": "Both are tokenizer bugs, but they affect different models and different behaviors (offset mapping vs BOS/EOS insertion)." + "reason": "Tokenizer offset_mapping regression vs BOS/EOS special-token insertion failure." }, { "left": "issue:22355", "right": "issue:27957", "accept": false, - "reason": "Missing transformers.onnx import is unrelated to XLM-Roberta Flash Attention support." + "reason": "Missing transformers.onnx module import and XLMRoberta FlashAttention2 support are unrelated." }, { "left": "issue:42757", "right": "issue:45278", "accept": false, - "reason": "Specific missing huggingface_hub symbol vs broad import errors after a version upgrade; not the same concrete defect." + "reason": "A specific missing import from huggingface_hub is not the same as broad v5 import breakages." }, { "left": "issue:27957", "right": "issue:30990", "accept": false, - "reason": "Flash Attention on XLMRoberta is unrelated to sentence-transformers loading hangs." + "reason": "FlashAttention2 model support issue vs sentence-transformers loading hang; different symptoms and causes." }, { "left": "issue:42886", "right": "issue:43066", "accept": false, - "reason": "Offline cache loading failure and wrong tokenizer decoder type are different tokenizer regressions." - }, - { - "left": "issue:22355", - "right": "issue:28218", - "accept": false, - "reason": "ONNX import failure is unrelated to added-token spacing behavior." + "reason": "Offline cache loading failure is unrelated to tokenizer decoder type selection in v5." }, { "left": "issue:42886", "right": "issue:42994", "accept": false, - "reason": "Tokenizer offline/cache loading is unrelated to quantized model saving failure." + "reason": "Tokenizer cache/offline bug vs quantized model saving failure." }, { - "left": "issue:42831", - "right": "issue:45753", + "left": "issue:22355", + "right": "issue:28218", "accept": false, - "reason": "FineGrainedFP8 accuracy issues and a Qwen3_5 image-generation loop are different model behaviors." + "reason": "ONNX import failure and extra-space tokenization behavior are unrelated." }, { "left": "issue:44368", "right": "issue:46305", "accept": false, - "reason": "A tied-embeddings warning and a sliding-window off-by-one bug are not the same underlying issue." + "reason": "A tied-weights warning is not the same as a streaming mask off-by-one bug." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue vs multimodal infinite-loop behavior; different failures." }, { "left": "issue:43408", "right": "issue:46369", "accept": false, - "reason": "Both are SAM3-related, but one is a config/model-class warning and the other is a memory overflow in a different path." + "reason": "Model-type warning on instantiation vs memory overflow in a different model path." }, { "left": "issue:43232", "right": "issue:45593", "accept": false, - "reason": "Generation kwargs handling after sync_gpus is unrelated to D-FINE auxiliary-loss behavior." + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss behavior are unrelated." }, { "left": "issue:43576", "right": "issue:43784", "accept": false, - "reason": "Broken transformers env command and a sentence-transformers import NameError are different failures." + "reason": "Broken CLI env command vs import-time NameError in sentence-transformers integration." }, { "left": "issue:44295", "right": "issue:44297", "accept": false, - "reason": "Reading registered position_ids and tokenizer_class mismatch are distinct bugs." + "reason": "position_ids buffer read error vs tokenizer_class metadata mismatch on save_pretrained." }, { - "left": "issue:43646", - "right": "issue:46002", + "left": "issue:43550", + "right": "issue:46132", "accept": false, - "reason": "Custom model initialization regression and NaNs in classification heads are different symptoms and code paths." + "reason": "torch.compile SDPA failure in Bamba and AttentionInterface.register behavior change do not share a concrete root cause." }, { - "left": "issue:43550", - "right": "issue:46132", + "left": "issue:43646", + "right": "issue:46002", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to AttentionInterface.register behavior changes." + "reason": "Custom model initialization breakage in v5 vs NaNs in classification heads are different regressions." }, { "left": "issue:44117", "right": "issue:44704", "accept": false, - "reason": "Tokenizer mapping None handling and AutoProcessor kwargs forwarding are separate loader bugs." + "reason": "TOKENIZER_MAPPING_NAMES returning None and AutoProcessor kwargs forwarding are separate loader bugs." }, { "left": "issue:42371", "right": "issue:44263", "accept": false, - "reason": "TF32 API warning and GlmMoeDsaIndexer split behavior are unrelated." + "reason": "TF32 settings warning/behavior is unrelated to torch.split output handling." }, { "left": "issue:43756", "right": "issue:44060", "accept": false, - "reason": "Smollm3 RoPE-layer mismatch and Qwen3-Next tied-weights warning are different model configuration issues." + "reason": "Smollm3 RoPE-layer mismatch vs Qwen3-Next tied-weights warning; not the same issue." }, { "left": "issue:30990", "right": "issue:33357", "accept": false, - "reason": "A loading hang and a MacOS bus error are not the same bug." + "reason": "Loading hang and MacOS bus error are different failure modes with no shared concrete bug." }, { "left": "issue:44188", "right": "issue:44373", "accept": false, - "reason": "torch.compile attention-kernel divergence and a docstring typo are unrelated." + "reason": "Torch.compile attention-kernel divergence is unrelated to a docstring typo." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Dequantized model save failure and pipeline-parallel output correctness are different code paths." - }, - { - "left": "issue:44261", - "right": "issue:44373", - "accept": false, - "reason": "MLA layernorm precision/config issue is unrelated to a docstring bug." + "reason": "Quantized/dequantized save failure vs pipeline parallelism output correctness are different problems." }, { "left": "issue:43122", "right": "issue:44521", "accept": false, - "reason": "Tokenization differences across versions and multimodal assistant_mask generation are different behaviors." + "reason": "Tokenization drift across versions vs multimodal assistant_masks all-zero bug are unrelated." }, { "left": "issue:39401", "right": "issue:44488", "accept": false, - "reason": "Wrong offset_mapping and failure to load a specific model are unrelated tokenizer/model-loading issues." + "reason": "Qwen3 offset-mapping bug and failure to load cjvt/sleng-bert are not the same underlying issue." }, { - "left": "issue:43122", - "right": "issue:46326", + "left": "issue:41628", + "right": "issue:45278", "accept": false, - "reason": "Version-to-version tokenization changes are unrelated to continue_final_message behavior with mistral_common." + "reason": "Single missing AutoImageProcessor import vs broad post-upgrade import errors are too different to merge." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:43761", + "right": "issue:44373", "accept": false, - "reason": "Specific AutoImageProcessor import failure is broader than, but not the same as, the many import errors regression." + "reason": "A CLIPVision forward regression is not a docstring issue." }, { - "left": "issue:42907", - "right": "issue:43937", + "left": "issue:44261", + "right": "issue:44373", "accept": false, - "reason": "Quantized model saving and invalid GenerationConfig for GLM-5 are unrelated." + "reason": "MLA q_a_layernorm precision/config bug is unrelated to a docstring mistake." } ] }, @@ -13687,6 +13715,7 @@ "issue:45753" ], "soft_pairs": [ + "issue:42907|issue:43937", "issue:43784|issue:45600", "issue:43976|issue:44945", "issue:44704|issue:45290", @@ -13704,9 +13733,9 @@ "issue:42907|issue:43452", "issue:45216|issue:45478", "issue:43824|issue:44351", - "issue:36296|issue:43116", "issue:22355|issue:29159", "issue:42757|issue:44351", + "issue:36296|issue:43116", "issue:43278|issue:44945", "issue:42175|issue:44938", "issue:43493|issue:45753", @@ -13717,172 +13746,214 @@ "issue:38617|issue:45278", "issue:43937|issue:44485" ], - "prepared_review_unit_hash": "a549e622308d7fcd217ec6f15797ebf2f1249bef197f4b0bb13929c3f733a20c", + "prepared_review_unit_hash": "ad0c029c9c7a4ca920560f3f8e3fdf856e028e09a99df3a46c074b02b5160d96", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 35463, - "estimated_input_tokens": 8866, - "estimated_eval_tokens": 17988 + "soft_pair_count": 30, + "serialized_chars": 35759, + "estimated_input_tokens": 8940, + "estimated_eval_tokens": 18136 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a grab bag of unrelated Transformers issues. The soft pairs share broad themes like tokenization, imports, or Qwen/SAM model families, but they do not describe the same underlying bug closely enough to deduplicate. I would reject every soft edge.", - "confidence": 0.91, - "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report in the set, covering multiple import failures after an upgrade and serving as the best anchor for the small import-regression subset.", - "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the most general, self-contained symptom among the issues and the one most likely to represent a broader duplicate cluster, even though the overall set is still heterogeneous.", - "best_pr_reason": null, + "summary": "All soft pairs are false positives; these items span unrelated bugs across imports, tokenizers, model loading, chat templates, and training, not one duplicate cluster.", + "confidence": 0.97, + "canonical_issue_reason": "No canonical issue: the cluster is too heterogeneous to map to one underlying bug.", + "canonical_pr_reason": "No PRs present in the cluster.", + "best_issue_reason": "No issue is a strong global representative because the accepted-overlap signal is absent and every soft edge is a mismatch.", + "best_pr_reason": "No PRs present.", "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Dequantized save failure for Ministrals/Devstrals vs GLM-5 generation config validation; different models and code paths." + }, { "left": "issue:43784", "right": "issue:45600", "accept": false, - "reason": "Both are import-related, but they fail on different symbols in different code paths; not the same bug." + "reason": "sentence-transformers import NameError vs Sam3LiteText auto-mappings CI breakage; unrelated import/mapping bugs." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python-version support regression and pipeline-parallel output corruption are unrelated failures." + "reason": "Python version compatibility issue vs incorrect output under pipeline parallelism; different failure modes." }, { "left": "issue:44704", "right": "issue:45290", "accept": false, - "reason": "Both involve processor/chat-template plumbing, but the APIs and failure modes are different." + "reason": "AutoProcessor kwarg forwarding to cached_file vs chat-template crash on tool-call assistant messages; different subsystems." }, { "left": "issue:42907", "right": "issue:44964", "accept": false, - "reason": "One is a save-pretrained issue for dequantized models; the other is a model load failure for a different model family." + "reason": "Save/dequantization bug vs inability to load Phi-4 multimodal model; different model families and paths." }, { "left": "issue:39401", "right": "issue:44521", "accept": false, - "reason": "Both are Qwen-related tokenization/multimodal issues, but offset_mapping and assistant_mask bugs are distinct." + "reason": "Qwen3 offset_mapping regression vs multimodal assistant_mask generation; tokenization offsets are not the same bug as chat-template masking." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "A Qwen3-VL-Moe regression and a tokenizer metadata mismatch are not the same underlying problem." + "reason": "Qwen3-VL-Moe from_pretrained error vs tokenizer_config tokenizer_class mismatch on save; different load/save issues." }, { "left": "issue:43116", "right": "issue:45200", "accept": false, - "reason": "Different subsystems and symptoms: example script multi-label output vs Gemma 4 token-type handling." + "reason": "Example script multi-label classification returning empty results vs Gemma 4 mm_token_type_ids defaulting; unrelated example vs model-input bug." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "A discussion about DSA indexing is unrelated to GLM-5 RoPE implementation behavior." + "reason": "DSA indexer/ReLU discussion vs GLM-5 RoPE implementation; entirely different technical areas." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "Config propagation for Qwen3.5 is unrelated to flash-attention training degeneration." + "reason": "Qwen3.5 num_labels propagation vs flash_attention_3 degenerate training; config plumbing and attention-kernel bugs are distinct." }, { "left": "issue:44038", "right": "issue:44295", "accept": false, - "reason": "A model load/runtime issue and a position_ids buffer access bug are different code paths." + "reason": "Qwen3-VL-Moe load failure vs position_ids buffer access error; not the same underlying bug." }, { "left": "issue:43122", "right": "issue:44514", "accept": false, - "reason": "Tokenizer regression across versions and batched apply_chat_template crash are different concrete bugs." + "reason": "Tokenizer regression across versions vs batched apply_chat_template crash with padding=False; different code paths." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "Both are SAM 3 video related, but one is a model-type warning and the other is a missing attribute in an output object." + "reason": "SAM3 model-type warning vs missing fpn_position_embeddings attribute in vision encoder output; related family, but not one bug." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Both mention tokenizer regressions, but they affect different models and different observable behavior." + "reason": "Generic tokenization differences in v5 vs Deepseek-specific incorrect tokenization; too broad to be the same defect." }, { "left": "issue:44038", "right": "issue:44222", "accept": false, - "reason": "A Qwen3-VL-Moe load bug is not the same as FP8 save_pretrained behavior." + "reason": "Qwen3-VL-Moe from_pretrained error vs FP8 save_pretrained issue; load-time and save-time bugs differ." }, { "left": "issue:42907", "right": "issue:43452", "accept": false, - "reason": "Saving dequantized models and loading gguf through AutoTokenizer/AutoModel are distinct failure modes." + "reason": "Dequantized saving bug vs gguf_file loading breakage for AutoTokenizer/AutoModelForCausalLM; different serialization paths." }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Both concern Qwen3.5 and version 5.4+, but one is save_pretrained correctness and the other is from_pretrained loading." + "reason": "Qwen3.5 save_pretrained checkpoint regression vs Qwen3.5 Moe from_pretrained error; same model family, but different save/load failures." }, { "left": "issue:43824", "right": "issue:44351", "accept": false, - "reason": "These are separate missing-export/import errors for unrelated symbols." - }, - { - "left": "issue:36296", - "right": "issue:43116", - "accept": false, - "reason": "Tensor-parallel training and the classification example script are unrelated bugs." + "reason": "Missing Qwen2_5_VLForConditionalGeneration export vs missing HybridCache export; both import errors, but for different symbols and modules." }, { "left": "issue:22355", "right": "issue:29159", "accept": false, - "reason": "Missing module import and slow/fast tokenizer inconsistency are not the same underlying issue." + "reason": "Missing transformers.onnx module vs slow/fast tokenizer inconsistency; import failure and tokenizer behavior bug are unrelated." }, { "left": "issue:42757", "right": "issue:44351", "accept": false, - "reason": "One is a huggingface_hub import error, the other is a transformers export/import error." + "reason": "huggingface_hub missing is_offline_mode vs transformers missing HybridCache; different packages and missing symbols." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor-parallel training bug vs example script multi-label classification issue; distinct workloads and symptoms." }, { "left": "issue:43278", "right": "issue:44945", "accept": false, - "reason": "BF16-to-FP32 dtype drift and pipeline-parallel output corruption are different problems." + "reason": "BF16-to-FP32 embedding dtype change in evaluate vs incorrect pipeline-parallel outputs; different bugs." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Missing tensorflow in extras install vs Python 3.14 load failure; packaging/dependency issue versus interpreter compatibility." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 HF vs JAX discrepancy vs Qwen3_5 infinite loop on a specific image; unrelated model families." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 vs Deepseek tokenizer incorrect results; both tokenizer regressions, but not the same concrete defect." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "onnx import failure vs LayoutLMv3 error-message clarity; unrelated import and UX bugs." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM with multiple dataloader threads vs dtype mismatch CI failure; different mechanisms and symptoms." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder load vs AutoProcessor kwarg forwarding; unrelated model loading vs processor plumbing." }, { "left": "issue:38617", "right": "issue:45278", "accept": false, - "reason": "Both are import-related, but one is a missing symbol and the other is a broad set of import failures after upgrade; not clearly one bug." + "reason": "Specific missing layer_type_validation import vs broad 'many import errors' report; 45278 is an umbrella symptom, not the same bug." }, { "left": "issue:43937", "right": "issue:44485", "accept": false, - "reason": "Invalid GenerationConfig and GLM-5 RoPE implementation are different failure classes." + "reason": "GLM-5 generation config invalid vs GLM-5 RoPE implementation; same model family but different concrete bugs." } ] }, @@ -13961,238 +14032,238 @@ "issue:43116|issue:43493", "issue:42371|issue:43874", "issue:42222|issue:43441", - "issue:44075|issue:44936", - "issue:43976|issue:45446", "issue:44038|issue:45478", + "issue:43976|issue:45446", "issue:45357|issue:45478", + "issue:44075|issue:44936", "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:43937|issue:45563", "issue:44117|issue:45588", "issue:44380|issue:46077", - "issue:43827|issue:45446", "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:45405|issue:45446", "issue:43937|issue:44977", - "issue:44117|issue:45081", + "issue:45405|issue:45446", "issue:44410|issue:45310", - "issue:43122|issue:46298" + "issue:43122|issue:46298", + "issue:44117|issue:45081" ], - "prepared_review_unit_hash": "86323b0f92007f674ff08c686e4ddf703c1c95605cfd4484bcddd54b5b2b92ce", + "prepared_review_unit_hash": "56e71963df02f4a9f7ffd4bc90e52bbc33c45d97a06a153b736d75b5eee26cde", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 47, "item_count": 47, "soft_pair_count": 29, - "serialized_chars": 35697, - "estimated_input_tokens": 8925, - "estimated_eval_tokens": 18106 + "serialized_chars": 35703, + "estimated_input_tokens": 8926, + "estimated_eval_tokens": 18108 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster remains mostly heterogeneous regression reports. The Qwen3.5 Moe `from_pretrained` pair is only topically similar from the packet, not a confirmed duplicate bug, so it should stay conservative.", - "confidence": 0.87, - "canonical_issue_reason": "If a representative issue is needed, `issue:43122` is still the broadest tokenizer-regression report and best matches the tokenizer-related portion of the cluster.", + "summary": "This cluster is mostly heterogeneous Transformers v5 regression reports across unrelated subsystems. The only especially close pair is the two Qwen3.5 MoE `from_pretrained` reports, which look duplicate-like from the titles alone, but the metadata still does not prove a shared root cause. Issue:43122 is a reasonable representative tokenizer-regression issue, but the cluster as a whole is too mixed to treat as one bug.", + "confidence": 0.55, + "canonical_issue_reason": "issue:43122 is a reasonable representative tokenizer-regression report: its title clearly points to a version-dependent tokenization change, and it sits within the tokenizer-related subset of the cluster. I\u2019m not relying on unseen repro/body details.", "canonical_pr_reason": null, - "best_issue_reason": "`issue:43122` is the most general and representative issue here because it describes a version-to-version tokenization regression and is broader than the model-specific bugs.", + "best_issue_reason": "issue:43122 is the best representative issue to keep because it is one of the clearest tokenizer-regression reports in the set and serves as a sensible anchor for that small subgroup, even though the overall cluster remains mixed.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44188", "right": "issue:46328", "accept": false, - "reason": "Different subsystems: attention-kernel branching vs padding performance." + "reason": "Attention-kernel divergence and padding-performance regression are different bugs in different paths." }, { "left": "issue:43611", "right": "issue:46002", "accept": false, - "reason": "Model-loading `base_model_prefix` regression vs NaNs in classification heads." + "reason": "Model-loading `base_model_prefix` failure and NaNs in classification heads are distinct symptoms and likely distinct causes." }, { "left": "issue:44355", "right": "issue:46231", "accept": false, - "reason": "Compiled-file import errors vs a docs parameter-name mismatch." + "reason": "Compiled Python file errors and a docs parameter-name mismatch are unrelated." }, { "left": "issue:43116", "right": "issue:43493", "accept": false, - "reason": "Example-script multi-label output bug vs a SigLIP2 implementation discrepancy." + "reason": "Empty multi-label classification outputs and a SigLIP2 implementation discrepancy are different issues." }, { "left": "issue:42371", "right": "issue:43874", "accept": false, - "reason": "TF32 settings behavior vs missing processor method / AttributeError." + "reason": "TF32 API guidance is unrelated to a missing image-processor method." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "Vitpose breakage and Ministral FlashAttention failure are unrelated model-specific bugs." + "reason": "VitPose model breakage and Ministral-3 FlashAttention failure are different code paths." }, { - "left": "issue:44075", - "right": "issue:44936", + "left": "issue:44038", + "right": "issue:45478", "accept": false, - "reason": "Optimizer argument handling vs trainer evaluate-after-train behavior." + "reason": "Very similar Qwen3.5 MoE `from_pretrained` failures with identical-looking titles, but metadata alone does not prove the same underlying bug." }, { "left": "issue:43976", "right": "issue:45446", "accept": false, - "reason": "Python version compatibility report vs an AuxRequest import version check bug." + "reason": "Python-version compatibility and a PyTorch version-check bug are different problem classes." }, { - "left": "issue:44038", + "left": "issue:45357", "right": "issue:45478", "accept": false, - "reason": "Both are about Qwen3.5 Moe `from_pretrained` failures, but the packet only supports topical similarity, not a confirmed same underlying bug or duplicate report." + "reason": "Both are Qwen3.5-related, but one is a `save_pretrained` serialization issue and the other is a `from_pretrained` load failure." }, { - "left": "issue:45357", - "right": "issue:45478", + "left": "issue:44075", + "right": "issue:44936", "accept": false, - "reason": "Save-pretrained key serialization issue vs a load-time from-pretrained error." + "reason": "Ignored SGD args and `trainer.evaluate()` failing after `trainer.train()` are separate training bugs." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Empty multi-label example output vs an incorrect pipeline error message." + "reason": "Classification example output bug and pipeline translation messaging/docs issue are unrelated." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "Different helper-function behavior bug vs PyTorch version-check bug." + "reason": "Qwen3 omni feature-length helper and flex_attention version checking are different subsystems." }, { "left": "issue:43064", "right": "issue:43673", "accept": false, - "reason": "Optimizer-state issue under FSDP/PEFT vs GenerationMixin cache absence in chunked prefill." + "reason": "FSDP2/PEFT optimizer-state corruption and GenerationMixin cache handling are distinct issues." }, { "left": "issue:33453", "right": "issue:39401", "accept": false, - "reason": "Both tokenizer-related, but one is a loading regression and the other is offset_mapping correctness." + "reason": "Both tokenizer-related, but the loading regression and the Qwen3 offset-mapping bug are not the same concrete failure." }, { - "left": "issue:43937", - "right": "issue:45563", + "left": "issue:39401", + "right": "issue:44779", "accept": false, - "reason": "Invalid generation config vs a stale warning for `num_return_sequences`." + "reason": "Offset-mapping correctness and Deepseek tokenization output changes are different tokenizer bugs." }, { - "left": "issue:39401", - "right": "issue:44779", + "left": "issue:43937", + "right": "issue:45563", "accept": false, - "reason": "Different model-specific tokenizer regressions with different symptoms." + "reason": "GenerationConfig validation and a stale warning about `num_return_sequences` are unrelated." }, { "left": "issue:44117", "right": "issue:45588", "accept": false, - "reason": "Tokenizer mapping assumption bug vs FlashAttention crash on `s_aux=None`." + "reason": "Tokenizer mapping lookup handling and a flash-attention `s_aux=None` crash are not the same bug." }, { "left": "issue:44380", "right": "issue:46077", "accept": false, - "reason": "Attention backend scaling bug vs causal-mask construction bug." + "reason": "Different attention-mask/scaling problems in different models and backends." }, { - "left": "issue:43827", - "right": "issue:45446", + "left": "issue:42673", + "right": "issue:43906", "accept": false, - "reason": "Docs still referencing removed pipeline APIs vs an import/version-check issue." + "reason": "VRAM leakage across dataloader threads and an isolated reproduction issue are only loosely related at best." }, { - "left": "issue:42673", - "right": "issue:43906", + "left": "issue:43827", + "right": "issue:45446", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM threads vs an isolated reproduction of another issue." + "reason": "Docs still referencing removed pipelines and an import/version check bug are different issues." }, { "left": "issue:43450", "right": "issue:43986", "accept": false, - "reason": "Batched video shape bug vs missing torchvision crash path." + "reason": "Batched video-processor shape mismatch and missing `torchvision` crash are distinct failure modes." }, { "left": "issue:43673", "right": "issue:46328", "accept": false, - "reason": "Generation cache bug vs SequenceFeatureExtractor padding inefficiency." + "reason": "Generation cache missing and feature-extractor padding performance are unrelated subsystems." }, { "left": "issue:44466", "right": "issue:46002", "accept": false, - "reason": "Serialization inconsistency vs NaNs in classification heads." + "reason": "Tied-weight serialization inconsistency and NaNs after checkout/init do not appear to share a root cause." }, { "left": "issue:44291", "right": "issue:45071", "accept": false, - "reason": "Unexpected init argument error vs PretrainedConfig type-checking breakage." - }, - { - "left": "issue:45405", - "right": "issue:45446", - "accept": false, - "reason": "Dependency version bump issue vs an unrelated import-version check." + "reason": "Unexpected `_is_hf_initialized` argument handling and `PretrainedConfig` type checking are different bugs." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "Invalid generation config vs FlashAttention generation failure." + "reason": "Invalid generation config and Qwen3.5 flash-attention generation failure are unrelated." }, { - "left": "issue:44117", - "right": "issue:45081", + "left": "issue:45405", + "right": "issue:45446", "accept": false, - "reason": "Tokenizer mapping None bug vs Mistral regex patch crash." + "reason": "A packaging/version pinning issue is not the same as a PyTorch version check bug." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Different Qwen family bug reports with different failing code paths." + "reason": "Different Qwen-family load problems with different symptoms and likely different code paths." }, { "left": "issue:43122", "right": "issue:46298", "accept": false, - "reason": "Tokenizer regression vs a noisy logging message." + "reason": "Tokenizer regression and noisy logging are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping-name handling and the Mistral regex patch crash are different tokenizer bugs." } ] }, "evaluator_result": { "accept": false, - "feedback": "The cluster summary is mostly conservative, but one soft-edge verdict is too strong: `issue:44038` vs `issue:45478` is not grounded as a clear duplicate from the packet. Their titles are not the same, and the evidence only supports topical similarity around Qwen/Moe from_pretrained failures, not a confirmed same-bug duplicate. Please downgrade that pair to a conservative false/uncertain verdict." + "feedback": "The clustering is broadly plausible, but a few claims are too strong for the evidence available in the packet. In particular, calling issue:44038 and issue:45478 the \"same underlying bug\" is not well-grounded from the metadata alone; they are clearly similar, but the packet does not prove duplication. Also, the canonical/representative reasoning for issue:43122 leans on reproduction/body details that are not present in the packet. Please soften those verdicts and keep the summary more conservative, e.g. \"most likely duplicate-like pair\" and \"reasonable representative tokenizer-regression issue.\"" }, "evaluator_used": true, "retried": true, @@ -14264,8 +14335,8 @@ ], "soft_pairs": [ "issue:45341|issue:45561", - "issue:45341|issue:46328", "issue:38617|issue:41628", + "issue:45341|issue:46328", "issue:44829|issue:45216", "issue:42222|issue:46178", "issue:42617|issue:45341", @@ -14293,20 +14364,20 @@ "issue:41628|issue:45071", "issue:44263|issue:44485" ], - "prepared_review_unit_hash": "1b4f50528d2cf6faf2d312f945c885c829cbf562d012dc05e3ed516135ad59c7", + "prepared_review_unit_hash": "52cf101f77fbda400c12258bd6a023b52c547964b1e89a230d4660bfb266307d", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 35948, + "serialized_chars": 35945, "estimated_input_tokens": 8987, "estimated_eval_tokens": 18230 }, @@ -14315,186 +14386,186 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is mostly a loose collection of unrelated Transformers issues; the only near-cluster around import regressions is best represented by the broad version-upgrade import-breakage report. I found no PRs to evaluate.", - "confidence": 0.79, - "canonical_issue_reason": "issue:45278 is the broadest, most umbrella-like report in the set: it describes widespread import errors after upgrading from 4.57.0 to 5.5.0, which best subsumes the narrower import-regression complaints.", - "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best representative issue because it is the only one framed as a general import-regression problem rather than a single symbol/model-specific failure.", - "best_pr_reason": null, + "summary": "This cluster is heterogeneous: it mixes unrelated import errors, tokenizer regressions, backend/config issues, model-specific failures, and performance/behavior bugs. None of the soft pairs looks like the same underlying issue with a plausible single fix, so there is no strong canonical duplicate here.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue: the items are not one bug family and the closest-looking pairs still point to different code paths or symptoms.", + "canonical_pr_reason": "No pull requests were provided in the cluster.", + "best_issue_reason": "No best representative issue: this is not a tight duplicate cluster, so any one issue would misrepresent the rest.", + "best_pr_reason": "No pull requests were provided in the cluster.", "soft_edge_verdicts": [ { "left": "issue:45341", "right": "issue:45561", "accept": false, - "reason": "Both involve testing utilities, but one is a generic bug and the other is a specific xdist race on captured_info.txt; not the same underlying defect." + "reason": "Both mention testing_utils, but one is a vague \"little bug\" and the other is a specific xdist race on captured_info.txt; not the same concrete bug." }, { - "left": "issue:45341", - "right": "issue:46328", + "left": "issue:38617", + "right": "issue:41628", "accept": false, - "reason": "Unrelated areas: testing_utils vs SequenceFeatureExtractor.pad performance." + "reason": "Both are import errors, but they involve different symbols and code paths (`layer_type_validation` vs `AutoImageProcessor`)." }, { - "left": "issue:38617", - "right": "issue:41628", + "left": "issue:45341", + "right": "issue:46328", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols and likely different code paths." + "reason": "One is a testing_utils bug and the other is a SequenceFeatureExtractor.pad performance issue; unrelated functionality." }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "FlashAttention training degeneration and Qwen3.5 save_pretrained checkpoint regression are different bugs." + "reason": "Flash-attention training degeneration and Qwen3.5 save_pretrained checkpoint corruption are different regressions." }, { "left": "issue:42222", "right": "issue:46178", "accept": false, - "reason": "VitPose breakage and tensorflow_text backend validation are unrelated." + "reason": "VitPose breakage and a tensorflow_text backend mapping error are unrelated failures." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "Different reports with no shared failure mode beyond being in the test/docs ecosystem." + "reason": "3d_parallel.py execution failure is unrelated to a generic testing_utils bug." }, { "left": "issue:45362", "right": "issue:46328", "accept": false, - "reason": "Qwen3.5 chat crashes and padding performance are unrelated." + "reason": "Qwen3.5 chat crashes and feature-extractor padding inefficiency are different code paths and symptoms." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "Different model-loading problems: sam3 tracker/model-type warning versus MobileLLM key error." + "reason": "Model-type warning for sam3_tracker vs KeyError loading MobileLLM are unrelated model-loading issues." }, { "left": "issue:42673", "right": "issue:43582", "accept": false, - "reason": "VRAM leak in Qwen3ForCausalLM and Apple Silicon warmup TypeError are distinct problems." + "reason": "VRAM leak in multi-threaded Qwen3.5 inference is not the same as an Apple Silicon TypeError in caching_allocator_warmup." }, { "left": "issue:39401", "right": "issue:43421", "accept": false, - "reason": "Both are tokenizer-related, but offset_mapping is a different bug from runtime post-processor updates." + "reason": "Offset mapping bug and runtime post-processor update bug are both tokenizer-related but distinct underlying problems." }, { "left": "issue:44877", "right": "issue:46231", "accept": false, - "reason": "Strict config loading and a documentation parameter-name mismatch are unrelated." + "reason": "Strict config loading for granite_speech and a docs parameter-name mismatch are unrelated." }, { "left": "issue:42490", "right": "issue:42673", "accept": false, - "reason": "Local-path import warning and VRAM leak are different code paths." + "reason": "Local-path trailing-slash import warning is unrelated to the Qwen3.5 VRAM leak." }, { "left": "issue:42371", "right": "issue:44261", "accept": false, - "reason": "TF32 API guidance and MLA rms_norm_eps precision issue are unrelated." + "reason": "TF32 API warning and MLA rms_norm_eps precision issue do not share the same code path." }, { "left": "issue:44484", "right": "issue:45310", "accept": false, - "reason": "Save_pretrained shard-size question and Qwen3.5 Moe from_pretrained error are not the same bug." + "reason": "A save_pretrained shard-size question and a Qwen3.5 MoE from_pretrained error are not the same bug." }, { "left": "issue:38617", "right": "issue:44220", "accept": false, - "reason": "ImportError for layer_type_validation and fbank feature extraction issue are unrelated." + "reason": "ImportError for configuration_utils and `_torch_extract_fbank_features()` are unrelated issues." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "Invalid GenerationConfig and unrecognized model config are different validation failures." + "reason": "GenerationConfig validation for GLM-5 and an unrecognized model/config.json `model_type` error are different problems." }, { "left": "issue:44933", "right": "issue:45278", "accept": false, - "reason": "These are both import-related, but 44933 is a specific missing import while 45278 is a broad upgrade-wide regression; not enough to call them the same defect." + "reason": "A missing import from image_utils and broad import errors after upgrade are not specific enough to treat as the same bug." }, { "left": "issue:43066", "right": "issue:43122", "accept": false, - "reason": "Tokenizer decoder type regression and changed tokenization output are related by theme but not the same concrete bug." + "reason": "Tokenizer decoder type issue and tokenization output differences are related to tokenizers generally but not the same concrete defect." }, { "left": "issue:42222", "right": "issue:45542", "accept": false, - "reason": "VitPose model breakage and tensorboard/TensorFlow backend mismatch are unrelated." + "reason": "VitPose model breakage and a tensorboard-without-TF backend error are unrelated package/backend issues." }, { "left": "issue:44410", "right": "issue:45478", "accept": false, - "reason": "Missing qwen3next projections and Qwen3.5 Moe from_pretrained error are different model-loading/model-architecture issues." + "reason": "Missing layer projections in qwen3next and Qwen3.5 MoE from_pretrained failure are different model-specific issues." }, { "left": "issue:42371", "right": "issue:42831", "accept": false, - "reason": "TF32 warning behavior and FineGrainedFP8 accuracy are unrelated." + "reason": "TF32 behavior API warning and FineGrainedFP8 accuracy regression are not the same underlying bug." }, { "left": "issue:44075", "right": "issue:45656", "accept": false, - "reason": "SGD args ignored and optimizer step called twice with deepspeed are different optimizer bugs." + "reason": "SGD args not used and optimizer.step called twice with deepspeed are distinct optimizer bugs." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "Both mention float32/dtype, but one is a compile-time warning and the other is a loading dtype propagation bug." + "reason": "A compile-time float32 warning for bf16 models is different from model.dtype staying float32 after loading Blip2." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "General chat/completions rejection and streaming inference failure on Qwen3.5-0.8B are not the same concrete issue." + "reason": "Both touch chat/completions, but one is a general request-acceptance failure and the other is streaming inference on Qwen3.5-0.8B; not the same concrete issue." }, { "left": "issue:42175", "right": "issue:42222", "accept": false, - "reason": "Backend packaging/install issue versus VitPose model breakage." + "reason": "TensorFlow missing from a torch install and broken vitpose models are unrelated." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Tie-embedding API removal and unexpected position-id key warnings are unrelated." + "reason": "Removed tie_embeddings_and_encoder_decoder and unexpected position-id keys are different regression classes." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache missing during chunked prefill is unrelated to testing_utils." + "reason": "Generation cache missing in chunked_prefill and a testing_utils bug are unrelated." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "AutoImageProcessor import error and PreTrainedConfig type-checking regression are different import/API issues." + "reason": "ImportError for AutoImageProcessor and PreTrainedConfig type-checking breakage are different API regressions." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "GLM-MoE torch.split behavior and GLM-5 RoPE implementation are separate model internals problems." + "reason": "torch.split return-value handling and GLM-5 RoPE implementation are not the same bug." } ] }, @@ -14571,8 +14642,8 @@ "issue:43874|issue:44261", "issue:45071|issue:45542", "issue:27957|issue:43352", - "issue:44263|issue:46172", "issue:44484|issue:45478", + "issue:44263|issue:46172", "issue:44360|issue:46172", "issue:43931|issue:44410", "issue:43441|issue:44829", @@ -14585,8 +14656,8 @@ "issue:43673|issue:46123", "issue:45753|issue:46369", "issue:44117|issue:44987", - "issue:43874|issue:44263", "issue:43986|issue:44484", + "issue:43874|issue:44263", "issue:43531|issue:45341", "issue:45440|issue:46172", "issue:43986|issue:45372", @@ -14597,20 +14668,20 @@ "issue:38617|issue:42757", "issue:43502|issue:43992", "issue:44062|issue:44117", - "issue:44038|issue:44410", "issue:44410|issue:45375", + "issue:44038|issue:44410", "issue:42890|issue:46298", "issue:38617|issue:44351", "issue:43097|issue:43873" ], - "prepared_review_unit_hash": "846b1548e376e2024c20df18aacb9cd416e110322c4459e046c1e5ea0a7e7b64", + "prepared_review_unit_hash": "9e425f4e53b74e13853e24ee46aadf00204bf2d389b49431f7778b16b98d6f09", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, @@ -14625,216 +14696,216 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All soft pairs look like superficially related but distinct reports; none are close enough to merge as duplicates. The set is a mixed bag of model-specific failures, unrelated import regressions, and independent documentation/test issues.", - "confidence": 0.95, - "canonical_issue_reason": "issue:43064 is the most concrete and high-impact bug report in the set, with a specific failing configuration and a clear reproduction path, so it is the best anchor for this mixed cluster.", + "summary": "The cluster is very heterogeneous: most items are distinct model-specific regressions, import errors, or unrelated docs/tests issues. I did not find any soft pair that looks like the same underlying bug/change, so all soft edges are rejected.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43097 is the best anchor because it\u2019s a broad v5 release regression with downstream impact, whereas most other items are narrower model- or subsystem-specific failures.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43064 is the strongest representative issue overall because it is specific, actionable, and describes a real runtime failure rather than a vague symptom or narrow API mismatch.", + "best_issue_reason": "issue:43097 is the most representative and highest-level issue in the set: a release-wide API removal/regression that plausibly sits near the center of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Different bugs: GLM46V image-patch API vs MLA RMS epsilon precision/config handling." + "reason": "Different bugs in different multimodal models: missing image-patch helper vs missing rms_norm_eps in a layernorm config." }, { "left": "issue:45071", "right": "issue:45542", "accept": false, - "reason": "Unrelated regressions: PretrainedConfig type-checking vs TensorFlow backend detection." + "reason": "Unrelated failures: PretrainedConfig type-checking regression vs TensorFlow backend import handling." }, { "left": "issue:27957", "right": "issue:43352", "accept": false, - "reason": "Both mention Flash Attention 2, but they concern different models and different support/status issues." + "reason": "Both mention Flash Attention 2, but they concern different models and different support paths." }, { - "left": "issue:44263", - "right": "issue:46172", + "left": "issue:44484", + "right": "issue:45478", "accept": false, - "reason": "Different code paths: GlmMoeDsaIndexer split semantics vs DeepSeek-V4 tensor transpose/cast." + "reason": "save_pretrained shard-size question is unrelated to Qwen3.5 MoE loading errors." }, { - "left": "issue:44484", - "right": "issue:45478", + "left": "issue:44263", + "right": "issue:46172", "accept": false, - "reason": "Shard-size documentation question vs Qwen3.5 MoE from_pretrained failure are not the same bug." + "reason": "Different code paths: GlmMoeDsaIndexer torch.split behavior vs deepseek-v4 dtype/transpose logic." }, { "left": "issue:44360", "right": "issue:46172", "accept": false, - "reason": "DSA indexer missing ReLU is unrelated to the DeepSeek-V4 comb transpose issue." + "reason": "DSA indexer ReLU omission is unrelated to deepseek-v4 comb transpose behavior." }, { "left": "issue:43931", "right": "issue:44410", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and qwen3next missing projections are different model-loading problems." + "reason": "Both are Qwen-family, but one is a weight-shape load mismatch and the other is missing attention projections." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "Both involve FlashAttention, but one is model support failure and the other is degenerate training with FA3." + "reason": "Both involve flash attention, but the models, versions, and failure modes are different enough to be separate bugs." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Different Qwen vision-family regressions with different symptoms and likely different fixes." + "reason": "Different Qwen-VL/MoE loading problems with no clear shared underlying defect." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "Removed config field vs gguf_file loading breakage are unrelated loader regressions." + "reason": "Tie-embeddings removal is a release API regression; gguf_file loading is a separate tokenizer/model-loading issue." }, { "left": "issue:43335", "right": "issue:45071", "accept": false, - "reason": "SwitchTransformers sparse-layer construction bug is unrelated to the PretrainedConfig typing regression." + "reason": "SwitchTransformers sparse-layer misconfiguration is unrelated to config type-checking failures." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "Unsafe sys.modules access and a testing_utils bug are different issues in different files." + "reason": "sys.modules access bug and a testing utility bug are different code paths and symptoms." }, { "left": "issue:43986", "right": "issue:45341", "accept": false, - "reason": "Video AutoProcessor crash without torchvision is unrelated to a testing helper bug." + "reason": "Video-model loading without torchvision is unrelated to the testing_utils bug." }, { "left": "issue:44933", "right": "issue:46231", "accept": false, - "reason": "Missing image_utils import and a docs parameter-name mismatch are not the same underlying change." + "reason": "Missing image_utils import and a docs parameter-name mismatch are not the same issue." }, { "left": "issue:43673", "right": "issue:46123", "accept": false, - "reason": "Generation cache handling in chunked_prefill is separate from MaskGenerationPipeline final-batch dropping." + "reason": "Cache handling in GenerationMixin and final-batch pipeline bookkeeping are separate bugs." }, { "left": "issue:45753", "right": "issue:46369", "accept": false, - "reason": "Infinite loop on one Qwen3_5 image and SAM3 litetext memory overflow are different model failures." + "reason": "Infinite-loop generation on one image is unrelated to memory overflow in a different model." }, { "left": "issue:44117", "right": "issue:44987", "accept": false, - "reason": "Tokenizer mapping None handling is unrelated to loading the physical-intelligence/fast model." + "reason": "Tokenizer mapping returning None and a specific model load failure are not clearly the same defect." }, { - "left": "issue:43874", - "right": "issue:44263", + "left": "issue:43986", + "right": "issue:44484", "accept": false, - "reason": "Missing image-patch counting method vs torch.split return-value handling are distinct API issues." + "reason": "No meaningful overlap: torchvision-dependent processor crash vs shard-size behavior in save_pretrained." }, { - "left": "issue:43986", - "right": "issue:44484", + "left": "issue:43874", + "right": "issue:44263", "accept": false, - "reason": "Video-model torchvision crash and save_pretrained shard-size behavior are unrelated." + "reason": "Different model internals: missing image patch count method vs torch.split return-value handling." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and a testing utility bug do not share a code path." + "reason": "Qwen3-MoE sliding-window bug is unrelated to a testing utility issue." }, { "left": "issue:45440", "right": "issue:46172", "accept": false, - "reason": "DeepseekV3MoE divergence from remote implementation is a broader modeling mismatch, not the DeepSeek-V4 transpose bug." + "reason": "DeepSeekV3 MoE divergence and deepseek-v4 comb transpose are different implementation problems." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "One is a missing torchvision dependency crash; the other is a mistral_common import regression." + "reason": "No torchvision crash and missing ReasoningEffort import are unrelated import/dependency failures." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "RMS eps precision in MLA and GLM-5 RoPE implementation are separate architecture bugs." + "reason": "Precision/epsilon issue in MLA q_a_layernorm is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:43097", "right": "issue:45588", "accept": false, - "reason": "Removed tie_embeddings flag is unrelated to the flash_attention.py s_aux=None crash." + "reason": "Tie-embeddings API removal is unrelated to flash_attention.py crashing on s_aux=None." }, { "left": "issue:43097", "right": "issue:46050", "accept": false, - "reason": "Config-flag removal and quantized save/load corruption are different bugs." + "reason": "API removal/regression is distinct from quantized model save/load corruption." }, { "left": "issue:43064", "right": "issue:44315", "accept": false, - "reason": "Trainer/FSDP2 optimizer-state corruption is unrelated to Liger Kernel not being applied with model_init." + "reason": "FSDP2+PEFT optimizer-state corruption and Liger application with model_init are different training paths." }, { "left": "issue:38617", "right": "issue:42757", "accept": false, - "reason": "Different missing imports from different modules and different causes." + "reason": "Both are import errors, but they involve different symbols from different packages and likely different causes." }, { "left": "issue:43502", "right": "issue:43992", "accept": false, - "reason": "local_files_only network leakage and missing embed_tokens.weight on UMT5Encoder are unrelated." + "reason": "local_files_only network leakage is unrelated to missing embed_tokens.weight on UMT5Encoder load." }, { "left": "issue:44062", "right": "issue:44117", "accept": false, - "reason": "AddedToken special-argument duplication and tokenizer mapping fallback are separate tokenizer bugs." + "reason": "AddedToken keyword collision and tokenizer mapping returning None are separate tokenizer bugs." }, { - "left": "issue:44038", - "right": "issue:44410", + "left": "issue:44410", + "right": "issue:45375", "accept": false, - "reason": "Qwen3-VL-Moe loading bug and qwen3next missing projections are different model-definition issues." + "reason": "Missing qwen3next projections and missing deepstack_visual_indexes are different config/model-definition issues." }, { - "left": "issue:44410", - "right": "issue:45375", + "left": "issue:44038", + "right": "issue:44410", "accept": false, - "reason": "Missing qwen3next projections and missing deepstack_visual_indexes are different config/model bugs." + "reason": "Qwen3-VL-Moe from_pretrained failure is not the same as qwen3next layer-0 projection omission." }, { "left": "issue:42890", "right": "issue:46298", "accept": false, - "reason": "Flaky SamHQ integration tests and noisy logging are not the same underlying problem." + "reason": "Flaky integration test from missing set_seed and noisy logging are unrelated." }, { "left": "issue:38617", "right": "issue:44351", "accept": false, - "reason": "layer_type_validation import error and HybridCache import error are distinct missing-symbol regressions." + "reason": "Different missing imports from different modules; no shared underlying bug is evident." }, { "left": "issue:43097", "right": "issue:43873", "accept": false, - "reason": "Removed tie_embeddings flag and quantization offloading failure are unrelated changes." + "reason": "Tie-embeddings API removal is unrelated to offloading/quantization behavior." } ] }, @@ -14915,8 +14986,8 @@ "issue:43824|issue:43986", "issue:44938|issue:46178", "issue:43066|issue:43906", - "issue:43992|issue:44484", "issue:43208|issue:43550", + "issue:43992|issue:44484", "issue:43012|issue:44368", "issue:43012|issue:43550", "issue:43825|issue:44655", @@ -14933,8 +15004,8 @@ "issue:44380|issue:46132", "issue:43064|issue:44380", "issue:44222|issue:45478", - "issue:36683|issue:43582", "issue:44483|issue:44514", + "issue:36683|issue:43582", "issue:43208|issue:43526", "issue:42175|issue:46365", "issue:42490|issue:43012", @@ -14942,232 +15013,220 @@ "issue:42757|issue:43824", "issue:38617|issue:43502" ], - "prepared_review_unit_hash": "5372afc6951c97c61f611bc50a4cf2de54f6393f6768492878dfe28d402ba6e6", + "prepared_review_unit_hash": "59431ae160898bac43f0c230dbe02f076158a241ffcc45b9a9ee42e4fc27c224", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37447, - "estimated_input_tokens": 9362, - "estimated_eval_tokens": 18980 + "serialized_chars": 37444, + "estimated_input_tokens": 9361, + "estimated_eval_tokens": 18978 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a heterogeneous set of Transformers issues with only superficial text overlap. The soft pairs mostly share keywords like import, save, warning, or backend, but they describe different bugs and code paths, so none should be merged.", - "confidence": 0.18, - "canonical_issue_reason": "Issue 43502 is the most discussed, self-contained runtime bug in the set, so it is the best weak anchor for triage; however, the cluster is too mixed to have a strong canonical issue.", + "summary": "The soft links are all false positives; this is a heterogeneous set of unrelated import, training, model, backend, and warning issues rather than one duplicate cluster.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue: the items do not share one underlying bug or code path closely enough to consolidate.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43502 has the clearest repro-style report and the highest discussion activity, making it the most actionable single issue to anchor this cluster, even though it is not a true representative of all items.", + "best_issue_reason": "No strong global representative exists because the issues span unrelated failures across different subsystems and model families.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44936", "right": "issue:46365", "accept": false, - "reason": "Trainer.train()/evaluate lifecycle regression vs. an import failure involving datasets; different failure modes and code paths." + "reason": "Different failures: evaluation breaks after training vs import-time Trainer conflict with datasets." }, { "left": "issue:44704", "right": "issue:45103", "accept": false, - "reason": "Passing kwargs to cached_file is unrelated to the auto_docstring crash under future annotations." + "reason": "Different code paths: cached_file kwargs forwarding vs auto-docstring annotation introspection." }, { "left": "issue:44380", "right": "issue:46421", "accept": false, - "reason": "GPT2 attention scaling under SDPA/FlashAttention is a different bug from cache lazy-init recompilation with chunked prefill." + "reason": "Attention scaling under SDPA/FlashAttention is unrelated to cache lazy initialization recompilation." }, { "left": "issue:43986", "right": "issue:44117", "accept": false, - "reason": "Missing torchvision in AutoProcessor/video loading is unrelated to TOKENIZER_MAPPING_NAMES returning None in from_pretrained." + "reason": "Missing torchvision for video AutoProcessor 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 exported model class and the other is a torchvision dependency crash." + "reason": "Both are import/loading problems, but one is a missing exported class and the other is a missing optional dependency." }, { "left": "issue:44938", "right": "issue:46178", "accept": false, - "reason": "Python 3.14 load failure is unrelated to the tensorflow_text backend-mapping error." + "reason": "Python 3.14 load failure is unrelated to a bad BACKENDS_MAPPING backend name." }, { "left": "issue:43066", "right": "issue:43906", "accept": false, - "reason": "Wrong tokenizer decoder type in v5 is not the same as an isolated reproduction of another bug." - }, - { - "left": "issue:43992", - "right": "issue:44484", - "accept": false, - "reason": "Missing embed_tokens.weight on load is unrelated to the save_pretrained max_shard_size question." + "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue are not the same bug." }, { "left": "issue:43208", "right": "issue:43550", "accept": false, - "reason": "xLSTM small-model training blockers and Bamba SDPA compile failures are different model-specific bugs." + "reason": "xLSTM training blockers and Bamba torch.compile+SDPA failure affect different models and paths." }, { - "left": "issue:43012", - "right": "issue:44368", + "left": "issue:43992", + "right": "issue:44484", "accept": false, - "reason": "A bfloat16 compile warning is unrelated to the tie_word_embeddings config warning." + "reason": "Missing embed_tokens.weight on load is unrelated to max_shard_size behavior in save_pretrained." }, { "left": "issue:43012", - "right": "issue:43550", + "right": "issue:44368", "accept": false, - "reason": "Precision-warning behavior during compile is not the same as the Bamba torch.compile/SDPA failure." + "reason": "A bfloat16 precision warning is not the same as the tie_word_embeddings warning." }, { "left": "issue:43825", "right": "issue:44655", "accept": false, - "reason": "A pipeline error message about translations is unrelated to inability to save Pipeline objects." + "reason": "Pipeline task messaging in v5 is unrelated to saving Pipeline objects." }, { "left": "issue:45712", "right": "issue:46298", "accept": false, - "reason": "Leftover dummy classes in no-torch builds is repo hygiene, not the same as a noisy logging message." - }, - { - "left": "issue:42907", - "right": "issue:44655", - "accept": false, - "reason": "Saving dequantized models and saving Pipeline objects are both save-related, but they are different APIs and failure modes." + "reason": "Dummy PT object leakage is unrelated to a noisy logging message." }, { "left": "issue:42907", "right": "issue:45753", "accept": false, - "reason": "Saving dequantized Ministrals/Devstrals has nothing to do with a Qwen3.5 image-specific infinite loop." + "reason": "Saving dequantized models and a Qwen3_5 image loop are different bugs." }, { "left": "issue:43097", "right": "issue:46298", "accept": false, - "reason": "A removed tie_embeddings_and_encoder_decoder warning is unrelated to generic noisy logging." + "reason": "Config API removal is unrelated to noisy logging." }, { "left": "issue:43352", "right": "issue:44380", "accept": false, - "reason": "Nemotron FlashAttention support and GPT2 attention-scaling behavior are distinct attention issues with different code paths." + "reason": "Nemotron FlashAttention 2 support gap is not the same as GPT2 attention scaling ignored under backends." }, { - "left": "issue:43352", - "right": "issue:43976", + "left": "issue:43976", + "right": "issue:43352", "accept": false, - "reason": "FlashAttention support for Nemotron is unrelated to the Python-version load failure." + "reason": "Python version incompatibility is unrelated to a model-specific FlashAttention support issue." }, { "left": "issue:44162", "right": "issue:46172", "accept": false, - "reason": "ESM2 breakage and a deepseek tensor transpose bug are unrelated model-specific failures." + "reason": "ESM2 breakage and deepseek-v4 transpose behavior are different failures." }, { "left": "issue:45584", "right": "issue:46365", "accept": false, - "reason": "Whisper empty-transcription generation failure is unrelated to the Trainer import problem." + "reason": "Whisper empty-transcription generation failure is unrelated to importing Trainer with datasets installed." }, { "left": "issue:44077", "right": "issue:46365", "accept": false, - "reason": "patchtsmixer post_init restrictions are unrelated to the Trainer import failure caused by datasets." + "reason": "patchtsmixer post_init validation has nothing to do with the Trainer import conflict." }, { "left": "issue:44380", "right": "issue:45479", "accept": false, - "reason": "GPT2 attention scaling under backends is not the same as a degenerate zero-loss classification bug." + "reason": "Attention backend scaling bug and zero-loss classification bug are different code paths." }, { "left": "issue:44380", "right": "issue:46132", "accept": false, - "reason": "Attention scaling ignored under SDPA/FlashAttention is unrelated to AttentionInterface.register behavior changes." + "reason": "Generic AttentionInterface register behavior change is not the same as GPT2 attention scaling being ignored." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "FSDP2/PEFT optimizer-state corruption is a training-distributed bug, not the GPT2 attention-scaling issue." + "reason": "FSDP2/PEFT optimizer-state corruption is unrelated to GPT2 attention scaling." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "FP8 save_pretrained on MoE and Qwen3.5 from_pretrained errors are both model-family-related but involve different operations and failure points." + "reason": "FP8 save_pretrained on MoE models and Qwen3.5 MoE from_pretrained errors are not the same concrete bug." }, { - "left": "issue:36683", - "right": "issue:43582", + "left": "issue:44483", + "right": "issue:44514", "accept": false, - "reason": "Gemma3Config missing vocab_size is unrelated to the Apple Silicon caching_allocator_warmup TypeError." + "reason": "Chat completions API rejection is unrelated to Qwen2_5_VLProcessor batched padding crash." }, { - "left": "issue:44483", - "right": "issue:44514", + "left": "issue:36683", + "right": "issue:43582", "accept": false, - "reason": "Chat/completions request rejection is unrelated to the Qwen2.5 VL batched padding crash." + "reason": "Missing Gemma3Config vocab_size 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." + "reason": "xLSTM training blockers are unrelated to BeitImageProcessorFast reduce_labels behavior." }, { "left": "issue:42175", "right": "issue:46365", "accept": false, - "reason": "TensorFlow backend packaging and a Trainer import conflict with datasets are different dependency issues." + "reason": "Backend packaging for pip install '.[torch]' is unrelated to a Trainer import conflict with datasets." }, { "left": "issue:42490", "right": "issue:43012", "accept": false, - "reason": "Trailing-slash local model loading warnings are unrelated to bf16 compile precision warnings." + "reason": "Local-path trailing-slash warning and bfloat16 precision warning are different issues." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "A bf16 precision warning is not the same as compiled flex_attention failing on torch>=2.9." + "reason": "A precision warning is not the same as a torch>=2.9 flex_attention failure." }, { "left": "issue:42757", "right": "issue:43824", "accept": false, - "reason": "Missing is_offline_mode from huggingface_hub and missing Qwen2_5_VLForConditionalGeneration are both import errors, but they are different symbols with different causes." + "reason": "Different missing imports: is_offline_mode from huggingface_hub vs Qwen2_5_VLForConditionalGeneration from transformers." }, { "left": "issue:38617", "right": "issue:43502", "accept": false, - "reason": "The layer_type_validation import error and local_files_only still making API requests are unrelated issues." + "reason": "Importing layer_type_validation and making API requests despite local_files_only are unrelated." } ] }, @@ -15262,8 +15321,8 @@ "issue:43867|issue:45216", "issue:44625|issue:46077", "issue:43976|issue:44938", - "issue:43352|issue:45593", "issue:46077|issue:46082", + "issue:43352|issue:45593", "issue:36683|issue:43525", "issue:43618|issue:45216", "issue:36683|issue:44625", @@ -15274,226 +15333,226 @@ "issue:44355|issue:44855", "issue:43867|issue:46202" ], - "prepared_review_unit_hash": "ed7b0f3b5ee19249bec846af432b35bfd51bd76ad84b3324a9de9784f64cb969", + "prepared_review_unit_hash": "761607b7b3c05db5d856b202fe79e4c2cdf02fadbda4a393101a7b8da02c386c", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 32, - "serialized_chars": 36746, - "estimated_input_tokens": 9187, - "estimated_eval_tokens": 18630 + "serialized_chars": 36740, + "estimated_input_tokens": 9185, + "estimated_eval_tokens": 18626 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose collection of Transformers bugs across unrelated models, APIs, and regressions. The soft edges do not look like true duplicates; they mostly share generic symptoms like loading/import failures or v5 regressions without the same underlying code path.", - "confidence": 0.82, - "canonical_issue_reason": "Issue 45698 is the closest broad representative because it covers a general from_pretrained/save_pretrained loading regression, which is the most central theme among the more general items. That said, the cluster is heterogeneous, so any canonical choice is weak.", + "summary": "This cluster is a loose collection of Transformers issues around v5 regressions, import/load failures, model-specific config bugs, and docs/behavior mismatches. I don\u2019t see any soft pair that clearly represents the same underlying bug or change.", + "confidence": 0.39, + "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report in the set ('many import errors after update from 4.57.0 to 5.5.0'), so it is the best representative issue if one must be chosen.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45698 is the best available anchor for the cluster\u2019s loose load/save regression theme, but it is not a strong duplicate hub because many other items are model-specific or unrelated API regressions.", + "best_issue_reason": "issue:45278 is the most general, high-signal issue here and best fits the overall v5 import/regression theme.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45600", "right": "issue:46298", "accept": false, - "reason": "One is a CI break from removed SAM config references; the other is a vague noisy logging complaint. Different problems and code paths." + "reason": "One is a CI break from stale Sam3LiteText auto-mappings; the other is a noisy logging message. Different bug classes." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "Both mention Qwen3-VL-Moe/fp8, but one is a general transformers 5.0 load bug and the other is inference-only fp8 support. Not the same concrete bug." + "reason": "Both mention Qwen3-VL/Moe, but one is a loading bug and the other is fp8 inference support; not the same concrete defect." }, { "left": "issue:45698", "right": "issue:46172", "accept": false, - "reason": "General save/load wrong custom module issue versus a deepseek-v4 tensor transpose/dtype bug. Different failure modes and subsystems." + "reason": "Different problems: custom-module reload after save_pretrained versus a deepseek-v4 tensor transpose issue." }, { "left": "issue:43618", "right": "issue:44380", "accept": false, - "reason": "CLIP output attentions regression is unrelated to GPT2 attention scaling under SDPA/FlashAttention." + "reason": "CLIP attentions missing versus GPT2 attention scaling ignored under SDPA/FlashAttention are separate code paths and symptoms." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Both are CLIP-related, but one is an output-field assignment issue and the other is a forward regression returning hidden_states=None. Different concrete breakages." + "reason": "Both are CLIP regressions, but one concerns attentions and the other hidden_states; distinct missing outputs." }, { "left": "issue:44483", "right": "issue:45290", "accept": false, - "reason": "One is a v5 chat completions API acceptance problem; the other is a tokenizer/template crash on assistant tool-call messages. Different request paths." + "reason": "An API request acceptance regression is unrelated to chat template crashes on tool-call messages." }, { "left": "issue:44625", "right": "issue:46365", "accept": false, - "reason": "Qwen3.5 num_labels propagation is unrelated to importing Trainer being blocked by datasets." + "reason": "Qwen3.5 num_labels propagation is unrelated to a Trainer import conflict with datasets." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3VL-Moe loading regression versus generic state_dict-sorted load error. Similar surface area, but not the same bug." + "reason": "Qwen3VL-Moe loading breakage is not the same as a generic state_dict-sorting load error." }, { "left": "issue:42994", "right": "issue:44655", "accept": false, - "reason": "Quantized model saving failure and pipeline save_pretrained failure are both save-related, but the affected objects and code paths differ." + "reason": "Quantized model saving failure and pipeline save_pretrained failure are both save-related, but they are different failures." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Gemma3Config missing vocab_size is unrelated to SAM3 vision encoder output missing fpn_position_embeddings." + "reason": "Gemma3Config vocab_size missing and SAM 3 Video missing fpn_position_embeddings are unrelated model-specific attribute errors." }, { "left": "issue:43066", "right": "issue:46209", "accept": false, - "reason": "Tokenizer decoder type mismatch and FP8 inference failure are unrelated bugs." + "reason": "Tokenizer decoder type issue and FP8 inference on B200 are unrelated regressions." }, { "left": "issue:43756", "right": "issue:44485", "accept": false, - "reason": "Smollm3 RoPE layer dropping and GLM-5 RoPE implementation are both RoPE-themed but not the same concrete defect." + "reason": "Smollm3 RoPE-layer mismatch and GLM-5 RoPE implementation are different model/configuration issues." }, { "left": "issue:45276", "right": "issue:46077", "accept": false, - "reason": "Gemma4 resize_token_embeddings propagation bug versus Gemma4 bidirectional attention mask construction bug. Same model family, different code paths." + "reason": "Gemma4 token embedding resize bug and bidirectional attention mask bug affect different mechanisms." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "Docs mismatch for return_all_scores is unrelated to pipeline save_pretrained serialization." + "reason": "Docs mismatch for return_all_scores is not the same as pipeline serialization failure." }, { "left": "issue:36010", "right": "issue:42757", "accept": false, - "reason": "Both are import errors, but they concern different symbols from different dependencies and unrelated breakages." + "reason": "Different import errors from different dependencies: transformers.generation vs huggingface_hub." }, { "left": "issue:43611", "right": "issue:44655", "accept": false, - "reason": "base_model_prefix loading regression is separate from inability to save Pipeline objects." + "reason": "base_model_prefix load failure and pipeline save_pretrained failure are distinct serialization/load paths." }, { "left": "issue:43352", "right": "issue:43937", "accept": false, - "reason": "Nemotron flash-attn support gap and GLM-5 generation config validation are different model-specific issues." + "reason": "Nemotron FlashAttention 2.0 support and GLM-5 GenerationConfig validation are unrelated model-specific bugs." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "Gemma3 vocab_size absence is unrelated to patchtsmixer post_init API constraints." + "reason": "Both are attribute/config issues, but Gemma3Config vocab_size and patchtsmixer post_init allowance are unrelated." }, { "left": "issue:43867", "right": "issue:45216", "accept": false, - "reason": "Generic sorted-state_dict load error is not the same as Qwen3.5 save_pretrained regression." + "reason": "A generic load error with sorted state_dict is not the same as Qwen3.5 checkpoint corruption on save_pretrained." }, { "left": "issue:44625", "right": "issue:46077", "accept": false, - "reason": "Qwen3.5 num_labels propagation and Gemma4 attention masking are unrelated model configuration bugs." + "reason": "Qwen3.5 num_labels propagation and Gemma4 attention-mask behavior are unrelated." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Python-version packaging issue for Transformers 5.1.0 is distinct from Python 3.14 import/load failures." + "reason": "Python version compatibility for 5.1.0 and Python 3.14 load failures are different environment/version regressions." }, { - "left": "issue:43352", - "right": "issue:45593", + "left": "issue:46077", + "right": "issue:46082", "accept": false, - "reason": "Flash Attention 2 support for Nemotron and missing auxiliary losses in D-FINE are unrelated." + "reason": "Gemma4 attention mask bug and LlamaConfig head_dim validation are separate config behaviors." }, { - "left": "issue:46077", - "right": "issue:46082", + "left": "issue:43352", + "right": "issue:45593", "accept": false, - "reason": "Gemma4 bidirectional attention masking and LlamaConfig head_dim validation are different model/config behaviors." + "reason": "FlashAttention support limitation for NemotronH and D-FINE auxiliary-loss handling are unrelated." }, { "left": "issue:36683", "right": "issue:43525", "accept": false, - "reason": "Gemma3Config vocab_size and Llama4Config pad_token_id are separate missing-attribute regressions." + "reason": "Gemma3Config lacks vocab_size, while Llama4Config lacks pad_token_id; similar shape, but different models and fixes." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP attentions regression is unrelated to Qwen3.5 checkpoint save corruption." + "reason": "Missing CLIP attentions and Qwen3.5 checkpoint save regression are unrelated issues." }, { "left": "issue:36683", "right": "issue:44625", "accept": false, - "reason": "Different config propagation bugs in different models: missing Gemma3 vocab_size versus Qwen3.5 num_labels not copied." + "reason": "Both are config propagation issues, but for different models and different fields." }, { "left": "issue:43992", "right": "issue:45310", "accept": false, - "reason": "UMT5Encoder missing embed_tokens.weight is unrelated to Qwen3.5 MoE from_pretrained failure." + "reason": "Missing embed_tokens.weight in UMT5Encoder is not the same as Qwen3.5 MoE from_pretrained loading error." }, { "left": "issue:43064", "right": "issue:44492", "accept": false, - "reason": "FSDP2/PEFT optimizer-state mismatch is unrelated to a cache strategy typo." + "reason": "Distributed optimizer-state corruption and a cache-strategy typo are unrelated." }, { "left": "issue:44075", "right": "issue:46365", "accept": false, - "reason": "Optimizer SGD args ignored is not the same as Trainer import being affected by datasets installation." + "reason": "SGD args not used in optimizer code and Trainer import conflict with datasets are different bugs." }, { "left": "issue:43010", "right": "issue:45278", "accept": false, - "reason": "Torch no_grad decoration for cache updates and broad import errors after upgrade are different issues." + "reason": "Cache/layer update decoration and broad import errors after upgrade are different issue areas." }, { "left": "issue:44355", "right": "issue:44855", "accept": false, - "reason": "Compiled Python file import errors are unrelated to the Python 3.13 torch.jit.script indentation/parsing bug." + "reason": "Compiled Python file loading errors and Python 3.13 torch.jit script parsing errors are different failure modes." }, { "left": "issue:43867", "right": "issue:46202", "accept": false, - "reason": "Generic sorted-state_dict load failure is not the same as Gemma pipeline roundtrip behavior on the \u2581 token." + "reason": "Generic sorted-state_dict load error and Gemma pipeline roundtrip failure are not the same underlying bug." } ] }, @@ -15569,15 +15628,15 @@ "soft_pairs": [ "issue:43867|issue:45375", "issue:44938|issue:45071", - "issue:36683|issue:43064", "issue:41669|issue:44246", + "issue:36683|issue:43064", "issue:42890|issue:45561", "issue:43010|issue:45850", "issue:43992|issue:45478", "issue:43867|issue:45357", - "issue:36296|issue:43575", - "issue:43010|issue:43976", "issue:43618|issue:44077", + "issue:43010|issue:43976", + "issue:36296|issue:43575", "issue:28282|issue:33453", "issue:43618|issue:46285", "issue:43618|issue:44857", @@ -15591,8 +15650,8 @@ "issue:44222|issue:45125", "issue:43010|issue:43065", "issue:43576|issue:46132", - "issue:43502|issue:44485", "issue:42548|issue:44351", + "issue:43502|issue:44485", "issue:45125|issue:45478", "issue:44355|issue:44749", "issue:30333|issue:33453", @@ -15601,232 +15660,232 @@ "issue:43502|issue:44222", "issue:30990|issue:31795" ], - "prepared_review_unit_hash": "f3f1acbd45a1282d2a9ec9cb1efc0a1268c318942dbbb19826b602824a2c522a", + "prepared_review_unit_hash": "020e1b4ae0f443f97ad3bb476c062223778daeed01725a7dff62a4d41ceccc0d", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 37120, - "estimated_input_tokens": 9280, - "estimated_eval_tokens": 18816 + "serialized_chars": 37116, + "estimated_input_tokens": 9279, + "estimated_eval_tokens": 18814 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is very loose: most pairs are only superficially similar because they mention loading, v5, or a specific model name. I do not see any true duplicate pair among the soft edges. The cluster is best treated as a mixed bag of unrelated Transformers regressions and documentation/test issues.", - "confidence": 0.19, - "canonical_issue_reason": "issue:43723 is the clearest representative of the main recurring theme in this cluster: a concrete v5-era loading failure affecting end users, without being overly niche.", + "summary": "The cluster is mostly heterogeneous: many items are about unrelated loading, model-specific, CI, performance, or API-regression topics. I found no soft pair that looks like the same underlying bug/change with a plausible single-PR merge path, so all candidate edges are rejected.", + "confidence": 0.77, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43723 is the most suitable global issue here because it describes a direct, user-facing loading regression and is easier to anchor than the many narrowly scoped or unrelated reports.", + "best_issue_reason": "issue:43502 is the clearest standalone bug report in the set (offline loading still making API calls), but it is not a duplicate center for the rest of the items.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "Different bugs: one is a generic model load/state_dict ordering error, the other is a Qwen3.5 vision config field being dropped." + "reason": "Different failures: a general state_dict loading error vs a Qwen3.5 vision config field being dropped." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Unrelated failures: Python 3.14 import/load issue vs PretrainedConfig type-checking regression." + "reason": "Both are v5-related breakages, but one is Python 3.14 import failure and the other is PretrainedConfig type-checking; not the same bug." }, { - "left": "issue:36683", - "right": "issue:43064", + "left": "issue:41669", + "right": "issue:44246", "accept": false, - "reason": "Different code paths and symptoms: Gemma3Config missing vocab_size vs Trainer/FSDP2 optimizer-state corruption." + "reason": "Both mention import slowdown, but 41669 is a specific import-* cleanup proposal while 44246 is an intermittent import performance complaint." }, { - "left": "issue:41669", - "right": "issue:44246", + "left": "issue:36683", + "right": "issue:43064", "accept": false, - "reason": "Both mention import performance, but one is about removing import * from models while the other is about slow imports; not the same change or bug." + "reason": "Unrelated: Gemma3Config missing vocab_size versus FSDP2/PEFT optimizer-state corruption." }, { "left": "issue:42890", "right": "issue:45561", "accept": false, - "reason": "Test flakiness in one SAM HQ test is unrelated to xdist race conditions in patched testing utilities." + "reason": "Test flakiness from missing set_seed is unrelated to xdist races on captured_info.txt." }, { "left": "issue:43010", "right": "issue:45850", "accept": false, - "reason": "Decorating cache/layer update methods with no_grad is a specific autograd fix, not a broad claim that v5 breaks too many models." + "reason": "A no_grad decorator change for cache updates is not the same as a broad 'too many models break' complaint." }, { "left": "issue:43992", "right": "issue:45478", "accept": false, - "reason": "Different models and missing weights vs a Qwen3.5 from_pretrained error; no shared concrete code path." + "reason": "Both are from_pretrained problems, but they involve different models and different missing/error paths." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "Generic load error vs Qwen3.5 visual-encoder key regression; same broad area but not the same bug." + "reason": "State-dict loading error and incorrect save_pretrained visual-encoder keys are different directions of failure." }, { - "left": "issue:36296", - "right": "issue:43575", + "left": "issue:43618", + "right": "issue:44077", "accept": false, - "reason": "Tensor-parallel training bug and tp OOM are both distributed-training themed, but they are distinct failures." + "reason": "CLIPOutput attentions regression is unrelated to patchtsmixer post_init validation rules." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "no_grad decorator fix is unrelated to Python version compatibility failures." + "reason": "No-grad decoration for cache updates is unrelated to a Python-version compatibility issue." }, { - "left": "issue:43618", - "right": "issue:44077", + "left": "issue:36296", + "right": "issue:43575", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to patchtsmixer post_init semantics." + "reason": "Tensor-parallel training bug and tensor-parallel OOM on a specific model are different concrete problems." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "PyTorch missing ImportError is not the same as tokenizer-loading regression." + "reason": "PyTorch-missing import error is unrelated to tokenizer loading regression." }, { "left": "issue:43618", "right": "issue:46285", "accept": false, - "reason": "Both concern CLIP-related structure, but one is attentions assignment and the other is text model nesting; different bugs." + "reason": "Different CLIP regressions: attentions field assignment vs model structure inconsistency." }, { "left": "issue:43618", "right": "issue:44857", "accept": false, - "reason": "CLIP attentions regression and float16 AMP crash in LwDetrImageLoss are unrelated." + "reason": "CLIPOutput attentions and float16 CUDA crash in LwDetrImageLoss are unrelated bugs." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "DSA indexer/ReLU discussion is a different subsystem from DeepseekV3MoE implementation divergence." + "reason": "A DSA indexer ReLU discussion is unrelated to DeepseekV3 implementation divergence." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "local_files_only network behavior and AddedToken special-argument duplication are unrelated." + "reason": "Offline API calls and AddedToken special-argument duplication are unrelated loading/tokenization bugs." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "A flaky SAM HQ test is unrelated to loading Phi-4 multimodal model failures." + "reason": "Seed-related test flakiness is unrelated to Phi-4 multimodal model loading failure." }, { "left": "issue:42890", "right": "issue:45600", "accept": false, - "reason": "Test seeding flakiness is unrelated to auto_mappings referencing removed configs." + "reason": "SamHQ test seeding issue is unrelated to auto_mappings referencing removed Sam3LiteText configs." }, { "left": "issue:42890", "right": "issue:44112", "accept": false, - "reason": "Test randomness/stale device override failure is not the same as missing seeds in SAM HQ tests." + "reason": "One is a seed-driven integration-test flake; the other is a stale GraniteSpeech device-override test." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Tokenizer loading failure and v4.57-to-v5.2 checkpoint reload failure are both loading-related but distinct problems." + "reason": "Tokenizer loading regression and checkpoint reload incompatibility across versions are related only at a broad level, not the same bug." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "Int4 quantization OOM is unrelated to checkpoint reload compatibility across versions." + "reason": "Int4 quantization memory regression is different from checkpoint reload failure after upgrade." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "FP8 save_pretrained behavior for MoE is different from missing _tp_plan for tensor parallelism." + "reason": "FP8 save_pretrained on MoE is not the same as missing _tp_plan for tensor parallelism." }, { "left": "issue:43010", "right": "issue:43065", "accept": false, - "reason": "no_grad cache update fix and dummy Conv2d in Sam3PixelDecoder are unrelated implementation changes." + "reason": "Cache no_grad decoration and Sam3PixelDecoder dummy Conv2d are unrelated code changes." }, { "left": "issue:43576", "right": "issue:46132", "accept": false, - "reason": "Broken env command in v5 is unrelated to AttentionInterface.register changing callback behavior." + "reason": "Broken env command in v5 and AttentionInterface.register behavior change are separate CLI/API issues." }, { - "left": "issue:43502", - "right": "issue:44485", + "left": "issue:42548", + "right": "issue:44351", "accept": false, - "reason": "local_files_only network leakage and GLM-5 RoPE implementation are not the same bug." + "reason": "Cannot import PreTrainedModel and cannot import HybridCache are different export/import regressions." }, { - "left": "issue:42548", - "right": "issue:44351", + "left": "issue:43502", + "right": "issue:44485", "accept": false, - "reason": "cannot import PreTrainedModel and cannot import HybridCache are separate export/regression issues." + "reason": "Offline loading making API calls is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:45125", "right": "issue:45478", "accept": false, - "reason": "Missing _tp_plan for tensor parallelism and Qwen3.5 from_pretrained errors are related only by model family, not by the same underlying fix." + "reason": "Missing _tp_plan and Qwen3.5 from_pretrained error are model-specific but not the same concrete failure." }, { "left": "issue:44355", "right": "issue:44749", "accept": false, - "reason": "Compiled Python file errors and performance slowdown after upgrade are different failure modes." + "reason": "Compiled Python file errors are unrelated to a post-upgrade data-processing slowdown." }, { "left": "issue:30333", "right": "issue:33453", "accept": false, - "reason": "MLflow reporting on failed training jobs is unrelated to tokenizer loading regression." + "reason": "MLFlow reporting on failed training and tokenizer loading regression do not share the same underlying bug." }, { "left": "issue:45125", "right": "issue:45375", "accept": false, - "reason": "Tensor-parallel plan missing and dropped vision config fields affect different parts of Qwen3.5 MoE." + "reason": "Tensor-parallel plan missing is unrelated to a missing deepstack_visual_indexes field in vision config." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Cache/layer no_grad decoration is unrelated to generation kwargs synchronization after sync_gpus." + "reason": "Cache update no_grad decoration is not the same as _update_model_kwargs_for_generation after sync_gpus." }, { "left": "issue:43502", "right": "issue:44222", "accept": false, - "reason": "API requests despite local_files_only and FP8 save_pretrained are distinct loading vs serialization bugs." + "reason": "Local-files-only API leakage and FP8 save_pretrained behavior are different loading vs saving issues." }, { "left": "issue:30990", "right": "issue:31795", "accept": false, - "reason": "Sentence Transformers loading hang and forward doc confusion are unrelated issues." + "reason": "Sentence Transformers load hanging and confusing forward docs are unrelated." } ] }, @@ -15915,12 +15974,12 @@ "issue:43576|issue:43824", "issue:31795|issue:38617", "issue:44961|issue:45468", - "issue:43576|issue:44938", "issue:44222|issue:44230", + "issue:43576|issue:44938", "issue:41628|issue:42548", + "issue:46143|issue:46144", "issue:44908|issue:44961", "issue:44485|issue:44961", - "issue:46143|issue:46144", "issue:36296|issue:43901", "issue:36296|issue:43526", "issue:44315|issue:46209", @@ -15934,88 +15993,88 @@ "issue:44273|issue:45003", "issue:43576|issue:44556" ], - "prepared_review_unit_hash": "8872d0e5d7b800106aa3a22477e7b8a350a6ee3e9a8f70ffefed3aa9ecceb972", + "prepared_review_unit_hash": "edf4b57a04afba782c8fb91d005ffe201c727a81cc700af12f0e1909d8a04255", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 33, - "serialized_chars": 36717, - "estimated_input_tokens": 9180, - "estimated_eval_tokens": 18616 + "serialized_chars": 36721, + "estimated_input_tokens": 9181, + "estimated_eval_tokens": 18618 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The candidate pairs are mostly superficial similarities (often just shared version/model/subsystem keywords) rather than the same underlying bug. I would not merge any of the soft edges, and there is no clear canonical issue or PR for this heterogeneous set.", - "confidence": 0.92, - "canonical_issue_reason": null, + "summary": "This set is mostly a noisy similarity cluster of unrelated Transformer regressions, import errors, docs issues, and model-specific bugs. I did not find any soft pair that clearly describes the same underlying issue.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45850 is the closest umbrella report: it broadly captures the theme of v5 breaking multiple models, which is the only common thread across the set. The other issues are narrower and mostly unrelated symptoms.", "canonical_pr_reason": null, - "best_issue_reason": "No strong global best issue: the items span unrelated loading, import, FP8, docs, scheduler, and model-specific bugs, so no single issue safely represents the cluster.", + "best_issue_reason": "Issue 45850 is the best representative of the cluster\u2019s loose theme because it is broad and generic enough to subsume many of the v5-regression style complaints, though it is still not a true duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both involve model loading and version changes, but one is Qwen3-VL-MoE loading while the other is checkpoint reload after upgrading to v5.2/v5.3; too different to be the same bug." + "reason": "Both are model-loading regressions in v5, but one is Qwen3VL MoE-specific and the other is a generic checkpoint reload failure; different underlying bugs." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "Local-files-only network requests and a random issue title are unrelated." + "reason": "The first is about unwanted API calls with local_files_only, while the second is an unrelated, underspecified issue with no clear overlap." }, { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "FP8 accuracy problems and a missing image-patch API on GLM46V are different code paths and symptoms." + "reason": "Accuracy problems with FP8 training are unrelated to a missing image-processor method causing an AttributeError." }, { "left": "issue:30333", "right": "issue:36296", "accept": false, - "reason": "MLflow reporting on failed jobs is unrelated to tensor-parallel training behavior." + "reason": "MLflow reporting after failed jobs is unrelated to a tensor-parallel training bug." }, { "left": "issue:44355", "right": "issue:45071", "accept": false, - "reason": "Compiled Python file errors and PretrainedConfig type-checking regressions are unrelated." + "reason": "Compiled Python file errors and PretrainedConfig type-checking regressions are different failure modes." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing pipeline() and an unrelated issue title are not duplicates." + "reason": "Docs still referencing pipeline() is unrelated to the other issue, which does not identify the same bug." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "One is about running compiled Python files, the other about save_pretrained on Pipeline objects; different failures." + "reason": "One is about compiled Python files, the other about save_pretrained on Pipeline objects; different code paths." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Both are loading-related regressions, but tokenizer loading and Llama 3.2 Vision model loading are not the same bug." + "reason": "Both concern loading, but one is tokenizer loading and the other is Llama Vision model loading; not the same underlying defect." }, { "left": "issue:44315", "right": "issue:44360", "accept": false, - "reason": "Liger Kernel not applying in model_init and a DSA indexer ReLU discussion are unrelated." + "reason": "Liger kernel application in model_init has nothing to do with a DSA indexer missing ReLU." }, { "left": "issue:27957", @@ -16027,139 +16086,139 @@ "left": "issue:42757", "right": "issue:43576", "accept": false, - "reason": "An import error from huggingface_hub and a broken transformers env CLI command are different issues." + "reason": "An import error from huggingface_hub is not the same as a broken transformers env CLI command." }, { "left": "issue:43576", "right": "issue:45071", "accept": false, - "reason": "Broken env CLI in v5 and PretrainedConfig type-checking breakage are not the same code-path problem." + "reason": "A broken env subcommand and a PretrainedConfig type-checking regression are distinct issues." }, { "left": "issue:43576", "right": "issue:43824", "accept": false, - "reason": "The env command and Qwen2_5_VL import failure are separate regressions." + "reason": "The env CLI bug is unrelated to a specific model import failure." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion around model.forward arguments is unrelated to an ImportError for layer_type_validation." + "reason": "Documentation confusion around forward arguments is unrelated to an ImportError for a missing symbol." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "Random title vs Gemma4 audio positional encoding bug are unrelated." + "reason": "No evidence these describe the same bug; one is empty/unclear, the other is a Gemma audio positional encoding issue." }, { - "left": "issue:43576", - "right": "issue:44938", + "left": "issue:44222", + "right": "issue:44230", "accept": false, - "reason": "Broken env CLI and Python 3.14 load failure are both version-related, but not the same concrete defect." + "reason": "Both mention FP8/MoE, but one is about save_pretrained and the other about inference support; not the same concrete bug." }, { - "left": "issue:44222", - "right": "issue:44230", + "left": "issue:43576", + "right": "issue:44938", "accept": false, - "reason": "Both mention FP8/MoE, but one is save_pretrained behavior and the other is inference support; too broad to merge." + "reason": "A broken CLI command is unrelated to Python 3.14 load failures." }, { "left": "issue:41628", "right": "issue:42548", "accept": false, - "reason": "Two different top-level import errors for different symbols; same broad area, not the same bug." + "reason": "Different top-level import errors for different symbols; likely separate missing-export problems." }, { - "left": "issue:44908", - "right": "issue:44961", + "left": "issue:46143", + "right": "issue:46144", "accept": false, - "reason": "Scheduler kwargs handling is unrelated to the unrelated title." + "reason": "Both are about RoFormer, but one is kwargs passthrough and the other is attention-interface usage; related area, not the same bug." }, { - "left": "issue:44485", + "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation and an unrelated issue are not duplicates." + "reason": "Scheduler kwarg handling is unrelated to the other issue." }, { - "left": "issue:46143", - "right": "issue:46144", + "left": "issue:44485", + "right": "issue:44961", "accept": false, - "reason": "Both are RoFormer-related, but one is missing kwargs passthrough and the other is attention-interface usage; distinct change requests." + "reason": "A GLM-5 RoPE implementation discussion is not the same as the other issue." }, { "left": "issue:36296", "right": "issue:43901", "accept": false, - "reason": "Tensor-parallel training bug and pipeline docs mentioning return_all_scores are unrelated." + "reason": "Tensor-parallel training and TextClassificationPipeline docs are unrelated." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor parallel training and BeitImageProcessorFast reduce_labels behavior are unrelated." + "reason": "Tensor-parallel training does not match a BeitImageProcessorFast reduce_labels bug." }, { "left": "issue:44315", "right": "issue:46209", "accept": false, - "reason": "Liger Kernel with model_init and FP8 inference on B200 are different problems." + "reason": "Liger Kernel model_init behavior and FP8 inference on B200 are different model/runtime problems." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Missing GenerationMixin cache in chunked_prefill and a typo in cache strategies are not the same issue." + "reason": "A cache-missing runtime bug is unrelated to a typo in cache strategy naming." }, { "left": "issue:44623", "right": "issue:45230", "accept": false, - "reason": "Missing files in processor.save_pretrained is not enough to merge with a generic bug report." + "reason": "Missing files from processor.save_pretrained is not the same as an unspecified bug report." }, { "left": "issue:45362", "right": "issue:45850", "accept": false, - "reason": "A Qwen3.5-35B crash and a broad complaint that v5 breaks many models are not the same concrete bug." + "reason": "A Qwen3.5 chat crash may fit the broader v5-breakage theme, but it is a distinct model-specific failure and not the same underlying bug." }, { "left": "issue:44273", "right": "issue:45542", "accept": false, - "reason": "Lazy loading issues and a TensorBoard-without-TensorFlow backend error are unrelated." + "reason": "Lazy loading issues are unrelated to TensorBoard/TensorFlow backend setup problems." }, { "left": "issue:44038", "right": "issue:45125", "accept": false, - "reason": "A general Qwen3-VL-Moe bug report and a missing _tp_plan for tensor parallelism are related only at a very broad model-family level." + "reason": "Both touch Qwen/MoE/tensor parallelism, but they describe different model families and different concrete failures." }, { "left": "issue:44230", "right": "issue:46095", "accept": false, - "reason": "FP8 inference support for Qwen models and a DeepSeek weight-conversion question are unrelated." + "reason": "FP8 inference support and a weight-conversion script request are different changes." }, { "left": "issue:44393", "right": "issue:45071", "accept": false, - "reason": "Qwen3-VL bbox hallucination/error and PretrainedConfig type checking are unrelated." + "reason": "Qwen3-VL bounding-box output issues are unrelated to PretrainedConfig type checking." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading problems and unsafe sys.modules access are different code paths." + "reason": "Lazy loading and unsafe sys.modules access are distinct bugs." }, { "left": "issue:43576", "right": "issue:44556", "accept": false, - "reason": "Broken transformers env CLI and checkpoint reload failures after upgrading versions are both version regressions, but not the same defect." + "reason": "The env CLI issue is unrelated to checkpoint reload failures after upgrading to v5.2/v5.3." } ] }, @@ -16277,226 +16336,226 @@ "issue:43704|issue:44485", "issue:44485|issue:45468" ], - "prepared_review_unit_hash": "ed8f3f4b94cfb254d675fe1a804f93ce1af499b4b12e23752df1a0753fd2eaee", + "prepared_review_unit_hash": "1273ea2cfb737f9561a9b2f07e8de97686a4861819c5a0b28baa61b6c8eaf09f", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 43, - "serialized_chars": 39861, - "estimated_input_tokens": 9966, - "estimated_eval_tokens": 20188 + "serialized_chars": 39866, + "estimated_input_tokens": 9967, + "estimated_eval_tokens": 20190 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is dominated by unrelated issue reports that only loosely share a Transformers-v5 regression theme. I would not merge any soft-edge pair; the safest representative issue is the broad but concrete v5 import-regression report.", - "confidence": 0.69, - "canonical_issue_reason": "issue:45278 is the clearest broad regression report here: it describes widespread import failures after upgrading to Transformers 5.5.0, which matches the recurring post-v5 breakage pattern better than the vaguer umbrella complaints.", + "summary": "Mostly noisy cluster of unrelated Transformers issues. There are at least two believable duplicate pairs: 44556/45092 around v5 checkpoint-loading/meta-init incompatibility, and the stronger exact-title pair 45310/45478 about the Qwen3.5 MoE from_pretrained failure. The rest are only loosely theme-related.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45310 is the strongest canonical duplicate anchor because 45478 has the exact same title and they describe the same Qwen3.5 MoE from_pretrained failure; 44556/45092 is a separate duplicate-like pair, not the same one.", "canonical_pr_reason": null, - "best_issue_reason": "Best single anchor for the cluster because it is concrete, versioned, and broad enough to represent the repeated model/import loading regressions without being as vague as the generic 'breaks too many models' report.", + "best_issue_reason": "issue:44556 is the best overall cluster representative because it is a concrete, user-facing v5 loading/reload regression and is more descriptive than the broader noisy issues.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44393", "right": "issue:45753", "accept": false, - "reason": "Same model family, but different symptoms and likely different bugs: bbox-output hallucination vs infinite loop on a specific image." + "reason": "Same model family, but bbox-output hallucination and an infinite-loop/image-specific failure are different bugs." }, { "left": "issue:41669", "right": "issue:43673", "accept": false, - "reason": "Different problems: import-time `*` usage slowdown vs missing GenerationMixin cache in chunked prefill." + "reason": "Import-time slowdown from wildcard imports is unrelated to missing cache state in chunked prefill." }, { "left": "issue:43525", "right": "issue:45850", "accept": false, - "reason": "Specific missing `pad_token_id` attribute on Llama4Config is not the same concrete bug as a generic v5 breakage complaint." + "reason": "Missing pad_token_id in Llama4Config is specific and concrete; the other is a vague broad-breakage complaint." }, { "left": "issue:44556", "right": "issue:45092", - "accept": false, - "reason": "Both are checkpoint-loading regressions, but one is a general reload failure and the other is a model-specific remote-code/meta-init incompatibility." + "accept": true, + "reason": "Both point to the same v5 checkpoint-loading/meta-init compatibility regression and could plausibly be fixed together." }, { "left": "issue:43572", "right": "issue:45850", "accept": false, - "reason": "Specific StableLmConfig missing `pad_token_idx` is not the same underlying issue as the broad v5 model-breakage report." + "reason": "Model-specific pad_token_idx regression is not the same bug as a broad 'many models break' report." }, { "left": "issue:44492", "right": "issue:45850", "accept": false, - "reason": "A cache-strategy typo is unrelated to a broad model-loading regression." + "reason": "A typo in cache-strategy text is unrelated to broad model-loading breakage." }, { "left": "issue:45278", "right": "issue:45850", "accept": false, - "reason": "Both mention v5 regressions, but one is import failures and the other is a vague umbrella complaint; not the same concrete bug." + "reason": "Both are upgrade-related complaints, but one is import errors and the other is a vague generalization; not the same concrete defect." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API guidance and cache-strategy typo are unrelated." + "reason": "TF32 API settings and a cache-strategy typo are unrelated." }, { "left": "issue:41669", "right": "issue:46218", "accept": false, - "reason": "Import-star cleanup slowdown is unrelated to a missing `Qwen3_5ForCausalLM` import." + "reason": "Import slowdown from wildcard imports is not the same as a Qwen3_5 import failure." }, { "left": "issue:43408", "right": "issue:44315", "accept": false, - "reason": "Different subsystems and symptoms: model-type mismatch warning vs Liger Kernel not applying in `model_init`." + "reason": "Type-mismatch warning for sam3 models and Liger Kernel model_init behavior are different code paths." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Test flakiness from missing seeds is unrelated to scheduler kwargs being ignored." + "reason": "Missing set_seed in tests and inverse_sqrt scheduler kwargs handling are unrelated." }, { "left": "issue:44297", "right": "issue:45230", "accept": false, - "reason": "Tokenizer class mismatch in save_pretrained is not the same as an unspecified bug report." + "reason": "Tokenizer metadata mismatch is not the same as a generic placeholder bug report." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "Position-ids buffer access error is unrelated to a generic bug report." + "reason": "position_ids buffer error and a placeholder bug report are unrelated." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Both concern import/lazy-loading behavior, but the titles do not indicate the same concrete failure path." + "reason": "Both touch loading, but import slowness and lazy-loading failure are not the same concrete bug." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Unsupported flash-attention-2 for Nemotron is unrelated to Qwen3-VL output errors." + "reason": "Flash-Attention support gating and Qwen3-VL bbox hallucination are unrelated." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "Broken `transformers env` command is unrelated to inverse-sqrt scheduler kwargs handling." + "reason": "Broken env CLI and scheduler kwargs handling are different issues." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python-version compatibility issue is unrelated to a Qwen3-VL bounding-box bug." + "reason": "Python version support mismatch is unrelated to Qwen3-VL bbox output errors." }, { "left": "issue:44908", "right": "issue:45071", "accept": false, - "reason": "Scheduler kwarg handling and `PretrainedConfig` type checking are different bugs." + "reason": "Scheduler kwargs being ignored and PretrainedConfig type checking are different bugs." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "`AutoProcessor.from_pretrained` kwarg forwarding is unrelated to scheduler behavior." + "reason": "Both involve kwargs, but they affect different APIs and failure modes." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-star performance cleanup is unrelated to `HybridCache` import failure." + "reason": "Import-time slowdown and missing HybridCache import are not the same issue." }, { "left": "issue:44492", "right": "issue:46328", "accept": false, - "reason": "Cache-strategy typo and `SequenceFeatureExtractor.pad` inefficiency are unrelated." + "reason": "Cache-strategy typo and a SequenceFeatureExtractor.pad performance regression are unrelated." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-star cleanup is unrelated to a testing_utils bug." + "reason": "Import slowdown from wildcard imports has nothing to do with a testing_utils bug." }, { "left": "issue:43976", "right": "issue:45850", "accept": false, - "reason": "Python 3.9 compatibility is not the same concrete problem as broad v5 model breakage." + "reason": "Python compatibility issues are not the same as broad model-breakage complaints." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Different models and different failure modes: Qwen3-VL bbox issue vs Qwen3.5 MoE `from_pretrained` error." + "reason": "Different symptoms in the same family: bbox output issue versus Qwen3.5 MoE from_pretrained failure." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "PatchTSMixer post-init validation and tied-weight-key AttributeError are unrelated code paths." + "reason": "Optional post_init validation and tied-weights key handling are separate bugs." }, { "left": "issue:43976", "right": "issue:46218", "accept": false, - "reason": "Python-version support issue is unrelated to a missing Qwen3_5 import." + "reason": "Python support mismatch is unrelated to a specific Qwen3_5 import error." }, { "left": "issue:43827", "right": "issue:44908", "accept": false, - "reason": "Docs still referencing pipeline() after removals is unrelated to scheduler kwargs." + "reason": "Docs still referencing pipeline() is unrelated to scheduler kwargs handling." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion and tokenizer-loading regression are unrelated." + "reason": "Forward-doc confusion and tokenizer loading regression are unrelated." }, { "left": "issue:44315", "right": "issue:44829", "accept": false, - "reason": "Liger Kernel not applying in `model_init` is unrelated to flash-attention-3 degenerate training." + "reason": "Liger Kernel application with model_init and flash_attention_3 training degeneration are different problems." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Broken env command and lazy-loading issues are different problems." + "reason": "Broken env CLI and lazy loading not working are too broad to call the same bug." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "A placeholder bug report is not the same as the Qwen3.5 MoE `from_pretrained` failure." + "reason": "A generic placeholder bug report is not the same as the Qwen3.5 MoE from_pretrained error." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Pipeline save_pretrained failure and tied-weight-key AttributeError do not share a concrete code path." + "reason": "Pipeline save_pretrained failure and _get_tied_weight_keys AttributeError are unrelated." }, { "left": "issue:44360", @@ -16508,68 +16567,71 @@ "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading malfunction and inverse-sqrt scheduler kwargs are unrelated." + "reason": "Lazy loading regressions and scheduler kwargs handling are different areas." }, { "left": "issue:43576", "right": "issue:46218", "accept": false, - "reason": "Env command breakage and Qwen3_5 import failure are different issues." + "reason": "Broken env CLI and Qwen3_5 import failure are different bugs." }, { "left": "issue:46018", "right": "issue:46172", "accept": false, - "reason": "Both are DeepSeek-V4 related, but shared-expert gating and a transpose bug are different concrete defects." + "reason": "DeepSeek-V4 shared-expert gating and a tensor transpose issue are not the same concrete defect." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are unrelated." + "reason": "Timestamp calculation in a processor is unrelated to GLM-5 RoPE implementation." }, { "left": "issue:43519", "right": "issue:44908", "accept": false, - "reason": "Timestamp computation in a processor is unrelated to scheduler kwargs." + "reason": "Timestamp calculation and scheduler kwargs handling are unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-star cleanup is unrelated to a multi-thread VRAM leak." + "reason": "Import slowdown and VRAM leak in dataloader threads are different bugs." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Documentation confusion is unrelated to a model-loading regression for Llama 3.2 Vision." + "reason": "Forward-doc confusion is unrelated to a Llama model-loading breakage." }, { "left": "issue:31515", "right": "issue:31795", "accept": false, - "reason": "Slow checkpoint loading and forward-doc confusion are different issue types." + "reason": "Slow checkpoint loading and confusing forward documentation are not the same issue." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leak in dataloader threads and GLM-5 RoPE behavior are unrelated." + "reason": "VRAM leak in Qwen3ForCausalLM and GLM-5 RoPE implementation are unrelated." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "Different models and different implementation areas; not the same bug." + "reason": "Different model families and different implementation areas; not a duplicate." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, - "accepted_nontrivial_soft_edge": false, + "evaluator_result": { + "accept": false, + "feedback": "The summary is a bit too strong for the packet. In particular, the cluster is not limited to a single believable duplicate pair: issue:45310 and issue:45478 have the same title and are a much more obvious duplicate-like match than the narrative suggests. Because of that, the claim that 44556/45092 is the only believable soft edge is not fully grounded, and the canonical-anchor reasoning is not conservative enough." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -16670,268 +16732,268 @@ "issue:37934|issue:43668", "issue:42936|issue:43668" ], - "prepared_review_unit_hash": "b93ef340566e52613028540886377d34423786504b321ba8887e6146fa56663b", + "prepared_review_unit_hash": "39f46537d12dbf654b70087c4ffe56c7dc02cf62eeee273b3d521a7b6765b4d3", "component_budget": { "node_count": 451, "item_count": 451, "soft_pair_count": 1666, - "serialized_chars": 862631, - "estimated_input_tokens": 215658, - "estimated_eval_tokens": 431572 + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 }, "budget": { "node_count": 42, "item_count": 42, "soft_pair_count": 39, - "serialized_chars": 35128, - "estimated_input_tokens": 8782, - "estimated_eval_tokens": 17820 + "serialized_chars": 35120, + "estimated_input_tokens": 8780, + "estimated_eval_tokens": 17816 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous. A small subset touches tokenizer/v5 regressions, but most pairs are different bugs in different code paths, so I would not merge any soft edges.", - "confidence": 0.37, - "canonical_issue_reason": "Issue #42913 is the clearest representative of the recurring tokenizer/v5 regression theme and is more concrete than the broader umbrella reports.", - "canonical_pr_reason": null, - "best_issue_reason": "#42913 is the best single issue to anchor the cluster, but only weakly: it matches the most plausible duplicate subgroup while the rest of the items span unrelated failures.", - "best_pr_reason": null, + "summary": "The items are mostly unrelated issues that only share superficial wording. None of the soft pairs look like the same underlying bug or change, so no duplicate merges are justified.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue: this cluster is an overbroad lexical grouping spanning unrelated bugs, regressions, and feature requests.", + "canonical_pr_reason": "No PRs are present in the cluster.", + "best_issue_reason": "No global best issue stands out; none of these issues can safely represent the cluster as a whole.", + "best_pr_reason": "No PRs are present in the cluster.", "soft_edge_verdicts": [ { "left": "issue:43298", "right": "issue:46095", "accept": false, - "reason": "Different problems: continuous batching support for audio models vs. DeepSeek-V4 weight conversion/readability." + "reason": "Different problems: continuous batching support for audio models vs a DeepSeek-V4 weight conversion question." }, { "left": "issue:45850", "right": "issue:46218", "accept": false, - "reason": "Broad v5 breakage report vs. an import failure for a specific Qwen3.5 module." + "reason": "A vague version-wide breakage report is not the same as a specific import failure for Qwen3.5." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 API guidance is unrelated to a Qwen3ForCausalLM VRAM leak." + "reason": "TF32 API guidance and a VRAM leak in Qwen3ForCausalLM are unrelated." }, { "left": "issue:43502", "right": "issue:43519", "accept": false, - "reason": "local_files_only network violation vs. Qwen3VL timestamp calculation bug; different code paths." + "reason": "One is unwanted network requests with local_files_only; the other is an incorrect timestamp calculation." }, { "left": "issue:43519", "right": "issue:44623", "accept": false, - "reason": "Processor timestamp math and missing save_pretrained files are separate issues." + "reason": "Timestamp logic in Qwen3VL is unrelated to missing files in processor.save_pretrained." }, { "left": "issue:43519", "right": "issue:43704", "accept": false, - "reason": "Unrelated processor bug vs. dataloader-thread VRAM leak." + "reason": "These describe different failures: timestamp math vs VRAM leakage across dataloader threads." }, { "left": "issue:44263", "right": "issue:45468", "accept": false, - "reason": "torch.split return handling is not the same as Gemma4AudioRelPositionalEncoding." + "reason": "A torch.split return-value issue is not the same as Gemma4 audio positional encoding behavior." }, { "left": "issue:43606", "right": "issue:45468", "accept": false, - "reason": "CPU offload device mismatch vs. Gemma-4 audio positional encoding issue." + "reason": "CPU offload device mismatch and audio positional encoding are different code paths and bugs." }, { "left": "issue:44261", "right": "issue:45468", "accept": false, - "reason": "MLA q_a_layernorm epsilon/config precision issue is distinct from Gemma-4 audio positional encoding." + "reason": "RMS norm epsilon precision discussion does not match Gemma4AudioRelPositionalEncoding." }, { "left": "issue:44351", "right": "issue:46218", "accept": false, - "reason": "Importing HybridCache and importing Qwen3_5ForCausalLM are different missing-symbol problems." + "reason": "HybridCache import failure and Qwen3_5ForCausalLM import failure are distinct module issues." }, { "left": "issue:43352", "right": "issue:46218", "accept": false, - "reason": "Flash Attention 2.0 support gap is unrelated to Qwen3.5 module import failure." + "reason": "Flash Attention 2 support for Nemotron and a Qwen3.5 import error are unrelated." }, { "left": "issue:42913", "right": "issue:43066", "accept": false, - "reason": "Both are tokenizer/v5-related, but one reports behavior drift and the other a wrong decoder type; not the same concrete bug." + "reason": "Both involve tokenizer changes, but they are different concrete regressions: behavior drift vs decoder type selection." }, { "left": "issue:44908", "right": "issue:44998", "accept": false, - "reason": "Scheduler kwargs handling is unrelated to the unrelated 'Unemployment' issue." + "reason": "Scheduler keyword handling is unrelated to the unrelated 'Unemployment' placeholder issue." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "Cache strategy typo does not match the unrelated placeholder/title mismatch." + "reason": "A cache-strategy typo is not the same as the unrelated 'Unemployment' issue." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the stray title." + "reason": "CLIPOutput attentions regression and the unrelated 'Unemployment' issue do not match." }, { "left": "issue:44998", "right": "issue:45850", "accept": false, - "reason": "The generic 'breaks too many models' complaint is not the same underlying bug as the placeholder issue." + "reason": "A nonsensical placeholder issue is not the same as a broad 5.0 compatibility complaint." }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "_tied_weights_keys AttributeError is unrelated to chat template crashes on tool-call messages." + "reason": "Tied-weight key handling and chat template crashes are different failures." }, { "left": "issue:44938", "right": "issue:44998", "accept": false, - "reason": "Python 3.14 import failure is unrelated to the stray title." + "reason": "Python 3.14 import failure is unrelated to the 'Unemployment' issue." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id on Llama4Config is not the same failure as _get_tied_weight_keys on list values." + "reason": "Missing pad_token_id in Llama4Config is distinct from _tied_weight_keys list handling." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "StableLmConfig pad_token_idx regression is a different config bug than tied-weight-key handling." + "reason": "StableLmConfig pad_token_idx regression is unrelated to tied-weight key crashes." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow checkpoint loading is unrelated to the missing layer_type_validation import." + "reason": "Slow from_pretrained checkpoint loading and a missing layer_type_validation import are different bugs." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 inference support request is different from processor.save_pretrained missing files." + "reason": "FP8 inference support for Qwen3-VL/Qwen3.5 MoE is not the same as processor.save_pretrained file loss." }, { "left": "issue:30333", "right": "issue:31515", "accept": false, - "reason": "MLflow reporting on failed training jobs is unrelated to slow from_pretrained checkpoint loading." + "reason": "MLFlow reporting on failed training jobs is unrelated to slow checkpoint loading." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint load slowness and tensor-parallel training bug are different problems." + "reason": "A slow checkpoint load is not the same as a tensor parallel training bug." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "Generic FP8 support request is not the same as a Qwen3.5 MoE from_pretrained error." + "reason": "FP8 support work and a Qwen3.5 MoE from_pretrained error are different concrete changes/failures." }, { "left": "issue:31515", "right": "issue:33453", "accept": false, - "reason": "from_pretrained slowness is distinct from tokenizer loading regression." + "reason": "Checkpoint loading performance and tokenizer loading regression are separate problems." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching for audio models and FP8 inference support are unrelated feature/problem reports." + "reason": "Continuous batching support for audio models is unrelated to FP8 inference support." }, { "left": "issue:43717", "right": "issue:46144", "accept": false, - "reason": "Mamba init_weights usage and RoFormer attention interface usage are different model implementation issues." + "reason": "Mamba init_weights usage and RoFormer attention-interface usage are different model implementation issues." }, { "left": "issue:42936", "right": "issue:46018", "accept": false, - "reason": "Mask2former ignore_value bug is unrelated to DeepSeek-V4 shared expert gating." + "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 on ModernBERTConfig.norm_eps is not the same as tokenizer regex warning behavior." + "reason": "A ModernBERT type-hint bug is unrelated to tokenizer regex warnings." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2former ignore_value and loss kwargs propagation are unrelated bugs." + "reason": "Mask2Former ignore_value and loss-kwargs forwarding are different code paths." }, { "left": "issue:43240", "right": "issue:46144", "accept": false, - "reason": "Loss kwargs not forwarded and RoFormer attention interface are different code paths." + "reason": "Loss kwargs forwarding and RoFormer attention-interface usage are unrelated bugs." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Mask2former ignore_value bug is unrelated to tokenizer regex warnings." + "reason": "Mask2Former ignore_value and tokenizer regex warnings are not the same issue." }, { "left": "issue:42936", "right": "issue:46144", "accept": false, - "reason": "Mask2former loss handling and RoFormer attention implementation are unrelated." + "reason": "Mask2Former ignore_value handling does not overlap with RoFormer attention implementation." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2former ignore_value and a typo in quantization operation text are unrelated." + "reason": "Ignore_value usage and a quantization_operation typo are unrelated." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Loss kwargs forwarding is unrelated to a config type-hint typo." + "reason": "Missing loss kwargs and ModernBERT norm_eps typing are unrelated." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "ModernBERTConfig type hint issue and quantization typo are different bugs." + "reason": "A type-hint issue and a typo in quantization_operation are different problems." }, { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Llama4TextL2Norm norm question is unrelated to ModernBERTConfig.norm_eps typing." + "reason": "Llama4TextL2Norm/RMS norm discussion is not the same as ModernBERT norm_eps typing." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Mask2former ignore_value and ModernBERTConfig type hint are unrelated." + "reason": "Mask2Former ignore_value and ModernBERT norm_eps typing are unrelated." } ] }, @@ -16964,7 +17026,7 @@ "issue:10519|issue:11005", "issue:11005|issue:11307" ], - "prepared_review_unit_hash": "c47c25a0f8bb48ed0e9261293059f5ab72420ac5aaf305b52b407f4ae1311685", + "prepared_review_unit_hash": "477f17e52149011010c89fc0015c0327625f6fafcc2fab8ae7e0a9817cf5a0d4", "component_budget": { "node_count": 3, "item_count": 3, @@ -16986,30 +17048,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three issues are not duplicates: they cover unrelated topics (token truncation behavior, a ReduceLROnPlateau-like scheduler feature request, and Wav2Vec2 word time offsets). The similarity is superficial and based on generic wording rather than the same underlying problem.", - "confidence": 0.98, - "canonical_issue_reason": "Issue 10519 is the most concrete and narrowly scoped request in the set, so it is the best representative if one issue must be chosen.", + "summary": "The three issues are unrelated feature requests in different areas: token truncation behavior, a ReduceLROnPlateau-like scheduler feature, and Wav2Vec2 word boundary offsets. This does not look like a duplicate cluster.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "10519 is the clearest standalone user-facing feature request with a specific behavior change and is less ambiguous than the other two.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:10519", "right": "issue:11307", "accept": false, - "reason": "Different problems in different model components: truncation strategy vs extracting word-level timestamps from Wav2Vec2. No shared underlying bug or change." + "reason": "Different topics and code paths: token truncation direction vs. speech model word offset extraction. No evidence of the same underlying bug or change." }, { "left": "issue:10519", "right": "issue:11005", "accept": false, - "reason": "One asks for truncation-from-beginning behavior; the other asks for ReduceLROnPlateau-like scheduler functionality. These are unrelated feature requests." + "reason": "One is about truncation strategy in tokenization; the other asks for learning-rate scheduler functionality. They are unrelated feature requests." }, { "left": "issue:11005", "right": "issue:11307", "accept": false, - "reason": "Scheduler behavior request and Wav2Vec2 timing offsets are unrelated; they do not describe the same code-path issue or mergeable change." + "reason": "Scheduler behavior request vs. Wav2Vec2 time-offset extraction. No shared concrete bug or mergeable code-path change." } ] }, @@ -17041,12 +17103,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "d93c640c87d3fd724c96a934cfa2fa3bce00241672ccd17362ef36b029ae1c2c", + "prepared_review_unit_hash": "749c4db89dabf7abc5e75e3519ce18024c184de211f12d9bbc07dec6e6ecba2a", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2478, + "serialized_chars": 2477, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -17054,7 +17116,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2478, + "serialized_chars": 2477, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -17063,24 +17125,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These three issues do not appear to be duplicates: one is about TAPAS tokenization mismatching TensorFlow, one about a misleading auth/error path for private model downloads, and one about failing to load bert-base-uncased. They share only broad model/loading language, not the same underlying bug.", - "confidence": 0.92, - "canonical_issue_reason": "Issue 16618 is the broadest and most user-facing of the set, with the highest discussion activity; if one issue must represent the cluster, it is the closest fit. That said, the cluster is heterogeneous rather than truly duplicate.", + "summary": "This does not look like a single duplicate bug cluster. 16618 is the broadest model-loading issue, but 15980 is about private-model login/error messaging and 13244 is about Tapas tokenization behavior, so the three are likely separate problems.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 16618 is the best cluster representative because it is the most general model-loading complaint; the other issues are narrower and point to different failure modes.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 16618 is the best global issue candidate because it concerns a general model-loading failure and drew the most discussion, making it the most prominent representative of the group.", + "best_issue_reason": "16618 is the best issue to anchor the cluster only because it is the closest to a generic loading problem, but it is not a strong duplicate of the others.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Different problem domains: TAPAS tokenization mismatch versus a bert-base-uncased model loading failure. Similarity is only superficial around model behavior." + "reason": "Tokenization behavior for Tapas and failing to load bert-base-uncased are different code paths and different bugs." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "Auth/download error message for private models is a different code path from failing to load bert-base-uncased. Not the same underlying bug." + "reason": "Private-model login/error-message handling is not the same underlying issue as a generic public model loading failure." } ] }, @@ -17110,7 +17172,7 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "1b4d2bc07e2631661f01468e6f7f3f91429e2a72f712f2f6f9bb216e981b651c", + "prepared_review_unit_hash": "5f4ee6f66f7a79ba543d2007bc735686053324d94ac9acc8beaf7f6c9d02aa7b", "component_budget": { "node_count": 2, "item_count": 2, @@ -17132,18 +17194,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both community contribution requests about expanding model-ecosystem support, but they target different features: Flash Attention 2 versus GGUF. They are related in theme, not duplicates of the same underlying change.", + "summary": "These are two unrelated community contribution issues: one for Flash Attention 2 support on more architectures and one for GGUF support on more architectures. They share a similar framing but not the same underlying feature request.", "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the two issues request different capabilities and should remain separate.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 26350 is the more established thread with much higher discussion activity, so it is the better representative if one must be chosen.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Different feature requests for different support paths (Flash Attention 2 vs GGUF); they are not the same underlying bug or change." + "reason": "Both are broad community contribution requests, but they target different features and code paths (Flash Attention 2 vs GGUF support). Similar wording is not enough to treat them as duplicates." } ] }, @@ -17173,7 +17235,7 @@ "soft_pairs": [ "issue:26413|issue:31501" ], - "prepared_review_unit_hash": "48396163bcd742dd9da7443fbed466dd4c0325d3f591c6d16e281471cff23b84", + "prepared_review_unit_hash": "81e6c243e340ee80a42f56d93a6a91f81bfa9f913bb122fb640baa53c4b69f3b", "component_budget": { "node_count": 2, "item_count": 2, @@ -17195,18 +17257,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues do not look like the same underlying bug. One is a specific checkpoint-resume failure with a clear error message; the other is a broad report that FineWeb SLM training does not start, with no explicit evidence it stems from the same code path.", - "confidence": 0.71, - "canonical_issue_reason": "issue:26413 is the more concrete and actionable report: it names the failing API (`resume_from_checkpoint`) and the exact error symptom, making it a better anchor for a duplicate cluster if one were needed.", + "summary": "The two issues both concern training failures, but they describe different symptoms and likely different root causes: one is a specific `resume_from_checkpoint` epoch-iterator error, the other is a broader report that FineWeb SLM training does not start. I would not deduplicate them.", + "confidence": 0.92, + "canonical_issue_reason": "Issue #26413 is the more specific and technically concrete report, with an explicit failure message and a clearly scoped code path (`resume_from_checkpoint`).", "canonical_pr_reason": null, - "best_issue_reason": "issue:26413 is the best issue candidate because it describes a specific reproducible failure rather than a generic training startup problem.", + "best_issue_reason": "#26413 is the best representative because it names the exact exception and failure point, making it more actionable as the canonical issue for triage.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26413", "right": "issue:31501", "accept": false, - "reason": "The titles point to different problem scopes: a checkpoint-resume iterator failure versus a general training startup issue. There is not enough evidence that they share the same concrete bug or code path." + "reason": "The reports are too different to treat as the same bug: #26413 is a checkpoint-resume iterator crash, while #31501 is a generic startup/training-init failure. Same broad area, but not the same underlying issue." } ] }, @@ -17265,187 +17327,187 @@ "issue:44370|issue:46226", "issue:46222|issue:46226" ], - "prepared_review_unit_hash": "35a514ee1ea286beebbd980a43b6bf34239b49650d475ca28cb44a8a76c8311f", + "prepared_review_unit_hash": "970d8ddb664c0143ff2b746ee7f92a0d1ee54aed40a6841a0b248e525d6556ae", "component_budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 24, - "serialized_chars": 13527, - "estimated_input_tokens": 3382, - "estimated_eval_tokens": 7020 + "serialized_chars": 13531, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 24, - "serialized_chars": 13527, - "estimated_input_tokens": 3382, - "estimated_eval_tokens": 7020 + "serialized_chars": 13531, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly a set of i18n translation-request issues, but it actually contains separate language requests (Turkish, Spanish, Romanian) plus two small duplicate groups: the three generic placeholder-title issues, and the identical Thai-titled pair. The high similarity is mostly template overlap, not the same underlying request.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 43834 is the earliest clean representative of the near-identical generic-template translation-request subgroup, and it is the best candidate to stand in for that duplicate set.", + "summary": "This is a mixed i18n cluster: 27088, 28936, and 38435 are clearly separate language-labeled translation requests; 43834/44365/44370 form a highly similar placeholder-title subgroup that may be duplicates; and 46222/46226 are an exact duplicate pair. The safest representative is 43834, but the cluster is not one single duplicate set.", + "confidence": 0.9, + "canonical_issue_reason": "43834 is the best representative of the dense placeholder-title subgroup and sits near the center of the cluster, but it should be treated conservatively because the cluster also contains separate language-specific requests and an exact Thai duplicate pair.", "canonical_pr_reason": null, - "best_issue_reason": "43834 is the clearest representative of the duplicate subgroup; the concrete Turkish/Spanish/Romanian requests are distinct and should not be merged into it.", + "best_issue_reason": "43834 is the strongest global issue candidate because it belongs to the most internally consistent near-duplicate subgroup and is more representative than the exact two-item Thai pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different language request; similarity is from the shared translation template only." + "reason": "Different language-specific translation requests; only the broad docs/i18n theme overlaps." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different language request; not the same underlying issue." + "reason": "Different language-specific translation requests; only the broad docs/i18n theme overlaps." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Turkish request vs generic placeholder template; no evidence they are the same task." + "reason": "Turkish-specific request vs a generic placeholder-title issue; not enough evidence of the same report." }, { "left": "issue:27088", "right": "issue:38435", "accept": false, - "reason": "Different target language (Turkish vs Romanian); separate requests." + "reason": "Turkish and Romanian translation requests are distinct issues." }, { "left": "issue:27088", "right": "issue:28936", "accept": false, - "reason": "Different target language (Turkish vs Spanish); separate requests." + "reason": "Turkish and Spanish translation requests are distinct issues." }, { "left": "issue:28936", "right": "issue:43834", "accept": false, - "reason": "Spanish request vs generic placeholder template; template similarity is not enough." + "reason": "Spanish-specific request vs a generic placeholder-title issue; not the same concrete issue." }, { "left": "issue:28936", "right": "issue:44365", "accept": false, - "reason": "Different target language request; same broad i18n theme only." + "reason": "Different translation requests; same broad i18n theme is not enough to merge." }, { "left": "issue:28936", "right": "issue:38435", "accept": false, - "reason": "Spanish and Romanian translation requests are distinct." + "reason": "Spanish and Romanian translation requests are distinct issues." }, { "left": "issue:28936", "right": "issue:44370", "accept": false, - "reason": "Different target language request; not the same issue." + "reason": "Different translation requests; same broad i18n theme is not enough to merge." }, { "left": "issue:38435", "right": "issue:44370", "accept": false, - "reason": "Romanian request vs generic placeholder template; not enough to merge." + "reason": "Romanian-specific request vs a generic placeholder-title issue; not enough evidence of the same report." }, { "left": "issue:38435", "right": "issue:44365", "accept": false, - "reason": "Romanian request vs generic placeholder template; distinct underlying request." + "reason": "Romanian-specific request vs a generic placeholder-title issue; not enough evidence of the same report." }, { "left": "issue:38435", "right": "issue:43834", "accept": false, - "reason": "Romanian request vs generic placeholder template; not the same issue." + "reason": "Romanian-specific request vs a generic placeholder-title issue; not enough evidence of the same report." }, { "left": "issue:38435", "right": "issue:46222", "accept": false, - "reason": "Romanian request and Thai-titled issue are unrelated." + "reason": "Different titles and no evidence they describe the same underlying report." }, { "left": "issue:38435", "right": "issue:46226", "accept": false, - "reason": "Romanian request and Thai-titled issue are unrelated." + "reason": "Different titles and no evidence they describe the same underlying report." }, { "left": "issue:43834", "right": "issue:44370", "accept": true, - "reason": "Same generic placeholder title/body pattern; looks like the same duplicated template request." + "reason": "Identical placeholder-style title and extremely high similarity; these look like the same underlying issue." }, { "left": "issue:43834", "right": "issue:44365", "accept": true, - "reason": "Near-identical generic-template issue; same underlying request." + "reason": "Identical placeholder-style title and extremely high similarity; these look like the same underlying issue." }, { "left": "issue:43834", "right": "issue:46222", "accept": false, - "reason": "Generic template issue vs a different Thai-titled issue; not the same request." + "reason": "Only broad docs-translation similarity is visible; the titles do not show the same concrete report." }, { "left": "issue:43834", "right": "issue:46226", "accept": false, - "reason": "Generic template issue vs a different Thai-titled issue; not the same request." + "reason": "Only broad docs-translation similarity is visible; the titles do not show the same concrete report." }, { "left": "issue:44365", "right": "issue:44370", "accept": true, - "reason": "Near-identical generic-template translation issue; same underlying request." + "reason": "Identical placeholder-style title and near-identical bodies; this is a strong duplicate match." }, { "left": "issue:44365", "right": "issue:46222", "accept": false, - "reason": "Generic template issue vs a different Thai-titled issue; not the same request." + "reason": "Only broad docs-translation similarity is visible; not enough to treat as the same issue." }, { "left": "issue:44365", "right": "issue:46226", "accept": false, - "reason": "Generic template issue vs a different Thai-titled issue; not the same request." + "reason": "Only broad docs-translation similarity is visible; not enough to treat as the same issue." }, { "left": "issue:44370", "right": "issue:46222", "accept": false, - "reason": "Generic template issue vs a different Thai-titled issue; not the same request." + "reason": "Only broad docs-translation similarity is visible; not enough to treat as the same issue." }, { "left": "issue:44370", "right": "issue:46226", "accept": false, - "reason": "Generic template issue vs a different Thai-titled issue; not the same request." + "reason": "Only broad docs-translation similarity is visible; not enough to treat as the same issue." }, { "left": "issue:46222", "right": "issue:46226", "accept": true, - "reason": "Identical title and near-identical metadata indicate a duplicate pair." + "reason": "Same exact title and identical report shape; this is a true duplicate pair." } ] }, "evaluator_result": { - "accept": true, - "feedback": "Grounded and conservative overall. The summary matches the packet: Turkish/Spanish/Romanian are distinct language requests, the three placeholder-template issues form one near-identical subgroup, and 46222/46226 is an identical Thai-titled pair. The soft-edge verdicts are appropriately cautious and do not overstate equivalence beyond the available evidence." + "accept": false, + "feedback": "The duplicate call on 46222/46226 is well grounded, but the cluster summary overreaches by describing the rest as distinct issues in different languages. The packet only directly supports language-specific separation for 27088/28936/38435; it does not support that claim for the identical placeholder-title trio 43834/44365/44370, which have extremely high similarity and could also be near-duplicates. Please make that distinction more conservative and avoid asserting different-language separation where the packet does not show it." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -17518,136 +17580,136 @@ "issue:43837|issue:45177", "issue:45322|issue:46133" ], - "prepared_review_unit_hash": "1eba4c595d12ba4ffc51353c132b10a76a664148de01bd592bb3a81c9fdaca41", + "prepared_review_unit_hash": "6d1d0b11348ba24ae1e66c649f1c97c32fef8659eb7ad8ee1796e0eb2ae87338", "component_budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 36, - "serialized_chars": 21657, - "estimated_input_tokens": 5415, - "estimated_eval_tokens": 11086 + "serialized_chars": 21662, + "estimated_input_tokens": 5416, + "estimated_eval_tokens": 11088 }, "budget": { "node_count": 16, "item_count": 16, "soft_pair_count": 36, - "serialized_chars": 21657, - "estimated_input_tokens": 5415, - "estimated_eval_tokens": 11086 + "serialized_chars": 21662, + "estimated_input_tokens": 5416, + "estimated_eval_tokens": 11088 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This is a loose thematic cluster of model-addition requests across different architectures and modalities, not a true duplicate set.", - "confidence": 0.94, - "canonical_issue_reason": "No single issue is a true duplicate hub; issue 43671 is only the closest representative because it is the most connected request in the cluster, but the items span different models and modalities.", + "summary": "This cluster is a loose collection of model-request issues that all say \u201cadd/support model X,\u201d but they describe different models and code changes, not the same underlying bug or feature. No PRs are present.", + "confidence": 0.95, + "canonical_issue_reason": "No true canonical duplicate exists here: the issues request different model additions/architectures and should remain separate.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43671 is the best single anchor for the cluster due to its high connectivity, but it should not be treated as a duplicate canonical for the others.", + "best_issue_reason": "Issue 45322 is the cleanest, most self-contained model-request example in the set, so it is the best representative if one issue must be chosen.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "EuroBert config vs DEIMv2 are different model requests and not the same change." + "reason": "Different model requests: EuroBert config vs DEIMv2. Same broad theme, not the same change." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBert config work is unrelated to Qwen3-TTS support." + "reason": "EuroBert config request is unrelated to Qwen3-TTS support." }, { "left": "issue:36817", "right": "issue:36879", "accept": false, - "reason": "EuroBert vs RF-DETR are different model families and unrelated updates." + "reason": "EuroBert and RF-DETR are unrelated model additions." }, { "left": "issue:36879", "right": "issue:41211", "accept": false, - "reason": "RF-DETR and DEIMv2 are distinct detection-model additions." + "reason": "RF-DETR and DEIMv2 are different models and different implementation paths." }, { "left": "issue:36879", "right": "issue:43909", "accept": false, - "reason": "Detection-model support is unrelated to adding LFM2.5 Audio." + "reason": "RF-DETR and LFM2.5 Audio are unrelated additions." }, { "left": "issue:36879", "right": "issue:44121", "accept": false, - "reason": "RF-DETR and a weight-sparse transformer are different code paths and model types." + "reason": "RF-DETR and OpenAI weight-sparse transformer are different model requests." }, { "left": "issue:36879", "right": "issue:43837", "accept": false, - "reason": "RF-DETR and Qwen3-ASR are unrelated model integrations." + "reason": "RF-DETR and Qwen3-ASR are different concrete models." }, { "left": "issue:36879", "right": "issue:43671", "accept": false, - "reason": "RF-DETR and Qwen3-TTS are different modalities and not the same underlying change." + "reason": "RF-DETR and Qwen3-TTS are unrelated model additions." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are unrelated model requests." + "reason": "NeoBERT and Qwen3-ASR are different model families and code changes." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT support and the sparse-transformer request are not the same fix." + "reason": "NeoBERT and circuitgpt/weight-sparse transformer are unrelated requests." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT is a text model request, while LFM2.5 Audio is an audio-model request." + "reason": "NeoBERT and LFM2.5 Audio are different models with different support needs." }, { "left": "issue:36879", "right": "issue:37015", "accept": false, - "reason": "RF-DETR and NeoBERT are different model families with no shared concrete change." + "reason": "RF-DETR and NeoBERT are unrelated model additions." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are both BERT-like, but they are distinct model additions and not one mergeable change." + "reason": "NeoBERT and ModernVBERT are distinct model requests, not duplicates." }, { "left": "issue:41115", "right": "issue:43837", "accept": false, - "reason": "MiniCPM3 architecture support is unrelated to Qwen3-ASR." + "reason": "MiniCPM3 architecture support is unrelated to Qwen3-ASR support." }, { "left": "issue:41115", "right": "issue:44121", "accept": false, - "reason": "MiniCPM3 support and the sparse-transformer request are different integration tasks." + "reason": "MiniCPM3 and circuit-sparse transformer are different concrete models." }, { "left": "issue:41115", "right": "issue:42503", "accept": false, - "reason": "MiniCPM3 and ModernVBERT are separate model requests, not the same bug or feature." + "reason": "MiniCPM3 and ModernVBERT are separate model additions." }, { "left": "issue:41115", "right": "issue:45027", "accept": false, - "reason": "MiniCPM3 and Voxtral TTS are different model integrations." + "reason": "MiniCPM3 and Voxtral-4B-TTS-2603 are different model integrations." }, { "left": "issue:41211", @@ -17659,37 +17721,37 @@ "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models and Qwen3-ASR target different model families." + "reason": "ModernVBERT and Qwen3-ASR are different models and code paths." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT models and Qwen3-TTS are unrelated additions." + "reason": "ModernVBERT and Qwen3-TTS are unrelated additions." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT is a specific model request; the RoPE issue is a broader architectural proposal, not the same change." + "reason": "ModernVBERT and generic BERT-like RoPE support are not the same concrete change." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like models with RoPE and Voxtral TTS have no concrete overlap." + "reason": "BERT-like models with RoPE and Voxtral TTS support are unrelated requests." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Qwen3-TTS and Qwen3-ASR are same family but different modalities and code paths, so they are not one duplicate change." + "reason": "Qwen3-TTS and Qwen3-ASR are related by name but are distinct speech models and separate changes." }, { "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Both are TTS-related, but they concern different models and likely separate integrations." + "reason": "Qwen3-TTS and Voxtral TTS are different model integrations." }, { "left": "issue:43671", @@ -17701,55 +17763,55 @@ "left": "issue:43671", "right": "issue:44121", "accept": false, - "reason": "TTS support and a sparse transformer request are unrelated." + "reason": "Qwen3-TTS and the weight-sparse transformer request are different model changes." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Both are audio-model requests, but they are different models with different implementation scope." + "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model support requests." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio are distinct audio-model integrations." + "reason": "Qwen3-ASR and LFM2.5 Audio are different model integrations." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "ASR support and EUPE are unrelated model requests." + "reason": "Qwen3-ASR and EUPE are unrelated model requests." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR and the sparse-transformer request are not the same underlying change." + "reason": "Qwen3-ASR and circuit-sparse transformer are different concrete changes." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio and SAM3-LiteText are different modalities and integration paths." + "reason": "LFM2.5 Audio and SAM3-LiteText are different model families." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "Weight-sparse transformer support and EUPE are unrelated." + "reason": "Weight-sparse transformer support and EUPE are unrelated model additions." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral TTS are different audio tasks and not one mergeable fix." + "reason": "Qwen3-ASR and Voxtral-4B-TTS-2603 are distinct speech-model requests." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral TTS are unrelated model additions." + "reason": "NeoBERT and Voxtral-4B-TTS-2603 are unrelated model support tasks." }, { "left": "issue:43837", @@ -17761,7 +17823,7 @@ "left": "issue:45322", "right": "issue:46133", "accept": false, - "reason": "EUPE and TIPSv2 are both vision-language related, but they are distinct model additions and not the same concrete change." + "reason": "EUPE and TIPSv2 are separate vision-language model requests from different authors and model families." } ] }, @@ -17791,7 +17853,7 @@ "soft_pairs": [ "issue:38740|issue:46443" ], - "prepared_review_unit_hash": "ea7b50799e223c0fdda5a65b624d92284f6908658581f066a92b62816f53f359", + "prepared_review_unit_hash": "e2a21758b76ea719607ee1ca2730a4c737b77ce94c670c6ba1e136a407248b14", "component_budget": { "node_count": 2, "item_count": 2, @@ -17813,18 +17875,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The pair is not a duplicate cluster: one item is a closed docs request about adding `pruna`, the other is an open Vision RoPE refactor. They do not describe the same underlying change or bug.", - "confidence": 0.98, - "canonical_issue_reason": "Issue 46443 is the better cluster anchor because it is a concrete open engineering task; issue 38740 is a separate closed docs request.", + "summary": "The two issues are unrelated: one is a docs request about adding `pruna`, the other is a Vision RoPE refactor. The soft similarity signal is not enough to treat them as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "46443 is the most substantive and active issue in the set, while 38740 is unrelated and documentation-focused.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:38740", - "right": "issue:46443", + "left": "issue:38740 ([DOCS] Add `pruna` as optimization framework)", + "right": "issue:46443 (Vision RoPE refactor)", "accept": false, - "reason": "Reject: the topics are unrelated (docs framework addition vs Vision RoPE refactor), with no shared targets/files or explicit linkage." + "reason": "Different topics and problem types: documentation/framework addition vs model architecture refactor; no shared code path or duplicate underlying issue." } ] }, @@ -17854,7 +17916,7 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "a5f155011f41c26ede4e86b0fc07c984bcc9ba3043f157af758336778a6c9bd0", + "prepared_review_unit_hash": "38168fa2b04c108ee3c9c3d0d24748561ce27798a199cceec46424fafedd67ca", "component_budget": { "node_count": 2, "item_count": 2, @@ -17876,18 +17938,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one is about block decoding, the other is about DeepSeek v3.2 model support.", + "summary": "The two issues are not duplicates: one is about block decoding, the other about DeepSeek v3.2 support. No PRs are present.", "confidence": 0.98, - "canonical_issue_reason": "No canonical issue: the titles and implied problem domains are unrelated, so they should not be clustered as duplicates.", + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a global best duplicate anchor; they describe different feature requests and should remain separate.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different underlying requests: block decoding vs DeepSeek v3.2 support. No evidence they describe the same bug or change." + "reason": "Unrelated topics and no shared concrete bug/change path; a decoding setting request is not the same as model-version support." } ] }, @@ -17917,12 +17979,12 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "2eb5b54e5b4e04fd5e96d13a74ff46f6f5b38f540a9be3e9ccc2f6d863e60850", + "prepared_review_unit_hash": "fe2e0c70170ddbcbf5d8d720ce0cb6cae7b63548a06a8d14c740cba5c8594740", "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 }, @@ -17930,7 +17992,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1584, + "serialized_chars": 1583, "estimated_input_tokens": 396, "estimated_eval_tokens": 1048 }, @@ -17939,18 +18001,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related only at a very broad level (Trainer extensibility), but they describe different problems and code paths: one about `remove_unused_column` behavior, the other about FLOPs estimation for custom trainers.", + "summary": "The two closed issues are related only by a broad theme of custom trainer compatibility, but they describe different problems: one is about `remove_unused_column` customizability in `Trainer`, the other about FLOPS estimation for custom trainers. They should not be deduplicated.", "confidence": 0.94, - "canonical_issue_reason": "Issue #43598 is the cleaner representative because it states a concrete Trainer customization concern; #43600 is a separate accounting/metrics issue.", + "canonical_issue_reason": "No single canonical issue: the topics overlap only at the level of custom trainer support, but the underlying bugs are different code paths and would need separate fixes.", "canonical_pr_reason": null, - "best_issue_reason": "#43598 is the better cluster anchor since it is more directly about a Trainer behavior change, while #43600 is a narrower FLOPs-estimation edge case.", + "best_issue_reason": "If a representative issue is needed, #43598 is the better anchor because it names a specific Trainer behavior and a concrete customization point, making the scope clearer than the FLOPS-accounting request.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "They both mention custom trainers, but one is about unused-column removal and the other is about FLOPs estimation. Different underlying bugs, different fixes, not mergeable as one issue." + "reason": "Different underlying issues: one targets `remove_unused_column` handling, the other FLOPS counting for custom trainers; same subsystem, not the same bug/change." } ] }, @@ -17980,12 +18042,12 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "96e31213173996d63de4a9bbc2eb283998fe6198519317daa6cdec06038532b3", + "prepared_review_unit_hash": "b7d62275c67c8a6ece832e4eb16a4182481eba65941907f289e78286224e8538", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1610, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -17993,7 +18055,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1610, + "serialized_chars": 1612, "estimated_input_tokens": 403, "estimated_eval_tokens": 1062 }, @@ -18002,18 +18064,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues share a broad Qwen-support theme, but they ask for different changes: Qwen3-Omni model registration vs. Qwen3.5 sequence-classification support. They are not duplicates.", - "confidence": 0.97, - "canonical_issue_reason": "Issue 43684 is the broader representative request in this pair, centered on model registration for Qwen3-Omni.", + "summary": "These two issues are both about adding missing model support, but they target different model families and different AutoModel heads, so they are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #43684 is the broader request, covering registration support for Qwen3-Omni in AutoModel and AutoModelForConditionalGeneration, so it is the better representative of the cluster.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43684 is the better cluster representative because it is the more general API-registration request; issue 44405 is a separate head-specific feature request.", + "best_issue_reason": "#43684 is the best issue to keep as the cluster representative because it describes the more general model-registration gap rather than a single task-specific head.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Different underlying changes: 43684 requests AutoModel/AutoModelForConditionalGeneration registration for Qwen3-Omni, while 44405 requests AutoModelForSequenceClassification support for Qwen3.5. Similar model family, but not the same bug or fix." + "reason": "Different underlying changes: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration versus AutoModelForSequenceClassification support for Qwen3.5. They involve different configs and different code paths, so they should not be merged as the same issue." } ] }, @@ -18050,77 +18112,80 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "c7f73a4d667a54579ca3f0e4f70fa388bbd1507ba3514dc0f1d7e77799c3fd1d", + "prepared_review_unit_hash": "d71da42388bf8ecba4b20ca1c37e5863f01a6dd86a6720cd96c2496544f37f74", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4264, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4265, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4264, - "estimated_input_tokens": 1066, - "estimated_eval_tokens": 2388 + "serialized_chars": 4265, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All four are closed issues with very little content and highly generic titles; there is not enough evidence to confidently confirm any pair as the same underlying issue. The cluster is most likely centered on the repeated \"huggingface\" reports, but the data is too sparse for a duplicate merge decision.", - "confidence": 0.33, - "canonical_issue_reason": "issue:43845 is the earliest of the near-identical \"huggingface\" issues, so it is the safest representative if one must be chosen.", + "summary": "The three \"huggingface \" issues appear to be duplicates of the same report; \"Fett\" looks unrelated. I would treat the earliest huggingface issue as the canonical issue and the other two as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue #43845 is the earliest of the near-identical \"huggingface \" reports and is the best representative of that duplicate cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43845 is the most representative issue in the cluster because it matches the repeated title pattern and appears earliest among the similar reports.", + "best_issue_reason": "Issue #43845 is the best global issue choice because it is the first clear member of the repeated \"huggingface \" group, while #43835 has a different title and no visible overlap in content.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Only vague title similarity; no body details, targets, or filenames to show the same underlying issue." + "reason": "Different titles and no content evidence beyond weak similarity; not enough to conclude the same underlying issue." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Only vague title similarity; insufficient evidence that these are duplicates of the same bug or request." + "reason": "Different titles and no shared details indicating the same bug." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Generic titles do not establish the same underlying issue, and there is no supporting discussion or target evidence." + "reason": "Different titles and no concrete overlap in the issue descriptions." }, { "left": "issue:43845", "right": "issue:43846", - "accept": false, - "reason": "The titles match closely, but the records are too sparse to confirm the same concrete issue with confidence." + "accept": true, + "reason": "Near-identical title and timing strongly suggest duplicate reports of the same underlying issue." }, { "left": "issue:43845", "right": "issue:43859", - "accept": false, - "reason": "Similar generic title only; no substantive content to prove they are the same issue." + "accept": true, + "reason": "Same title and same apparent topic; likely a duplicate of the same report." }, { "left": "issue:43846", "right": "issue:43859", - "accept": false, - "reason": "Similar generic title only; not enough evidence to safely merge as duplicates." + "accept": true, + "reason": "Same title and repeated topic with no distinguishing details, so they likely describe the same issue." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The three identical \"huggingface \" issues are reasonably grouped as a likely duplicate cluster based on the matching titles and high soft-similarity scores, while #43835 is plausibly separate given its different title and much weaker similarity. The earliest huggingface report (#43845) is a conservative canonical choice." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -18154,88 +18219,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "6767e9f96b96ff8682e2e52b0f456fde87869a7428c2b18c553a2cdb9417ff9f", + "prepared_review_unit_hash": "05e822059b0772b7fcced7205cbde2681ffebe8aac92aaebbb1ef50b1fe9dfa9", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "serialized_chars": 5791, + "estimated_input_tokens": 1448, + "estimated_eval_tokens": 3152 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5794, - "estimated_input_tokens": 1449, - "estimated_eval_tokens": 3154 + "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": "These are low-information issues, mostly spam placeholders, plus one generic documentation suggestion and one vague topic. There is no clear shared underlying bug or change to deduplicate.", - "confidence": 0.88, - "canonical_issue_reason": "#44418 is the only non-spam, minimally substantive issue, so it is the best representative by signal, but the cluster lacks enough detail to justify a true duplicate canonical.", + "summary": "The cluster is made up of low-signal, likely spam or placeholder issues with only superficial textual similarity. They do not appear to describe the same underlying bug or request, so they should not be deduplicated together.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "#44418 is the most legitimate-looking and topic-bearing issue in an otherwise spammy set, though it still does not provide a concrete bug or reproducible change.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are spam placeholders with no body detail; identical titles alone do not establish the same underlying issue." + "reason": "Both are generic spam-like placeholders; no evidence they report the same underlying issue." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders, but there is no concrete problem description to confirm they are duplicates of the same bug." + "reason": "Both are generic spam-like placeholders; superficial similarity is not enough to treat them as duplicates." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is spam and the other is a generic documentation suggestion; no shared issue topic is evident." + "reason": "One is spam-like, the other is a vague documentation suggestion; they do not match on a concrete problem." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "A spam placeholder and a vague name-related issue are not the same underlying problem." + "reason": "The titles are unrelated and do not indicate the same bug or request." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "A documentation enhancement request and a spam placeholder are unrelated at the substantive level." + "reason": "A vague documentation suggestion is not the same as a spam-like placeholder issue." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "The documentation suggestion does not share a concrete bug or change request with the spam placeholder." + "reason": "These do not describe the same concrete request or failure mode." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Matching spam titles are not enough to conclude they describe the same underlying issue." + "reason": "Both are spam-like placeholders, but there is no evidence they refer to the same underlying report." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "A spam placeholder and a vague company-name issue do not indicate the same bug or change." + "reason": "The titles suggest different content; no shared bug/request is apparent." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "No concrete shared code-path or problem description is present to justify a duplicate merge." + "reason": "No concrete overlap beyond superficial similarity, so they should not be merged." } ] }, @@ -18265,7 +18330,7 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "301aaffca25696a8bf2a4642862e6ef5a373312a46c9a352e85e0db61870992b", + "prepared_review_unit_hash": "c656e317459e162e74ff56ae3ed46844b4d8596d79e2711ee4a6d92d02138a57", "component_budget": { "node_count": 2, "item_count": 2, @@ -18287,28 +18352,28 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues look like likely duplicates: they have identical titles and very similar body sizes, but the packet only provides soft-similarity evidence, so the duplicate call should stay cautious rather than definitive. #45145 is the better canonical issue because it is earlier and has the inbound reference.", - "confidence": 0.89, - "canonical_issue_reason": "#45145 is earlier and already has an inbound reference, so it is the better canonical representative of this likely duplicate pair.", + "summary": "#45145 is the best canonical issue record because it was opened first and has the only inbound reference. #45151 looks highly similar, but the available evidence is limited to matching titles and soft similarity, so the duplicate call should stay conservative.", + "confidence": 0.88, + "canonical_issue_reason": "#45145 is earlier and has the only inbound reference, making it the best canonical issue to keep.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the strongest single issue to keep because it appears first and has the same apparent scope as #45151, while also having slightly more reference activity.", + "best_issue_reason": "#45145 is the strongest issue representative for the cluster because it is earliest and slightly better connected.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", - "accept": true, - "reason": "Likely duplicate: the titles are identical and the bodies are highly similar in length, but the evidence is only soft-similarity, so this should be treated as a probable duplicate rather than a certain one." + "accept": false, + "reason": "The issues have the same title and appear closely related, but only soft-similarity evidence is available here; without body text or other distinguishing detail, this is not enough to confirm the same underlying request." } ] }, "evaluator_result": { "accept": false, - "feedback": "The canonical-issue choice is reasonable, but the duplicate verdict is overstated. The packet only shows identical titles, similar body lengths, and a soft-similarity signal with deterministic_accept=false; it does not provide enough evidence to call them \"clear duplicates\" or confirm identical scope. Please soften the language to something like \"likely duplicate\" unless the packet includes stronger textual evidence." + "feedback": "The canonical selection is mostly grounded (#45145 is earlier and has one inbound reference vs. none). However, the duplicate-issue conclusion is overstated for the packet: the evidence shown is only soft similarity plus identical titles, and there is no body text or other content here to support \u201cnear-exact,\u201d \u201csame apparent content,\u201d or \u201cno distinguishing technical differences.\u201d The soft edge verdict should be more conservative, e.g. \u201clikely duplicate / likely same request,\u201d not a strong accept." }, "evaluator_used": true, "retried": true, - "accepted_nontrivial_soft_edge": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -18385,12 +18450,12 @@ "soft_pairs": [ "issue:45672|issue:45674" ], - "prepared_review_unit_hash": "5fce0123e820eca46677c6ad94a26c19773b62e57a96c722b6eb5bea8ffa3455", + "prepared_review_unit_hash": "2847beeb0078578be6da3098a2f55e1603fa1dd4b47e1e5a98c9ababbe29d676", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1671, + "serialized_chars": 1670, "estimated_input_tokens": 418, "estimated_eval_tokens": 1092 }, @@ -18398,7 +18463,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1671, + "serialized_chars": 1670, "estimated_input_tokens": 418, "estimated_eval_tokens": 1092 }, @@ -18407,18 +18472,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are both 4-bit inference bugs, but they appear to be different failures in different code paths: Gemma4 audio modules hit a torch.finfo() TypeError on uint8 weights, while BitsAndBytesConfig skip_modules clears the default lm_head exclusion and triggers an AssertionError. Not a duplicate pair.", + "summary": "These two issues are related to 4-bit inference, but they describe different bugs and should not be deduplicated. One is a torch.finfo TypeError triggered by uint8 audio-module weights in Gemma4; the other is a BitsAndBytesConfig regression where llm_int8_skip_modules drops the default lm_head exclusion and causes an AssertionError.", "confidence": 0.97, - "canonical_issue_reason": "No single canonical duplicate here: the two issues have different symptoms, causes, and affected paths. If forced, 45674 is the clearer report, but it does not subsume 45672.", + "canonical_issue_reason": "No canonical issue for the pair: issue 45672 is an audio-module uint8/torch.finfo failure, while issue 45674 is a separate BitsAndBytesConfig default-exclusion regression.", "canonical_pr_reason": null, - "best_issue_reason": "45674 is the stronger standalone issue report: it states a specific BitsAndBytesConfig regression, includes a concrete failure mode, and isolates the likely cause more directly.", + "best_issue_reason": "Neither issue is a good global duplicate target for the other; they have different root causes and would need separate fixes.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45672", "right": "issue:45674", "accept": false, - "reason": "Different bugs: 45672 is a Gemma4/audio-module torch.finfo() TypeError on uint8 weights, while 45674 is a BitsAndBytesConfig skip_modules regression that removes the default lm_head exclusion and causes an AssertionError." + "reason": "Different failure modes and code paths: 45672 is about torch.finfo on uint8 weights in audio modules, while 45674 is about llm_int8_skip_modules clearing the default lm_head exclusion in 4-bit inference." } ] }, @@ -18448,40 +18513,40 @@ "soft_pairs": [ "issue:45762|issue:45763" ], - "prepared_review_unit_hash": "3a6b36ef35e4aec2c24c77634fd39fc6454da0b28a278e9532e3c72ef9ef4180", + "prepared_review_unit_hash": "63e87cac3a1210700acfbc5c8b45439c303df3563eb4a25cf56c93d2c707a593", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1477, - "estimated_input_tokens": 370, - "estimated_eval_tokens": 996 + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1477, - "estimated_input_tokens": 370, - "estimated_eval_tokens": 996 + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues do not appear to describe the same bug or request; they are just two unrelated personal-name title entries with no body or shared technical context.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "Both items are closed issues with only person-name/date titles and no body, links, or shared targets; they do not appear to describe the same bug or request, so this is not a duplicate cluster.", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: neither issue contains enough technical detail, and they do not share a specific problem statement or target.", "canonical_pr_reason": null, - "best_issue_reason": "No clear canonical issue exists here; if forced to pick one representative, issue #45762 is slightly earlier, but the pair does not form a duplicate cluster.", + "best_issue_reason": "Neither issue is a good global representative for a technical cluster because both look like isolated, content-free entries rather than actionable bug reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45762", "right": "issue:45763", "accept": false, - "reason": "Different names and dates, no body text, no shared targets, and no technical signal that they refer to the same underlying issue." + "reason": "The titles are different personal-name/date strings with no shared code context, body, links, or target; they do not look like the same underlying issue." } ] }, @@ -18511,7 +18576,7 @@ "soft_pairs": [ "issue:45764|issue:45765" ], - "prepared_review_unit_hash": "7f3d58098359ba37f9cdb7600bb6da87005aee9c23852661436abcdf310a2ca1", + "prepared_review_unit_hash": "85f59221a84bcfbcef99891707844afb90e5d3b4fc596e3257676fe0010aff84", "component_budget": { "node_count": 2, "item_count": 2, @@ -18533,8 +18598,8 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two closed issues have different person-name titles and different dates, with no body text, links, or shared targets suggesting the same underlying bug or request. They do not look like duplicates.", - "confidence": 0.98, + "summary": "The two closed issues are just empty title-only reports with different personal-name/date strings and no shared technical context, so they should not be treated as duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, @@ -18544,7 +18609,7 @@ "left": "issue:45764", "right": "issue:45765", "accept": false, - "reason": "Different named subjects and dates; no shared context, description, or issue targets. These do not appear to report the same underlying problem." + "reason": "No evidence of the same underlying bug or request: different titles, no body, no linked targets, and no shared technical context." } ] }, @@ -18574,12 +18639,12 @@ "soft_pairs": [ "issue:45810|issue:46295" ], - "prepared_review_unit_hash": "8700b7e0b344efba85987743d1e60781111e1c8ee45a9853aadc92d6d098bee7", + "prepared_review_unit_hash": "6cae1f4b9275f29c0c333e4321b1a44927132884998a4d977974712a30a4bc12", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1541, + "serialized_chars": 1542, "estimated_input_tokens": 386, "estimated_eval_tokens": 1028 }, @@ -18587,7 +18652,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1541, + "serialized_chars": 1542, "estimated_input_tokens": 386, "estimated_eval_tokens": 1028 }, @@ -18596,18 +18661,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are not duplicates: one requests a new Qwen3_5 token-classification model class for value-model use, while the other asks for FP8 training support. They concern different features and code paths.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "The two issues are unrelated: one requests a new Qwen3_5 token-classification model for value modeling, while the other asks for FP8 training support. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the requests target different problems and would not be resolved by the same change.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "Neither issue is a good cluster representative because they are about separate feature requests with no shared underlying bug or fix.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45810", "right": "issue:46295", "accept": false, - "reason": "Different requests with no shared code path: model-class addition vs FP8 training support." + "reason": "Different topics and code paths: adding a model class versus enabling FP8 training. They do not describe the same underlying issue." } ] }, @@ -18647,91 +18712,91 @@ "issue:45871|issue:45873", "issue:45872|issue:45873" ], - "prepared_review_unit_hash": "5985f470f91912988129d91deb52b52fcc5de5dba789f7c53382b1a3abdb7aa3", + "prepared_review_unit_hash": "2da42aea1df98d74cd979b96ff096e9aa7a951cf9f3ab2b3620dfc6f5d37d9fb", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5721, - "estimated_input_tokens": 1431, - "estimated_eval_tokens": 3118 + "serialized_chars": 5727, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 8, - "serialized_chars": 5721, - "estimated_input_tokens": 1431, - "estimated_eval_tokens": 3118 + "serialized_chars": 5727, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster splits into two clear duplicate pairs: 45870/45871 and 45872/45873. Issue 45869 is related to financial data but is a different request/topic. The two example-style pairs are not the same underlying issue as each other.", - "confidence": 0.97, - "canonical_issue_reason": "45870 is a strong canonical for its duplicate pair because it is the earliest of the identical-title issues and reads like the more generic request. The other exact-duplicate pair (45872/45873) appears to be a separate topic, not a duplicate of 45870.", + "summary": "The cluster appears to split into two duplicate-looking/same-topic issue pairs: 45870/45871 and 45872/45873. Issue 45869 looks related thematically but not like the same issue as either pair.", + "confidence": 0.79, + "canonical_issue_reason": "45870 is a reasonable representative for the 45870/45871 same-topic pair because it is the earlier of the two and their titles/body lengths match closely; however, the cluster does not look like one single duplicate set overall.", "canonical_pr_reason": null, - "best_issue_reason": "45870 is the best single issue to represent one duplicate subgroup, but the cluster as a whole contains two distinct issue topics. If forced to pick one representative, 45870 is the most central and earliest of the exact duplicates.", + "best_issue_reason": "45870 is the best issue to keep as a representative for the strongest same-topic pair, since it is the earlier of the matching 45870/45871 issues and there is no stronger evidence tying 45869 into the same set.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45869", "right": "issue:45870", "accept": false, - "reason": "Both mention financial data, but 45869 is a dataset/sample offering while 45870 is an example/tutorial request. Same broad domain, different underlying issue." + "reason": "They share financial/DataForge context, but one is a dataset/free-sample post and the other is a prediction example; they do not look like the same issue." }, { "left": "issue:45869", "right": "issue:45871", "accept": false, - "reason": "Dataset/sample bundle versus example workflow; related theme, but not the same concrete request." + "reason": "They share financial/DataForge context, but one is a dataset/free-sample post and the other is a prediction example; they do not look like the same issue." }, { "left": "issue:45870", "right": "issue:45871", "accept": true, - "reason": "Identical title and matching context; these look like the same issue duplicated." + "reason": "Identical title/body-length pair with very strong soft-similarity; these appear to be duplicate or same-topic issues." }, { "left": "issue:45870", "right": "issue:45872", "accept": false, - "reason": "One is an example for financial time series prediction, the other adds a scraping service. Similar area, but different underlying request." + "reason": "Both are about DataForge finance-related content, but they describe different topics: an example versus a scraping-service request." }, { "left": "issue:45870", "right": "issue:45873", "accept": false, - "reason": "Same as 45870/45872: related financial/DataForge theme, but not the same concrete issue." + "reason": "Both are about DataForge finance-related content, but they describe different topics: an example versus a scraping-service request." }, { "left": "issue:45871", "right": "issue:45872", "accept": false, - "reason": "These are distinct requests despite overlapping financial/DataForge terminology." + "reason": "Both are about DataForge finance-related content, but they describe different topics: an example versus a scraping-service request." }, { "left": "issue:45871", "right": "issue:45873", "accept": false, - "reason": "Distinct topic from the exact-title duplicate pair; no evidence they are the same underlying issue." + "reason": "Both are about DataForge finance-related content, but they describe different topics: an example versus a scraping-service request." }, { "left": "issue:45872", "right": "issue:45873", "accept": true, - "reason": "Identical title and matching context; these appear to be the same issue duplicated." + "reason": "Identical title/body-length pair with very strong soft-similarity; these appear to be duplicate or same-topic issues." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The grouping is grounded in the packet: 45870/45871 and 45872/45873 have identical titles, while 45869 is only soft-related. The verdicts stay conservative overall, and the summary does not overreach beyond the available evidence." + "accept": false, + "feedback": "The cluster split is plausible, but the writeup overstates the evidence. The packet only shows identical titles and body lengths plus soft-similarity scores; it does not prove the pairs are exact duplicate title/body matches. Please soften the language to something like \"appear to be duplicate/same-topic issues\" and avoid claiming exact duplicates or same underlying issue unless the bodies or explicit links support it. The 45869 non-match call is reasonably conservative, but the cluster-level summary is too strong." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -18755,12 +18820,12 @@ "soft_pairs": [ "issue:46439|issue:46441" ], - "prepared_review_unit_hash": "f29119108535f31f39b95c8664c5fb06cf7665b3af88f2abd78ef5ece67cc450", + "prepared_review_unit_hash": "389a1a13a49eadbd479354e9053f3462e5bb239d236f61cea9a035fc294ee44b", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1634, + "serialized_chars": 1636, "estimated_input_tokens": 409, "estimated_eval_tokens": 1074 }, @@ -18768,7 +18833,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1634, + "serialized_chars": 1636, "estimated_input_tokens": 409, "estimated_eval_tokens": 1074 }, @@ -18777,24 +18842,27 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open issues both concern static-cache behavior around linear attention, but they describe different failure modes: one is shape corruption during early initialization, the other is a Qwen3.5 `KeyError` for `linear_attention`. They look related at a subsystem level, but not clearly the same underlying bug.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 46439 is the better canonical representative because it states a concrete, likely root-cause static-cache corruption affecting linear-attention layers, rather than a model-specific downstream crash.", + "summary": "The two issues are closely related around static-cache and linear-attention handling, but the packet does not establish that the Qwen3.5 `KeyError: 'linear_attention'` and the wrong-shape `RuntimeError` are the same root bug. They should be treated as related reports rather than confirmed duplicates.", + "confidence": 0.61, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it describes the broader mechanism (`early_initialization` corrupting linear-attention layers) instead of a model-specific symptom.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 46439 is the most general and actionable description of the apparent defect in the set.", + "best_issue_reason": "Issue 46439 is the best representative for the cluster since it is the more general diagnostic report, while 46441 appears to be a narrower failure report in the same area.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:46439", "right": "issue:46441", "accept": false, - "reason": "Both mention static cache and linear attention, but one reports wrong-shape corruption and the other reports a `KeyError: 'linear_attention'`; these are not clearly the same concrete bug or fix." + "reason": "The titles suggest a shared static-cache/linear-attention area, but the evidence is too thin to conclude the `KeyError` and wrong-shape `RuntimeError` come from the same underlying bug." } ] }, - "evaluator_result": null, - "evaluator_used": false, - "retried": false, + "evaluator_result": { + "accept": false, + "feedback": "The conclusion is a bit too strong for the evidence in the packet. We only have titles and a soft-similarity signal, so saying the two issues are the same underlying bug is not fully grounded. A more conservative verdict would be that they appear related and may involve the same static-cache/linear-attention area, but the packet does not establish that the KeyError and wrong-shape RuntimeError are manifestations of one root cause." + }, + "evaluator_used": true, + "retried": true, "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null @@ -18848,154 +18916,154 @@ "pull_request:43823|pull_request:45213", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "da4bdcd3bb9f0fd3a8d77dffd2b411aab51699f75350f42bba3e21db1542000c", + "prepared_review_unit_hash": "327b0e7c28b236270ae23f60ff7a574975905568302999540eee4b506fad9b7b", "component_budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 19, - "serialized_chars": 38143, - "estimated_input_tokens": 9536, - "estimated_eval_tokens": 19328 + "serialized_chars": 38083, + "estimated_input_tokens": 9521, + "estimated_eval_tokens": 19298 }, "budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 19, - "serialized_chars": 38143, - "estimated_input_tokens": 9536, - "estimated_eval_tokens": 19328 + "serialized_chars": 38083, + "estimated_input_tokens": 9521, + "estimated_eval_tokens": 19298 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is mostly unrelated PRs spanning new model additions, tokenizer/processor fixes, loss utilities, export tests, naming cleanup, and docs/serve parsing. The only clear duplicate-style pair is the two `fixed_cross_entropy` changes, which both target the same loss helper and the same tracked bug.", - "confidence": 0.95, + "summary": "This cluster is mostly a set of unrelated pull requests across different models, docs, and utilities. The only clear duplicate-like pair is the two `fixed_cross_entropy` changes, which address the same loss helper in the same code path.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical PR is truly representative of the whole cluster, but `pull_request:43251` is the best anchor because it directly addresses the tracked loss bug with the clearest, most concrete fix scope.", + "canonical_pr_reason": "PR 43251 is the best representative of the only true duplicate pair: it is directly tied to issue 43240, changes the same `fixed_cross_entropy` helper, and reads like the more complete fix for passing through `cross_entropy` kwargs.", "best_issue_reason": null, - "best_pr_reason": "`pull_request:43251` is the best global PR candidate because it has the strongest evidence of a specific bug fix: it explicitly targets issue 43240 and updates `fixed_cross_entropy` to pass the relevant `cross_entropy` kwargs. `pull_request:43254` is a near-duplicate variant of the same fix, but 43251 is the more complete candidate.", + "best_pr_reason": "PR 43251 is the strongest standalone candidate overall because it has the explicit issue target, a focused fix, and the clearest linkage to the underlying bug.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Unrelated: one adds the VideoPrism model stack and docs, the other adds an ExecuTorch dynamic-shape test." + "reason": "Different domains and code paths: VideoPrism model addition vs ExecuTorch dynamic-shape export test." }, { "left": "pull_request:41855", "right": "pull_request:43823", "accept": false, - "reason": "Unrelated tokenizer-method patch versus adding a new MobileLLM model implementation." + "reason": "Unrelated tokenizer utility fix vs new MobileLLM model implementation." }, { "left": "pull_request:41855", "right": "pull_request:43424", "accept": false, - "reason": "Unrelated tokenizer-method patch versus an ExecuTorch export test." + "reason": "Tokenizer method addition and ExecuTorch export testing are not the same underlying change." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Both touch processor/model auto plumbing, but they fix different models and different problems." + "reason": "Processor auto-loading robustness and MobileLLM model addition are different features." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Different subsystems and goals: loss-function kwarg handling versus new model addition." + "reason": "Loss helper kwargs fix is unrelated to adding a new model." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Loss utility kwarg support is unrelated to Mistral4 test and kernel/model cleanup changes." + "reason": "The loss helper fix and Mistral4 test/model changes are unrelated code paths." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "ExecuTorch dynamic-shape export testing is unrelated to Mistral4 test fixes." + "reason": "ExecuTorch export testing and Mistral4 test/model work do not fix the same bug." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "ExecuTorch export test versus a brand-new model implementation; not the same bug or change." + "reason": "Different concerns: dynamic export test vs new MobileLLM model." }, { "left": "pull_request:43823", "right": "pull_request:45655", "accept": false, - "reason": "New MobileLLM model implementation is unrelated to config-resolution ordering in modeling_utils." + "reason": "New model addition and `cls.config` resolution ordering are unrelated." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "New model addition versus Falcon output-interface refactor; different code paths and intent." + "reason": "MobileLLM model addition and Falcon output collection refactor are separate changes." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon refactor and model interface work are unrelated to the model-creation skill PR." + "reason": "Falcon refactor and model-creation skill docs are unrelated." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output-collection refactor is unrelated to Mistral4 test/kernel fixes." + "reason": "Falcon standardized output refactor and Mistral4 test fixes do not share the same concrete bug." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Mistral4 test fixes are unrelated to the layrnorm-to-layernorm naming cleanup across CLIP-like models." + "reason": "Mistral4 tests vs `layrnorm` typo normalization is a broad subsystem overlap only, not the same fix." }, { "left": "pull_request:44827", "right": "pull_request:46267", "accept": false, - "reason": "Mistral4 test fixes are unrelated to chat response parsing updates." + "reason": "Mistral4 test fixes and serve chat parsing updates are unrelated." }, { "left": "pull_request:44827", "right": "pull_request:46048", "accept": false, - "reason": "Mistral4 test fixes are unrelated to adding the TDT loss kernel integration." + "reason": "Mistral4 tests and adding a TDT loss kernel are different changes." }, { "left": "pull_request:44827", "right": "pull_request:45655", "accept": false, - "reason": "Mistral4 test fixes are unrelated to `cls.config` resolution ordering." + "reason": "Mistral4 test/model work and config resolution ordering are not the same underlying issue." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "They only share a broad release-era context; model test fixes and model-creation skill are not the same change." + "reason": "Model tests and the new-model skill documentation are not mergeable as one duplicate PR." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "New MobileLLM model implementation is unrelated to the model-creation skill/meta tooling PR." + "reason": "A model implementation PR and a repository skill/documentation PR are unrelated." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both patch the same `fixed_cross_entropy` bug in `src/transformers/loss/loss_utils.py`, and both are tied to the same issue 43240; they are near-duplicates of the same concrete fix." + "reason": "Both PRs modify `fixed_cross_entropy` in the same file to support passing `cross_entropy` kwargs for the same issue 43240; they are alternative fixes for the same code-path bug." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well grounded in the packet: most items are clearly unrelated across model additions, utilities, tests, and docs, and the 43251/43254 pair is a conservative duplicate-style match because both edit `src/transformers/loss/loss_utils.py` for the same issue 43240. No soft-edge verdict appears overstated." + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted pair, 43251/43254, is clearly mergeable into one PR for the same bugfix: both change `src/transformers/loss/loss_utils.py`, both target issue 43240, and both extend `fixed_cross_entropy` to forward supported `cross_entropy` kwargs. The other soft-edge rejections are also cautious and consistent with the file lists and titles." }, "evaluator_used": true, "retried": false, @@ -19024,7 +19092,7 @@ "pull_request:40546|pull_request:42781", "pull_request:42781|pull_request:46331" ], - "prepared_review_unit_hash": "5190db9af5e1ae990c63b78ddaededd6f0372ed646e5275b418a21559979a7cb", + "prepared_review_unit_hash": "10ac43826f39335e5faa28af853569cd4912fe54304608461c5cfd944844ba39", "component_budget": { "node_count": 3, "item_count": 3, @@ -19046,24 +19114,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate PRs: one base VibeVoice implementation, one VibeVoice realtime extension, and one Parakeet-RNNT addition. They share some auto-mapping scaffolding, but they do not fix or add the same concrete change.", + "summary": "The cluster contains three distinct pull requests, not duplicate changes: an initial VibeVoice implementation, a later VibeVoice Realtime add-on, and an unrelated Parakeet-RNNT integration. Both soft edges should be rejected.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 40546 is the clearest canonical representative: it is the original VibeVoice implementation, not a draft, and it establishes the core model/docs/auto-registry pieces that the later VibeVoice realtime PR builds on.", + "canonical_pr_reason": "PR 40546 is the most central representative: it introduces the core VibeVoice implementation and integration surface, while 42781 is a separate realtime extension and 46331 targets a different model family.", "best_issue_reason": null, - "best_pr_reason": "PR 40546 is the best overall PR for the cluster because it is the base, completed VibeVoice integration; the other PRs are either a later realtime variant of the same model family or a different model entirely.", + "best_pr_reason": "PR 40546 is the best standalone representative for the cluster because it is the original full VibeVoice implementation with the broadest scope and clearest subject matter.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Related model family, but not the same underlying change. 40546 adds core VibeVoice, while 42781 is a draft realtime extension with additional tokenizer/generation work and broader scope." + "reason": "Related topic, but not the same underlying change: 40546 implements core VibeVoice, while 42781 adds VibeVoice Realtime with acoustic-tokenizer and generation changes. They overlap in files but are not mergeable as one duplicate PR." }, { "left": "pull_request:42781", "right": "pull_request:46331", "accept": false, - "reason": "Different model families and different additions. Shared auto-registration files are incidental; VibeVoice realtime and Parakeet-RNNT are not the same concrete feature." + "reason": "Different model families and different concrete changes. The shared auto-mapping files are generic integration plumbing, not evidence of the same bug or feature." } ] }, @@ -19093,7 +19161,7 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "1458705be84a34a189f688db1550a648797012f43715b96a6617d75f89a3a3f6", + "prepared_review_unit_hash": "2d0021d37b7bf37e5e5cb7f94d5da50cb24b8b64f5891e4039a6e94d69f7d407", "component_budget": { "node_count": 2, "item_count": 2, @@ -19115,18 +19183,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification. They do not look like the same underlying change, so the soft edge should be rejected and there is no meaningful duplicate cluster representative among them.", + "summary": "The two pull requests are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They do not look like the same underlying bug or change.", "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs target different code paths, different models, and different user-facing features, so there is no duplicate anchor in this cluster.", "best_issue_reason": null, - "best_pr_reason": "No single canonical PR stands out because the changes target different codebases and user-facing capabilities; neither PR is a duplicate of the other.", + "best_pr_reason": "No global best PR: neither PR is a duplicate of the other, and each addresses a separate feature area with no meaningful mergeable overlap.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models, different tasks, and different code paths: TimesFM forecasting with covariates versus DINOv3 image classification support. They are not the same underlying bug or change." + "reason": "Different subsystems and different changesets: TimesFM forecasting with covariates versus DINOv3 image classification support. They are not the same underlying change and should not be merged as duplicates." } ] }, @@ -19221,12 +19289,12 @@ "pull_request:41797|pull_request:45186", "pull_request:45186|pull_request:45630" ], - "prepared_review_unit_hash": "ceb4eec8fdcb64152eecb8165d06f4cfc661fdc33d49918fc60aadebb647bdcd", + "prepared_review_unit_hash": "b50793534bd4dea6da266cde310c14ef12e916d865e74c016ef03637d4db9955", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10436, + "serialized_chars": 10435, "estimated_input_tokens": 2609, "estimated_eval_tokens": 5474 }, @@ -19234,7 +19302,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 10436, + "serialized_chars": 10435, "estimated_input_tokens": 2609, "estimated_eval_tokens": 5474 }, @@ -19243,24 +19311,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three distinct open PRs that all add new model integrations (DeepSeek OCR, Isaac, Kimi2-6). They share the usual Transformers scaffolding files, but they target different models, configs, processors, and tests, so this cluster is not a duplicate set.", - "confidence": 0.96, + "summary": "These are three separate model-addition PRs that only overlap on the standard Transformers boilerplate. The shared files and \"soft_similarity\" signals are not enough to make them duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "No single PR is a canonical duplicate: each one introduces a different model integration with separate codepaths and tests.", + "canonical_pr_reason": "No single canonical PR: 41797 adds DeepSeek OCR, 45186 adds Isaac, and 45630 adds Kimi2-6; they target different models and code paths.", "best_issue_reason": null, - "best_pr_reason": "If a representative PR is needed, #45630 (Kimi2-6) is the broadest standalone model-addition change, but it is not a duplicate of the others.", + "best_pr_reason": "45630 is the strongest representative of the group: it has the most review activity and a broad, well-scoped model integration across auto mappings and model files.", "soft_edge_verdicts": [ { "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different model additions (DeepSeek OCR vs Isaac). They only overlap on shared integration scaffolding, not the same underlying change." + "reason": "Different model additions (DeepSeek OCR vs Isaac). The overlap is only in shared Transformers scaffolding, not the same underlying bug or change." }, { "left": "pull_request:45186", "right": "pull_request:45630", "accept": false, - "reason": "Different model additions (Isaac vs Kimi2-6) with separate model-specific files and tests; same broad subsystem, but not the same fix or feature." + "reason": "Different model additions (Isaac vs Kimi2-6). Shared docs/auto-registration files are generic boilerplate, not evidence of one mergeable fix." } ] }, @@ -19290,7 +19358,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "123463cc18fbfa7170d4fa1e39676cdf9beb8c68b13b4ceadfe0c2767048ac54", + "prepared_review_unit_hash": "3d1f6d0fe6fa77a556b6d6f8155a4843dffb6ea33ac1f2f872635f86e0280831", "component_budget": { "node_count": 2, "item_count": 2, @@ -19312,18 +19380,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs both modernize GitHub Actions workflow dependencies, but they are different change sets: one targets Node 24 compatibility across several workflow actions, while the other upgrades docker/build-push-action across Docker image workflows. They are related maintenance work, not duplicates.", - "confidence": 0.94, + "summary": "Two open PRs both modernize GitHub Actions in workflow files, but they target different dependency families and different underlying maintenance issues. They are related thematically, not duplicates.", + "confidence": 0.92, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 42975 is the broader workflow-action upgrade and the better representative if one PR must stand in for this cluster; it updates multiple core GitHub Actions used across many workflows for Node 24 compatibility.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 42975 is the best single PR representative because it covers the wider cross-workflow Actions runtime compatibility update, whereas PR 42976 is a narrower Docker action version bump.", + "best_pr_reason": "PR 42975 is the broader representative of the cluster\u2019s theme because it upgrades multiple core workflow actions for Node 24 compatibility across many workflows. It is still not a duplicate of 42976, which focuses on docker/build-push-action upgrades and a different workflow subset.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Reject: these are different maintenance changes. PR 42975 upgrades checkout/cache/upload-artifact for Node 24 compatibility; PR 42976 upgrades docker/build-push-action. They touch overlapping workflow files, but not the same underlying bug or mergeable code-path fix." + "reason": "Reject: although both are workflow-action maintenance PRs, 42975 is about GitHub Actions runtime/version compatibility (checkout/cache/upload-artifact), while 42976 primarily upgrades docker/build-push-action and related workflow action versions. They do not fix the same concrete code-path problem and would not plausibly be merged as one PR." } ] }, @@ -19345,114 +19413,85 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-42978-5", + "cluster_id": "cluster-42978-4", "nodes": [ "pull_request:42978", "pull_request:43448", "pull_request:43451", - "pull_request:45490", - "pull_request:46411" + "pull_request:45490" ], "soft_pairs": [ "pull_request:42978|pull_request:43448", "pull_request:42978|pull_request:43451", "pull_request:42978|pull_request:45490", - "pull_request: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" + "pull_request:43451|pull_request:45490" ], - "prepared_review_unit_hash": "5689bb03695d4b4c2159569de38bd579ae95baa95a9e26c4600ccb2843fda216", + "prepared_review_unit_hash": "ec5d7773f7acf6fc5179e57c0fd0c94771f81bc68e444db0a7b8a150f0b38aed", "component_budget": { - "node_count": 5, - "item_count": 5, - "soft_pair_count": 10, - "serialized_chars": 17822, - "estimated_input_tokens": 4456, - "estimated_eval_tokens": 9168 + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14253, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 }, "budget": { - "node_count": 5, - "item_count": 5, - "soft_pair_count": 10, - "serialized_chars": 17822, - "estimated_input_tokens": 4456, - "estimated_eval_tokens": 9168 + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14253, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are all separate model add-ons or a demo script, not duplicates of the same underlying change. The soft-similarity is driven by shared Transformers boilerplate and docs/auto-registry edits, but each PR targets a different model or artifact.", - "confidence": 0.98, + "summary": "All four pull requests are new model additions for different architectures (ViT NEPA, Molmo, Molmo2, CTSM). They share some registration/docs boilerplate, but they do not fix the same concrete code-path problem, so none should be deduplicated against another.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43451 is the strongest representative if one must be chosen: it is the most complete and active model-addition PR, covering Molmo2 model, processor, image/video support, tests, docs, and registry updates.", + "canonical_pr_reason": "No single PR is a duplicate canonical for this cluster; if one must be chosen as a representative, Molmo2 is the most substantial and complete addition, but it is still a distinct model from the others.", "best_issue_reason": null, - "best_pr_reason": "PR 43451 is the best overall representative because it has the broadest scope and highest review activity, indicating the most mature and central change among these unrelated additions.", + "best_pr_reason": "PR 43451 (Molmo2) is the strongest representative on breadth and completeness, with the highest review activity and the most extensive model/processor coverage, but it is not a duplicate of the others.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo. They share Transformers scaffolding, but not the same code-path problem or mergeable change." + "reason": "Different model families: ViT NEPA vs Molmo. Shared auto/doc updates are generic model-integration boilerplate, not the same underlying change or bug fix." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo2. Shared registry/docs boilerplate does not make them the same change." + "reason": "ViT NEPA and Molmo2 are distinct architectures and code paths. Overlap is limited to common registration/docs files, so they are not mergeable as one PR." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model families: ViT NEPA vs CTSM. No evidence they fix the same bug or implement the same feature." - }, - { - "left": "pull_request:42978", - "right": "pull_request:46411", - "accept": false, - "reason": "ViT NEPA model addition and a MiniMax demo/code drop are unrelated changes." + "reason": "ViT NEPA is a vision model add-on, while CTSM is a time-series model add-on. They touch unrelated modeling code and do not address the same change." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are related by name, but they are distinct model implementations with different files and support surfaces." + "reason": "Molmo and Molmo2 are related by naming but still separate model implementations with different files and functionality; this is not the same concrete fix/change." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are unrelated model additions; shared auto/model registration files are not enough to merge them." - }, - { - "left": "pull_request:43448", - "right": "pull_request:46411", - "accept": false, - "reason": "Molmo model integration and a MiniMax demo script are not the same underlying change." + "reason": "Molmo and CTSM are unrelated model additions. Shared files are just integration points, which is insufficient for duplicate PR merging." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 vs CTSM are separate model integrations with different architectures and support code." - }, - { - "left": "pull_request:43451", - "right": "pull_request:46411", - "accept": false, - "reason": "Molmo2 model integration and a MiniMax demo/script are unrelated." - }, - { - "left": "pull_request:45490", - "right": "pull_request:46411", - "accept": false, - "reason": "CTSM model integration and a MiniMax demo/script do not address the same code-path problem." + "reason": "Molmo2 and CTSM solve different model-support tasks in different subsystems, with no shared bug or code-path target beyond generic auto/registry updates." } ] }, @@ -19489,70 +19528,70 @@ "pull_request:43665|pull_request:46380", "pull_request:44815|pull_request:46380" ], - "prepared_review_unit_hash": "ffb8e23ff9bc762bbd149adfb8d27a63dd43ef216f814b722e9ba7b6fca1f754", + "prepared_review_unit_hash": "efe380863eed3728627fb670d4657896a91793b2599323477e780ddff2c3f47c", "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": "All four pull requests appear to address different bugs or even unrelated test changes; none should be treated as duplicates of the others.", - "confidence": 0.94, + "summary": "No strong duplicate cluster here: the four PRs target different bugs or even look like unrelated test noise. I would not merge any soft-edge pair.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46380 is the clearest standalone bugfix in the set: it corrects a specific generation validation path where watermarking_config can be a dict and converts it before validation.", + "canonical_pr_reason": "PR 43291 is the clearest single bugfix: it fixes Whisper decode behavior in one concrete code path and has a focused scope.", "best_issue_reason": null, - "best_pr_reason": "PR 46380 is the strongest representative because it fixes a concrete runtime code path with a focused behavior change and matching tests, unlike the more diffuse or test-only PRs.", + "best_pr_reason": "PR 43291 is the best representative of a real, narrowly scoped defect fix. The others are either unrelated, much broader, or appear to be test-placeholder noise.", "soft_edge_verdicts": [ { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "Whisper tokenizer/timestamp decode changes vs. CLIP/ViT test edits; different models and different kinds of changes." + "reason": "Not the same underlying change: 43291 fixes Whisper decoding, while 43665 is just test-file edits with bogus assertions and no concrete bugfix path." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decode/test fix vs. FP8 dequant/loading fix; unrelated subsystems and code paths." + "reason": "Different bugs and code paths: Whisper tokenization/decoding versus FP8 dequantization and model loading." }, { "left": "pull_request:43291", "right": "pull_request:46380", "accept": false, - "reason": "Whisper tokenizer behavior vs. generation watermarking validation; no shared underlying bug." + "reason": "43291 is a Whisper tokenizer decode fix; 46380 is about generation config/watermarking validation and broader generation/model changes, not the same defect." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "Test-only assertions in CLIP/ViT are not the same concrete issue as the FP8 dequant fix." + "reason": "43665 looks like placeholder test breakage, not a real fix; 44815 is a real FP8 dequant fix. No shared underlying bug." }, { "left": "pull_request:43665", "right": "pull_request:46380", "accept": false, - "reason": "Artificial failing test edits are unrelated to the generation_config watermarking fix." + "reason": "43665 is nonsensical test-only noise, while 46380 is a substantive generation-validation change. Not duplicates." }, { "left": "pull_request:44815", "right": "pull_request:46380", "accept": false, - "reason": "FP8 dequant/model-loading work and generation watermarking validation are separate code paths and bugs." + "reason": "Both are broad PRs but they fix different concrete problems: FP8 dequant/padding versus generation watermarking validation and related model updates." } ] }, @@ -19582,7 +19621,7 @@ "soft_pairs": [ "pull_request:43488|pull_request:43532" ], - "prepared_review_unit_hash": "641ac6288cf45fd0a10c4d88064c907187f238617d85fd3fd6adc791bbdba63e", + "prepared_review_unit_hash": "13184443d3d1bdfe1f1e9568bb5fe84c5be14a0ca6e6fdfe0bcd33ed503d5789", "component_budget": { "node_count": 2, "item_count": 2, @@ -19604,18 +19643,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated open pull requests in the same repo: one is a small dependency/version-bump test PR, the other changes CI/workflow and testing utility behavior. They do not look like the same underlying change.", - "confidence": 0.98, + "summary": "Two unrelated non-merge PRs with only an incidental overlap in a Vit test file. One is a setup/versioning repo-bot check, the other is a workflow/CI diff demo touching multiple automation files.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43532 is the better representative because it has the broader, more substantive code change set and affects actual workflow/testing paths, while #43488 is a small repo-bot/version-format check.", + "canonical_pr_reason": "pull_request:43488 is the more self-contained and clearly scoped artifact; its change is centered on setup/versioning and a small test tweak, making it the cleaner representative of the cluster.", "best_issue_reason": null, - "best_pr_reason": "PR #43532 is the best overall representative of the cluster because its edits span multiple workflow and testing files and reflect a concrete code-path/process change; #43488 is more of a narrow test fixture/change.", + "best_pr_reason": "pull_request:43488 is the most suitable representative because its intent is explicit and its diff is narrower and easier to triage than the broader CI/workflow-oriented PR.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "Rejected: the diffs are not the same underlying bug/change. #43488 tweaks setup.py dependency/version metadata and a Vit test, while #43532 rewires GitHub workflows and testing utilities across several files." + "reason": "Reject: they do not fix the same underlying problem. The overlap in tests/models/vit/test_modeling_vit.py is incidental, while the rest of the diffs target different areas (setup/versioning vs CI workflows/testing utilities)." } ] }, @@ -19645,7 +19684,7 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "968b898827ebabccae9c6cb250569a73e1713cd27d7c8f06d147d9e60b2897c9", + "prepared_review_unit_hash": "8c40c3b9f45fa820e8bb889d3f0741cb0ee72dd968a224eb0478700b3ee36f1c", "component_budget": { "node_count": 2, "item_count": 2, @@ -19667,18 +19706,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs add new promptable segmentation pipelines, but they target different products and code paths: concept segmentation vs visual segmentation. They share some shared infrastructure/docs, yet they are not the same underlying change.", - "confidence": 0.92, + "summary": "These are two separate feature PRs for different segmentation pipelines: promptable concept segmentation vs promptable visual segmentation. They overlap in docs, registration, and pipeline scaffolding, but they target different models/code paths and should not be treated as duplicates.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43613 is the better representative of the cluster because it has the broader and more fully integrated visual-segmentation implementation, with more model/test coverage and discussion activity.", + "canonical_pr_reason": "No canonical PR: 43612 and 43613 implement different pipeline features with different task docs, pipeline modules, model integrations, and tests.", "best_issue_reason": null, - "best_pr_reason": "PR #43613 is the strongest standalone PR in the cluster for the same reason: it appears more complete and better integrated than #43612, even though it is not a duplicate.", + "best_pr_reason": "No single best PR for the cluster. If forced to choose a representative, 43613 is broader, but it is still a distinct change from 43612 rather than the same underlying fix/feature.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Reject: these are different pipeline additions with different scopes and underlying model families. Shared docs and pipeline registration are too broad to indicate the same concrete change." + "reason": "Different underlying features and code paths: concept segmentation vs visual segmentation. Shared docs/registration scaffolding is not enough to merge them as one PR." } ] }, @@ -19708,7 +19747,7 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "288f795ea74b49e5c9943d6e16f6441f7d9bb4e435bdb17ec74d50354f731f26", + "prepared_review_unit_hash": "59eec95683166ad01e427fa147482226fecbf2c7007e569c3acc347789bac692", "component_budget": { "node_count": 2, "item_count": 2, @@ -19730,18 +19769,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate Trainer.py changes: PR 43651 adjusts gradient-accumulation loss scaling via a new property, while PR 43663 refactors signature-column detection for custom/PEFT trainers. They touch the same file but fix different code paths and are not mergeable as one change.", - "confidence": 0.98, + "summary": "Both PRs touch `trainer.py` but address different trainer extension points: one changes gradient-accumulation loss scaling, the other extracts signature-column filtering into an overridable helper. They are not the same underlying change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #43651 is the more concrete bug-fix style change, with a direct training-step behavior adjustment and clearer user impact.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "PR #43651 is the strongest standalone representative of the cluster because it targets a specific runtime correctness issue rather than a refactor/extension hook.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Different underlying behavior: one changes loss scaling in training_step, the other changes dataset column filtering/signature inspection. Same file, but not the same bug or fix." + "reason": "Same file and similar theme, but they fix different code paths: loss scaling vs. signature column filtering. They could not plausibly be merged as one PR without creating an unrelated mixed change." } ] }, @@ -19771,12 +19810,12 @@ "soft_pairs": [ "pull_request:43656|pull_request:43842" ], - "prepared_review_unit_hash": "a203f83cd3ace383d47ea930d82c0e3b76486c82680fa8936ac3cbb5547cbe3e", + "prepared_review_unit_hash": "3d8f43278d2e928b21ece9f44d8659132c048757023342fffbc9de17d5a32731", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3663, + "serialized_chars": 3664, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -19784,7 +19823,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 3663, + "serialized_chars": 3664, "estimated_input_tokens": 916, "estimated_eval_tokens": 2088 }, @@ -19793,28 +19832,25 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests target the same runtime TypeAdapter NameError in `src/transformers/cli/serve.py` and are plausibly mergeable as one fix, despite using slightly different annotation/import strategies.", - "confidence": 0.86, + "summary": "These are related CLI type-annotation fixes in the same file, but they address different runtime failures: one broad annotation/forward-reference cleanup, the other a specific `TypeAdapter` import/annotation issue when `pydantic` is absent. They should not be treated as duplicates.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43842 is the more direct and minimal fix for the same bug: it makes `TypeAdapter` a TYPE_CHECKING-only import and quotes the annotation, which cleanly removes the runtime NameError path.", + "canonical_pr_reason": "PR #43842 is the more precise, concrete fix: it directly addresses the `TypeAdapter` NameError path by moving the import under `TYPE_CHECKING` and quoting the annotation. Its scope is narrower and more clearly aligned with the stated bug than #43656.", "best_issue_reason": null, - "best_pr_reason": "PR 43842 is the best standalone representative of the cluster because it targets the concrete import-time failure most explicitly and with the smallest surface-area change.", + "best_pr_reason": "PR #43842 is the best representative for this cluster because it cleanly describes a specific runtime NameError scenario and has a tight, self-contained code change. #43656 appears to be a broader annotation-related cleanup in the same module rather than the same exact bug.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", - "accept": true, - "reason": "Both PRs address the same underlying TypeAdapter NameError in the CLI serve module, with overlapping file scope and compatible fixes to annotation evaluation/import handling." + "accept": false, + "reason": "Same file and broadly similar theme, but the fixes are not the same underlying bug. #43656 is about annotation/forward-reference cleanup, while #43842 specifically fixes `TypeAdapter` import handling when `pydantic` is not installed. They would not plausibly be merged as one PR." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet: both PRs modify `src/transformers/cli/serve.py`, both are explicitly about a `TypeAdapter` NameError in the CLI, and the fixes are compatible enough to be merged into one PR for the same bug. The soft-edge verdict is conservative, and the conclusion does 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 }, @@ -19953,8 +19989,8 @@ "pull_request:44071|pull_request:44085", "pull_request:44059|pull_request:44086", "pull_request:44085|pull_request:44086", - "pull_request:44059|pull_request:44722", "pull_request: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", @@ -19999,322 +20035,292 @@ "pull_request:44044|pull_request:44129", "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "2443d3e4151aca65cb5c0bf332b7a00f649710315989a41657988726c564a316", + "prepared_review_unit_hash": "543940efd02bbd59903612f7e79baa212d62f122e4cb7c8de4d48e0128954719", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84900, - "estimated_input_tokens": 21225, - "estimated_eval_tokens": 42706 + "serialized_chars": 84888, + "estimated_input_tokens": 21222, + "estimated_eval_tokens": 42700 }, "budget": { "node_count": 27, "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 75098, - "estimated_input_tokens": 18775, - "estimated_eval_tokens": 37806 + "serialized_chars": 75083, + "estimated_input_tokens": 18771, + "estimated_eval_tokens": 37798 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of model-specific output-tracing refactors, not a single broad duplicate. The only strong duplicate families are the GPT-Neo pair and the GPT-J/CodeGen/GPT-J group around modeling_gptj.py; the rest are the same pattern applied to different model code paths.", - "confidence": 0.9, - "canonical_issue_reason": null, - "canonical_pr_reason": "44066 is the best representative PR: it is a concrete GPT-J/CodeGen output-tracing refactor, explicitly targets the shared tracking issue, and sits at the center of the strongest duplicate links in the cluster.", - "best_issue_reason": null, - "best_pr_reason": "44066 is the best single PR to keep as a representative because it spans the broadest concrete instance of the shared output-tracing refactor and connects directly to the strongest GPT-J/GPT-J+CodeGen duplicates.", + "summary": "This cluster is mostly a set of model-specific refactors to standardized output tracing/capture_outputs, not one single bug fix. The only clear duplicate-style matches are within the same concrete code paths (notably GPT-J/CodeGen and GPT-Neo), while most cross-model pairs should stay separate.", + "confidence": 0.83, + "canonical_issue_reason": "Issue #43979 is the shared upstream tracker for this output-tracing refactor work; the PRs in this cluster largely point back to it.", + "canonical_pr_reason": "PR #44722 is the best canonical representative: it is the latest, most complete GPT-J/CodeGen variant and appears to supersede earlier near-identical attempts on the same code path.", + "best_issue_reason": "No standalone issue artifact is present in the cluster; #43979 is the only common issue target and is the appropriate umbrella issue.", + "best_pr_reason": "PR #44722 is the strongest single PR in the cluster because it addresses the same concrete GPT-J/CodeGen tracing implementation as earlier duplicates, with the most evolved patch among them.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Different model families and code paths; they only share the general output-tracing refactor pattern." + "reason": "Different models and code paths; same umbrella refactor theme is not enough to merge them." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "Same standardized tracing theme, but the CVT/FNet and DeBERTa implementations are unrelated concrete changes." + "reason": "Both are output-tracing refactors, but they touch different model implementations and are not the same concrete fix." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both touch GPT-J output tracing in the same file family; 44066 is broader because it also folds in CodeGen, but the GPT-J change is the same underlying refactor." + "reason": "Both edit GPT-J output-tracing logic in the same file and look like the same underlying refactor." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "RegNet and EfficientNet are different model implementations; shared issue target alone is not enough." + "reason": "Different model families and distinct implementations; only the general pattern matches." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet and GPT-J/CodeGen are unrelated model-specific changes despite the same refactor theme." + "reason": "EfficientNet and GPT-J/CodeGen are unrelated code paths, so these are not duplicates." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "GPT-J/CodeGen and EfficientNet are different code paths and not the same underlying fix." + "reason": "GPT-J/CodeGen versus EfficientNet are separate model refactors, not one shared change." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are separate model refactors; only the tracing pattern overlaps." + "reason": "MobileNetV2 and DeBERTa V2 are separate model-specific output-tracing changes." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "Different model implementations and different file-level changes; not the same concrete bug or patch." + "reason": "GPT-J/CodeGen and MGP-STR are unrelated implementations." }, { "left": "pull_request:44066", "right": "pull_request:44071", "accept": false, - "reason": "GPT-J/CodeGen and MPT are distinct model-specific refactors." + "reason": "GPT-J/CodeGen and MPT are different code paths, despite similar decorator migration." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Same GPT-Neo output-tracing refactor on the same model/file family; these look like duplicate revisions of the same change." + "reason": "Both are GPT-Neo tracing refactors on the same file and appear to implement the same change." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different model code paths; same issue target is not enough." + "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are separate model refactors, not the same change." + "reason": "GPT-J/CodeGen and GPT-Neo are distinct refactors for different models." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are unrelated concrete changes." + "reason": "ResNet and GPT-J/CodeGen are unrelated code paths." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are different model implementations; same tracing style only." + "reason": "MGP-STR and GPT-J/CodeGen are different model changes, not a single fix." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet and GPT-J are different model families and code paths." + "reason": "ResNet and GPT-J are different implementations; same issue target does not make them duplicates." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are unrelated concrete refactors." + "reason": "GPT-Neo and ResNet are unrelated model refactors." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are different model implementations." + "reason": "ResNet and MPT are separate code paths." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert/XLM are separate model refactors." + "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." }, { "left": "pull_request:44018", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert/XLM are unrelated implementations." + "reason": "GPT-Neo and Flaubert/XLM are different model-specific changes." }, { "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "GPT-Neo and GPT-J/CodeGen are different model code paths." + "reason": "GPT-Neo and GPT-J/CodeGen are not the same underlying code-path fix." }, { "left": "pull_request:44071", "right": "pull_request:44722", "accept": false, - "reason": "MPT and GPT-J/CodeGen are distinct model-specific changes." + "reason": "MPT and GPT-J/CodeGen are different model refactors." }, { "left": "pull_request:44018", "right": "pull_request:44085", "accept": false, - "reason": "GPT-Neo and GPT-J are different model families." + "reason": "GPT-Neo and GPT-J are different concrete implementations." }, { "left": "pull_request:44018", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR are unrelated model changes." + "reason": "GPT-Neo and MGP-STR are unrelated changes." }, { "left": "pull_request:44068", "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo and MGP-STR are different concrete changes." + "reason": "GPT-Neo and MGP-STR are separate model paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT and GPT-J are distinct model implementations." }, { "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "GPT2 and MGP-STR are unrelated code paths." + "reason": "GPT-2 and MGP-STR are unrelated beyond sharing a decorator migration theme." }, { "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J and MGP-STR are different model implementations." + "reason": "GPT-J and MGP-STR do not share the same code path." }, { "left": "pull_request:44059", "right": "pull_request:44722", "accept": false, - "reason": "GPT2 and GPT-J/CodeGen are separate model refactors." - }, - { - "left": "pull_request:44098", - "right": "pull_request:44101", - "accept": false, - "reason": "ViLT and Flaubert/XLM are different model families." + "reason": "GPT-2 and GPT-J/CodeGen are different concrete refactors." }, { "left": "pull_request:44018", "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and MPT are unrelated model-specific changes." + "reason": "GPT-Neo and MPT are separate model implementations." }, { "left": "pull_request:44017", "right": "pull_request:44722", "accept": false, - "reason": "SegFormer and GPT-J/CodeGen are unrelated model refactors." + "reason": "SegFormer and GPT-J/CodeGen are unrelated model changes." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "SegFormer and LongT5 do not fix the same underlying code-path problem." - }, - { - "left": "pull_request:44017", - "right": "pull_request:44073", - "accept": false, - "reason": "SegFormer and VisualBert are different model implementations." - }, - { - "left": "pull_request:44068", - "right": "pull_request:44071", - "accept": false, - "reason": "GPT-Neo and MPT are separate model changes." - }, - { - "left": "pull_request:44071", - "right": "pull_request:44086", - "accept": false, - "reason": "MPT and MGP-STR are unrelated concrete changes." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44073", - "accept": false, - "reason": "RegNet and VisualBert are different model families." + "reason": "SegFormer and LongT5 are distinct implementations." }, { "left": "pull_request:44056", "right": "pull_request:44101", "accept": false, - "reason": "MPNet and Flaubert/XLM are unrelated implementations." + "reason": "MPNet and Flaubert/XLM are separate model-specific changes." }, { "left": "pull_request:44059", "right": "pull_request:44071", "accept": false, - "reason": "GPT2 and MPT are separate code paths." + "reason": "GPT-2 and MPT are not the same concrete fix." }, { - "left": "pull_request:44018", - "right": "pull_request:44059", + "left": "pull_request:44098", + "right": "pull_request:44101", "accept": false, - "reason": "GPT-Neo and GPT2 are different model refactors." + "reason": "ViLT and Flaubert/XLM are different model paths." }, { "left": "pull_request:44073", "right": "pull_request:44161", "accept": false, - "reason": "VisualBert and LongT5 are unrelated concrete changes." + "reason": "VisualBert and LongT5 are unrelated implementations." }, { "left": "pull_request:44056", - "right": "pull_request:44068", - "accept": false, - "reason": "MPNet and GPT-Neo are different model implementations." - }, - { - "left": "pull_request:44101", "right": "pull_request:44161", "accept": false, - "reason": "Flaubert/XLM and LongT5 do not share the same underlying fix." + "reason": "MPNet and LongT5 are separate model refactors." }, { - "left": "pull_request:44017", - "right": "pull_request:44018", + "left": "pull_request:43996", + "right": "pull_request:44017", "accept": false, - "reason": "SegFormer and GPT-Neo are unrelated model-specific changes." + "reason": "CVT/FNet and SegFormer are different model implementations." }, { - "left": "pull_request:43997", - "right": "pull_request:44056", + "left": "pull_request:43996", + "right": "pull_request:44161", "accept": false, - "reason": "RegNet and MPNet are different implementations; shared tracing theme only." + "reason": "CVT/FNet and LongT5 are unrelated code paths." }, { - "left": "pull_request:44071", - "right": "pull_request:44076", + "left": "pull_request:44017", + "right": "pull_request:44073", "accept": false, - "reason": "MPT and ImageGPT are unrelated concrete changes." + "reason": "SegFormer and VisualBert are different model-specific changes." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", + "left": "pull_request:44010", + "right": "pull_request:44071", "accept": false, - "reason": "TextNet and Flaubert/XLM are different model families." + "reason": "SqueezeBert and MPT are unrelated implementations." }, { "left": "pull_request:43997", - "right": "pull_request:44018", + "right": "pull_request:44074", "accept": false, - "reason": "RegNet and GPT-Neo are separate model refactors." + "reason": "RegNet and TextNet are different model refactors." }, { - "left": "pull_request:44056", - "right": "pull_request:44161", + "left": "pull_request:43997", + "right": "pull_request:44018", "accept": false, - "reason": "MPNet and LongT5 are unrelated code paths." + "reason": "RegNet and GPT-Neo are unrelated code paths." }, { "left": "pull_request:43997", "right": "pull_request:44161", "accept": false, - "reason": "RegNet and LongT5 do not represent the same underlying bug or change." + "reason": "RegNet and LongT5 are separate model changes." }, { "left": "pull_request:44074", @@ -20323,133 +20329,91 @@ "reason": "TextNet and LongT5 are different implementations." }, { - "left": "pull_request:44018", - "right": "pull_request:44161", - "accept": false, - "reason": "GPT-Neo and LongT5 are unrelated model changes." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44071", - "accept": false, - "reason": "RegNet and MPT are separate code paths." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44071", - "accept": false, - "reason": "SqueezeBert and MPT are unrelated model-specific refactors." - }, - { - "left": "pull_request:44018", + "left": "pull_request:44071", "right": "pull_request:44129", "accept": false, - "reason": "GPT-Neo and SpeechT5 are different model families." + "reason": "MPT and SpeechT5 are unrelated code paths." }, { "left": "pull_request:44010", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and VisualBert are unrelated implementations." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44074", - "accept": false, - "reason": "RegNet and TextNet are different concrete changes." - }, - { - "left": "pull_request:44071", - "right": "pull_request:44129", - "accept": false, - "reason": "MPT and SpeechT5 are unrelated code paths." + "reason": "SqueezeBert and VisualBert are separate model-specific refactors." }, { "left": "pull_request:44010", "right": "pull_request:44018", "accept": false, - "reason": "SqueezeBert and GPT-Neo are separate model refactors." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44129", - "accept": false, - "reason": "RegNet and SpeechT5 are different model implementations." + "reason": "SqueezeBert and GPT-Neo are different implementations." }, { - "left": "pull_request:43996", - "right": "pull_request:44161", - "accept": false, - "reason": "CVT/FNet and LongT5 are unrelated concrete changes." - }, - { - "left": "pull_request:43996", - "right": "pull_request:44017", + "left": "pull_request:44010", + "right": "pull_request:44044", "accept": false, - "reason": "CVT/FNet and SegFormer are separate model-specific refactors." + "reason": "SqueezeBert and DeBERTa V2 are unrelated model changes." }, { - "left": "pull_request:44010", - "right": "pull_request:44044", + "left": "pull_request:44044", + "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and DeBERTa are different code paths." + "reason": "DeBERTa V2 and SpeechT5 are different code paths." }, { "left": "pull_request:44044", - "right": "pull_request:44074", + "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and TextNet are unrelated model changes." + "reason": "DeBERTa V2 and VITS are unrelated model refactors." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Same GPT-J/CodeGen output-tracing refactor with the same concrete file family and issue target; these are plausibly mergeable as one PR." + "reason": "This is the same GPT-J/CodeGen output-tracing change as the later PR, just an earlier version." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Same GPT-J output-tracing refactor in modeling_gptj.py; 44085 and 44722 look like duplicate revisions of the same concrete change." + "reason": "Both appear to be the same GPT-J output-tracing refactor; one is an earlier duplicate of the other." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated implementations." + "reason": "GPT-J and SpeechT5 are different model implementations." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "CVT/FNet and RegNet/ResNet are different model refactors." + "reason": "CVT/FNet and RegNet/ResNet are unrelated model refactors." }, { "left": "pull_request:43996", "right": "pull_request:44129", "accept": false, - "reason": "CVT/FNet and SpeechT5 are separate concrete changes." + "reason": "CVT/FNet and SpeechT5 are different concrete changes." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and SpeechT5 are unrelated model-specific refactors." + "reason": "DeBERTa V2 and SpeechT5 are separate model-specific changes." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and VITS are different models; only the decorator pattern is shared." + "reason": "DeBERTa V2 and VITS are unrelated implementations." } ] }, "evaluator_result": { - "accept": false, - "feedback": "Reject: the cluster-level summary is broadly plausible, but several soft-edge rationales are not grounded in the packet. In particular, 44085 is a GPT-J PR in the packet, not GPT-Neo, so the explanations for 44085/44722, 44018/44085, and 44085/44129 misidentify the model family. Also, 44085 and 44722 share src/transformers/models/gptj/modeling_gptj.py and the packet flags that pair very strongly, so dismissing it as just a different model family is too strong. The accepted GPT-Neo pair and GPT-J/CodeGen pair are plausible, but the remaining verdicts need more conservative, packet-faithful wording." + "accept": true, + "feedback": "The clustering is grounded in the packet: it correctly treats this as a set of model-specific output-tracing refactors rather than one unified bugfix, and the canonical issue/PR choices are supported by the shared issue target and the later GPT-J/CodeGen patch. The soft-edge verdicts are conservative overall, and the accepted pairs are same-code-path duplicates that look mergeable into a single PR for the same fix." }, "evaluator_used": true, - "retried": true, + "retried": false, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -20484,70 +20448,70 @@ "pull_request:43997|pull_request:44010", "pull_request:44010|pull_request:44129" ], - "prepared_review_unit_hash": "05a215f9d2617a2fdcc016d057aa8d0826eaa781bf5142e487b48a88e54a245d", + "prepared_review_unit_hash": "079de8624b438fcca4070b100d5bf82c0693a9f5199430662af63b9ec6208e1e", "component_budget": { "node_count": 28, "item_count": 28, "soft_pair_count": 78, - "serialized_chars": 84900, - "estimated_input_tokens": 21225, - "estimated_eval_tokens": 42706 + "serialized_chars": 84888, + "estimated_input_tokens": 21222, + "estimated_eval_tokens": 42700 }, "budget": { "node_count": 8, "item_count": 8, "soft_pair_count": 6, - "serialized_chars": 17123, - "estimated_input_tokens": 4281, - "estimated_eval_tokens": 8818 + "serialized_chars": 17126, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All items are PRs for model-specific refactors to standardized output capture/tracing. They share a theme and similar helper imports, but they touch different model codepaths, so they should not be collapsed as duplicates.", - "confidence": 0.83, + "summary": "All items are model-specific refactors around standardized output tracing/capture_outputs, but they affect different architectures and code paths. They are thematically similar, not duplicates of the same underlying change.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44129 is the strongest single representative because it has the clearest scoped refactor, an explicit issue target, and concrete output-tracing changes in a model implementation.", + "canonical_pr_reason": "PR 44073 is the clearest representative of the cluster: it directly targets standardized output tracing/capture_outputs and shows the core decorator + recorder pattern in a typical model implementation.", "best_issue_reason": null, - "best_pr_reason": "PR 44129 best fits the cluster as a representative example of the standardized output-capture migration pattern while also tying back to a tracking issue.", + "best_pr_reason": "PR 44073 is the best global representative because its title and diff most clearly express the shared refactor pattern without being overly narrow or model-quirky.", "soft_edge_verdicts": [ { "left": "pull_request:44024", "right": "pull_request:44076", "accept": false, - "reason": "Same refactor theme, but FocalNet and ImageGPT are different model codepaths; these would not be merged into one PR." + "reason": "Both are output-tracing refactors, but for different models (FocalNet vs ImageGPT) and different implementation details; not the same concrete change." }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Both migrate output tracing, but they affect different models (VisualBert vs TextNet) and are separate changes rather than one underlying bug." + "reason": "Same general standardized-output migration, but VisualBert and TextNet are separate model implementations with distinct refactors." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "MPNet and VisualBert are distinct implementations; the similarity is only the generic output-tracing refactor pattern." + "reason": "Both touch output capture plumbing, but MPNet and VisualBert are unrelated model code paths and cannot plausibly be merged into one PR." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "Different model-specific refactors with no shared concrete codepath, so they are not the same change." + "reason": "Shared theme only; MPNet and TextNet are different models and the changes are not the same bug or fix." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "RegNet and SqueezeBert are unrelated model internals; same broad migration goal is not enough to treat them as duplicates." + "reason": "RegNet and SqueezeBert are separate models; these are parallel refactors, not duplicate fixes." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 are different codepaths and the PRs address separate model migrations, not one mergeable fix." + "reason": "Both are standardized output refactors, but SqueezeBert vs SpeechT5 targets different code paths and concrete behavior." } ] }, @@ -20632,12 +20596,12 @@ "pull_request:44028|pull_request:46438", "pull_request:43998|pull_request:46438" ], - "prepared_review_unit_hash": "3a5bed797fa8a3d9087609f9d6f8c4dc53a35d9add958831890264afa15bbf66", + "prepared_review_unit_hash": "4c9f3b49f4870ebdac8dae942e6c0efc9f1dd1052b11e375f68bb6f26aaafc29", "component_budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 44, - "serialized_chars": 43573, + "serialized_chars": 43575, "estimated_input_tokens": 10894, "estimated_eval_tokens": 22044 }, @@ -20645,7 +20609,7 @@ "node_count": 14, "item_count": 14, "soft_pair_count": 44, - "serialized_chars": 43573, + "serialized_chars": 43575, "estimated_input_tokens": 10894, "estimated_eval_tokens": 22044 }, @@ -20654,276 +20618,276 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are mostly separate model-specific PRs that share a common output-tracing refactor pattern, not duplicate fixes for the same code path. The modular-conversion speedup PR is unrelated. I would not merge any soft pair.", + "summary": "These PRs are thematically similar output-tracing refactors, but they target different model families and code paths, so they are not duplicates of the same underlying change.", "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical duplicate: the PRs are independent model-specific refactors across different forward paths, so there is no one PR that represents the same underlying change for the whole cluster.", + "canonical_pr_reason": "No issue is present; 44003 is the best representative PR because it applies the tracing refactor across two closely related Mamba-family files and shows the clearest use of the new helper infrastructure.", "best_issue_reason": null, - "best_pr_reason": "PR 44025 is the best representative of the shared refactor theme because it covers two related files and has the broadest scope among the model PRs, but it still is not a duplicate of the others.", + "best_pr_reason": "44003 is the strongest single PR in the set to anchor the cluster, but it should remain a representative example rather than absorbing the others as duplicates.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different model code paths (timm_backbone vs upernet); same refactor style, not the same underlying fix." + "reason": "Different models (timm_backbone vs upernet); same refactor pattern only." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different model families; both touch output handling, but they are separate forward-path refactors." + "reason": "timm_backbone vs vision_text_dual_encoder; unrelated code paths." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "Different architectures and code paths; similar mechanical change only." + "reason": "timm_backbone vs univnet; different model families and behavior." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "timm_backbone vs mamba/falcon_mamba are unrelated implementations, so this is not one duplicate bug." + "reason": "timm_backbone vs mamba/falcon_mamba; same tracing theme, not the same change." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Different model stacks; both are output-tracing refactors, but not the same concrete change." + "reason": "timm_backbone vs rwkv; distinct architectures and implementations." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "MobileNetV1 and UnivNet are unrelated model paths; same refactor pattern only." + "reason": "MobileNetV1 vs UnivNet; different model code paths." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different subsystems and forward methods; not the same underlying issue." + "reason": "MobileNetV1 vs UperNet; only the refactor style matches." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "MobileNetV1 vs Depth Anything are separate model-specific changes, not one mergeable fix." + "reason": "MobileNetV1 vs DepthAnything; separate models, not one underlying bug." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different encoder/decoder wrappers and distinct code paths; too broad to treat as duplicates." + "reason": "MobileNetV1 vs VisionEncoderDecoder; distinct implementations." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "MobileNetV1 and Speech Encoder-Decoder are unrelated models; same pattern only." + "reason": "MobileNetV1 vs SpeechEncoderDecoder; related naming only." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Vision-text dual encoder and UnivNet fix different forward implementations; not the same bug." + "reason": "vision_text_dual_encoder vs univnet; different model families." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "Different models and call paths; similar output-tracing cleanup only." + "reason": "vision_text_dual_encoder vs upernet; same tracing motif, different code." }, { "left": "pull_request:44000", "right": "pull_request:44003", "accept": false, - "reason": "Vision-text dual encoder vs mamba/codegen family are separate concrete changes." + "reason": "vision_text_dual_encoder vs mamba/falcon_mamba; not the same concrete change." }, { "left": "pull_request:44000", "right": "pull_request:44025", "accept": false, - "reason": "Different model families and different output plumbing; not a single duplicate PR." + "reason": "vision_text_dual_encoder vs depth_anything; separate model targets." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "UnivNet and UperNet are unrelated models; same refactor style, not same underlying fix." + "reason": "UnivNet vs UperNet; unrelated subsystems." }, { "left": "pull_request:44001", "right": "pull_request:44003", "accept": false, - "reason": "Separate code paths in unrelated models; do not merge as duplicates." + "reason": "UnivNet vs mamba/falcon_mamba; not one shared fix." }, { "left": "pull_request:44001", "right": "pull_request:44025", "accept": false, - "reason": "UnivNet vs Depth Anything are distinct model-specific output changes." + "reason": "UnivNet vs DepthAnything; different models and outputs." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "UperNet and RWKV are different model implementations; not one concrete bug." + "reason": "UperNet vs RWKV; distinct model families." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "Different model families and forward paths; only the refactor motif matches." + "reason": "UperNet vs mamba/falcon_mamba; only similar refactoring." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "Mamba/Falcon-Mamba vs CodeGen are separate code-path changes, not the same bug fix." + "reason": "Mamba/FalconMamba vs CodeGen; different code paths and semantics." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "Timm backbone and CodeGen are unrelated models; shared output-tracing theme is too broad." + "reason": "timm_backbone vs CodeGen; unrelated models." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "UnivNet and CodeGen touch unrelated forward logic; not duplicates." + "reason": "UnivNet vs CodeGen; separate architectures." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "UperNet and CodeGen are different model paths with only superficial similarity." + "reason": "UperNet vs CodeGen; not the same underlying change." }, { "left": "pull_request:44000", "right": "pull_request:44004", "accept": false, - "reason": "Different model families and different concrete forward signatures; not the same change." + "reason": "vision_text_dual_encoder vs CodeGen; different model families." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "Depth Anything and Vision-Encoder-Decoder are separate model wrappers, not one mergeable fix." + "reason": "DepthAnything vs VisionEncoderDecoder; distinct refactors on different models." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "Different model families and code paths; same broad refactor pattern only." + "reason": "DepthAnything vs SpeechEncoderDecoder; separate code paths." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "Depth Anything and SuperPoint are unrelated models, so this is not a duplicate PR." + "reason": "DepthAnything vs SuperPoint; unrelated implementations." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "Different models and different output APIs; not the same underlying change." + "reason": "DepthAnything vs RWKV; different model families." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "Depth Anything and DPR are separate model-specific refactors." + "reason": "DepthAnything vs DPR; not the same concrete bug/change." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "Vision-Encoder-Decoder and RWKV are unrelated forward paths; same theme only." + "reason": "VisionEncoderDecoder vs RWKV; unrelated subsystems." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "Different encoder-decoder wrappers with different implementations; not a duplicate." + "reason": "VisionEncoderDecoder vs SpeechEncoderDecoder; similar wrapper refactor only." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "Vision-Encoder-Decoder and SuperPoint are different model code paths." + "reason": "VisionEncoderDecoder vs SuperPoint; different model stacks." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "Different model families; this is a broad similarity, not the same fix." + "reason": "VisionEncoderDecoder vs DPR; separate code paths." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "Speech Encoder-Decoder and SuperPoint are unrelated models and cannot be merged as one PR." + "reason": "SpeechEncoderDecoder vs SuperPoint; distinct model implementations." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "Different concrete code paths in different models; same refactor motif only." + "reason": "SpeechEncoderDecoder vs RWKV; unrelated models." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "Speech Encoder-Decoder and DPR are separate model-specific changes." + "reason": "SpeechEncoderDecoder vs DPR; different functionality." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "SuperPoint and RWKV are unrelated implementations; not the same underlying bug." + "reason": "SuperPoint vs RWKV; no shared underlying bug." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "SuperPoint and DPR touch distinct forward paths; not duplicates." + "reason": "SuperPoint vs DPR; separate code paths." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "RWKV and DPR are different model families; same broad refactor pattern only." + "reason": "RWKV vs DPR; unrelated model families." }, { "left": "pull_request:44000", "right": "pull_request:46438", "accept": false, - "reason": "Vision-text dual encoder output tracing vs modular-conversion pool reuse are unrelated changes." + "reason": "Model refactor PR vs modular-conversion tooling; completely different domain." }, { "left": "pull_request:44027", "right": "pull_request:46438", "accept": false, - "reason": "Speech Encoder-Decoder refactor and modular-conversion speedup do not share a code path or bug." + "reason": "SpeechEncoderDecoder model change vs conversion script optimization; unrelated." }, { "left": "pull_request:44029", "right": "pull_request:46438", "accept": false, - "reason": "RWKV output tracing and modular-conversion pool reuse are unrelated." + "reason": "RWKV model refactor vs build/tooling speedup; not the same change." }, { "left": "pull_request:44028", "right": "pull_request:46438", "accept": false, - "reason": "SuperPoint output refactor vs modular-conversion utility change are different concerns." + "reason": "SuperPoint model refactor vs modular conversion pool reuse; different areas." }, { "left": "pull_request:43998", "right": "pull_request:46438", "accept": false, - "reason": "timm_backbone model refactor is unrelated to the modular-conversion process-pool optimization." + "reason": "timm_backbone model refactor vs tooling optimization; unrelated." } ] }, @@ -20953,12 +20917,12 @@ "soft_pairs": [ "pull_request:44662|pull_request:45401" ], - "prepared_review_unit_hash": "5b6d041a8d6c1f3729533b18a10322c4d85aa9b4da6d422a419d6b30de0f454c", + "prepared_review_unit_hash": "ed413d7592f76f70127184f39671161c7ffc6b3409d570471205d790eac9865a", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6552, + "serialized_chars": 6551, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -20966,7 +20930,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 6552, + "serialized_chars": 6551, "estimated_input_tokens": 1638, "estimated_eval_tokens": 3532 }, @@ -20975,18 +20939,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are unrelated model additions: one implements PenguinVL, the other adds Voxtral TTS. Their overlap is limited to shared Transformers plumbing and docs/auto-registry updates, which is incidental rather than a shared underlying change.", + "summary": "These are two separate model-addition PRs (PenguinVL vs Voxtral TTS). They share framework/auto-registration files, but they target different models and code paths, so they should not be clustered as duplicates.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45401 is the better representative because it is tightly scoped, explicitly tied to issue 45027, and clearly describes one concrete model integration.", + "canonical_pr_reason": "PR 44662 is the stronger cluster representative: it is a fuller, more mature implementation with substantial review activity and is not a draft. PR 45401 is a separate Voxtral TTS addition, not the same change.", "best_issue_reason": null, - "best_pr_reason": "PR #45401 is the strongest standalone PR in the cluster: it has a specific tracking issue, a focused code path, and a clearer change objective than the broader PenguinVL implementation PR.", + "best_pr_reason": "PR 44662 is the best overall PR in this set as a representative artifact, but only because it is the more developed standalone model implementation; it is not a duplicate of PR 45401.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "Different models and different feature work: PenguinVL vs Voxtral TTS. Shared framework files are expected for any model integration and do not indicate the same underlying bug or change." + "reason": "Different underlying features: PenguinVL implementation vs Voxtral TTS support. Shared registration/docs files are generic framework plumbing, not evidence of the same bug or change." } ] }, @@ -21020,58 +20984,58 @@ "pull_request:45363|pull_request:46339", "pull_request:45666|pull_request:46339" ], - "prepared_review_unit_hash": "05f72207339b41a1f225af945539ffb09dd500b4b611bc709af83d6aeaa65617", + "prepared_review_unit_hash": "d002d9b9d668bd7c75fe82ca7225f7c277eee45841f0121ed5e496efc52604e6", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 9635, - "estimated_input_tokens": 2409, - "estimated_eval_tokens": 5074 + "serialized_chars": 9632, + "estimated_input_tokens": 2408, + "estimated_eval_tokens": 5072 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 3, - "serialized_chars": 9635, - "estimated_input_tokens": 2409, - "estimated_eval_tokens": 5074 + "serialized_chars": 9632, + "estimated_input_tokens": 2408, + "estimated_eval_tokens": 5072 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster is centered on the same broader feature line: n-to-1 kernel fusion via `KernelConfig`, with one adjacent but distinct earlier module-fusion API PR. The latest PR is the best representative of the shared kernel-fusion work.", - "confidence": 0.88, + "summary": "These PRs form a related feature line around KernelConfig-based n-to-1 kernel fusion, but only the closest pair looks like the same concrete change path. The module_fusion API PR is foundational, not a duplicate of the KernelConfig fusion work.", + "confidence": 0.61, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46339 is the most complete and most recent iteration of the kernel-fusion feature: it keeps the same `KernelConfig` direction as the earlier PRs while broadening and simplifying the implementation across core loading, hub kernel integration, and tests.", + "canonical_pr_reason": "PR 46339 is the broadest and most mature representative of the KernelConfig fusion line: it touches core loading, hub kernel integration, model loading, and tests, and its title suggests a later simplification/extension of the same mechanism.", "best_issue_reason": null, - "best_pr_reason": "PR 46339 is the strongest global representative because it appears to subsume the earlier `KernelConfig` fusion work and presents the most polished, integrated version of that feature.", + "best_pr_reason": "PR 46339 is the best single representative for the cluster because it appears to subsume earlier KernelConfig fusion work and exercises the same kernel-loading path most comprehensively.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Related theme, but not the same underlying change: 44979 introduces a separate module-fusion API, while 45363 implements n-to-1 kernel fusion through `KernelConfig` and different integration points." + "reason": "44979 introduces the Module Fusion API scaffolding, while 45363 implements n-to-1 kernel fusion via KernelConfig. They are related, but not the same underlying change." }, { "left": "pull_request:45363", "right": "pull_request:46339", - "accept": true, - "reason": "Same concrete feature area and code path: both implement n-to-1 kernel fusion via `KernelConfig`, with 46339 clearly extending/simplifying the same mechanism rather than starting a different change." + "accept": false, + "reason": "Both are about KernelConfig fusion, but 46339 is a later extension/simplification that changes the core loading path and does not clearly duplicate the exact feature change in 45363." }, { "left": "pull_request:45666", "right": "pull_request:46339", "accept": true, - "reason": "These appear to be successive revisions of the same `KernelConfig`-based n-to-1 kernel fusion work, touching the same integration path and underlying feature rather than distinct changes." + "reason": "These look like the same concrete n-to-1 KernelConfig fusion change family, with heavy overlap in kernel-loading files; 46339 reads like a follow-on refinement of the same implementation path." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative about the one adjacent PR. The soft-edge decisions are plausible: 44979 is reasonably treated as distinct module-fusion work, while 45363/45666/46339 all look like closely related KernelConfig-based n-to-1 kernel-fusion revisions with overlapping files and matching titles." + "feedback": "The summary stays within the evidence in the packet, and the soft-edge calls are appropriately cautious overall. The only potentially close pair, 45666 vs 46339, is supported by the shared KernelConfig / n-to-1 fusion theme and overlapping kernel-loading files, while 44979 vs 45363 is clearly more foundational than duplicate. The representative PR choice is also grounded in the broader file coverage and the later 'extended & simplified' title of 46339." }, "evaluator_used": true, "retried": false, @@ -21098,12 +21062,12 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "7831c18ca81e5a253435463dbee041d20542945e35837018264ecfb315065405", + "prepared_review_unit_hash": "2256876bb6709f19d139d1af2390bd7e8c7606af4b76007cf107e64baf2b9dbf", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4413, + "serialized_chars": 4414, "estimated_input_tokens": 1104, "estimated_eval_tokens": 2464 }, @@ -21111,7 +21075,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4413, + "serialized_chars": 4414, "estimated_input_tokens": 1104, "estimated_eval_tokens": 2464 }, @@ -21120,18 +21084,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both pull requests target the same AttributeError in `_patch_mistral_regex` for the Mistral tokenizer path, with the same code change in `tokenization_utils_tokenizers.py`; the second PR adds tests but is still the same underlying fix.", + "summary": "Both PRs address the same Mistral regex patching AttributeError in the same code path and target the same issue; the later PR adds test coverage but does not change the underlying fix.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the better canonical PR because it has the same fix as 45086 plus an added test file, making it the more complete change set.", + "canonical_pr_reason": "PR #45317 is the better canonical PR because it contains the same fix as #45086 plus tests, making it the more complete implementation.", "best_issue_reason": null, - "best_pr_reason": "PR 45317 is the best representative for the cluster since it contains the core bug fix and accompanying regression coverage.", + "best_pr_reason": "PR #45317 is the strongest standalone candidate: same bug fix, same target issue, and it includes validation coverage in tests.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same concrete bug and same fix in the same code path (`_patch_mistral_regex`); the extra test changes in 45317 do not make it a different change." + "reason": "Same underlying AttributeError fix in _patch_mistral_regex, same file and issue target; the second PR is a superset with tests, so they are mergeable as one change." } ] }, @@ -21161,7 +21125,7 @@ "soft_pairs": [ "pull_request:45115|pull_request:45569" ], - "prepared_review_unit_hash": "e6eccea45fa0d8bcf221679f26d08b121a3a50f793492ac37b56eab2ccee6b2a", + "prepared_review_unit_hash": "38307df2659058477dfb0cd2b5fc856796fb1fe8e60d01e128b3ac350f2117e5", "component_budget": { "node_count": 2, "item_count": 2, @@ -21183,18 +21147,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open NemotronH PRs overlap in the same model area, but they target different changes: one is a refactor/inheritance and conversion-mapping cleanup, while the other is a broader split into nemotron_h_dense/nemotron_h_sparse with docs and auto-mapping updates.", - "confidence": 0.88, + "summary": "Two related but distinct NemotronH refactors: one focuses on inheriting GraniteMoeHybrid-style weight/config plumbing, the other splits NemotronH into dense/sparse model families with docs and auto-mapping updates. They overlap in shared NemotronH files but do not look like the same mergeable code change.", + "confidence": 0.86, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45569 is the broader, more complete NemotronH implementation change set, covering the dense/sparse model split, docs, and auto/model registration updates, whereas #45115 is a narrower refactor centered on inheritance and conversion mapping.", + "canonical_pr_reason": "PR 45569 is the broader and more complete NemotronH architecture change, covering the new dense/sparse model split plus docs and auto-mapping updates; PR 45115 is a narrower inheritance/refactor step.", "best_issue_reason": null, - "best_pr_reason": "PR #45569 is the best representative of the cluster because it introduces the substantive user-facing model packaging change; #45115 looks like a supporting refactor rather than the primary feature implementation.", + "best_pr_reason": "PR 45569 is the best representative of the cluster because it addresses the higher-level NemotronH redesign end-to-end, while PR 45115 is a partial implementation focused on conversion and inheritance plumbing.", "soft_edge_verdicts": [ { "left": "pull_request:45115", "right": "pull_request:45569", "accept": false, - "reason": "Related area, but not the same underlying change: #45115 focuses on refactoring NemotronH to inherit GraniteMoeHybrid-compatible structure and conversion mappings, while #45569 splits NemotronH into dense/sparse variants and updates docs/auto-mappings. They touch shared files, but the concrete code-path change is different and not something to merge as one PR." + "reason": "They share NemotronH infrastructure files, but one is a GraniteMoeHybrid inheritance/conversion refactor and the other is a broader dense/sparse model split. They are related, not the same concrete change and would not plausibly be merged as one PR." } ] }, @@ -21224,12 +21188,12 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "6a8dc0f1ee26318b3ffdba5cb0a6c7ee409fbd3e9cd3c52a8edcbce4c208fb94", + "prepared_review_unit_hash": "553acd11d5260b7dfd6ee716dff322970eb40b4e65ee5c20697e415706fecebe", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5466, + "serialized_chars": 5465, "estimated_input_tokens": 1367, "estimated_eval_tokens": 2990 }, @@ -21237,7 +21201,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5466, + "serialized_chars": 5465, "estimated_input_tokens": 1367, "estimated_eval_tokens": 2990 }, @@ -21246,18 +21210,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related heterogeneous-model efforts, but they are not the same concrete change: one PR adds per-layer config plumbing, while the other adds broader runtime/modeling support that builds on it.", - "confidence": 0.84, + "summary": "These are related feature PRs for heterogeneous support, but they implement different layers: one adds per-layer config handling, the other adds model/cache integration. They share helper files, yet they are not the same underlying change.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45332 is the more complete umbrella change: it includes the config plumbing plus the downstream cache/modeling updates needed to actually use heterogeneous per-layer settings.", + "canonical_pr_reason": "Pull request 45332 is the broader, more representative feature PR: it covers heterogeneous model support end-to-end and incorporates the downstream runtime changes that depend on per-layer config handling.", "best_issue_reason": null, - "best_pr_reason": "PR 45332 is the best representative of the cluster because it covers the actual end-to-end heterogeneous model support rather than only the configuration layer.", + "best_pr_reason": "PR 45332 is the better cluster representative because it is the more complete implementation and the one that would most naturally subsume the related config work, whereas 45333 is a narrower prerequisite/helper change.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "They overlap in heterogeneity/config files, but 45333 is limited to config representation/validation while 45332 also changes cache and modeling behavior for per-layer runtime support. They are related, not the same underlying bug/change, and would not cleanly merge into one PR." + "reason": "Related but distinct changes: 45333 adds per-layer configuration plumbing, while 45332 adds model/cache support and heterogeneous runtime behavior. They are not the same concrete bug/change and should not be merged as duplicates." } ] }, @@ -21287,12 +21251,12 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "098f4b14422a6fc83fe6810f1e28cbe638edd88b3e7feffa5e6946e2c6ad6d13", + "prepared_review_unit_hash": "c56df55e76430c8be3eaef468d0d3d05d7bc837aa874c16aef6a1f83a22f9a2d", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5514, + "serialized_chars": 5513, "estimated_input_tokens": 1379, "estimated_eval_tokens": 3014 }, @@ -21300,7 +21264,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5514, + "serialized_chars": 5513, "estimated_input_tokens": 1379, "estimated_eval_tokens": 3014 }, @@ -21309,18 +21273,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are type-checking/typing-related PRs, but they target different scopes: one is a broad typing pass across many files, while the other adds shared typing protocols and helpers for specific integrations. They are similar in theme, not duplicates.", - "confidence": 0.91, + "summary": "Both items are typing-related PRs in transformers, but they address different scopes: one is a broad repo-wide type-checking pass, the other adds specific typing helpers and related adjustments. They are not the same underlying change.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:45425 is the more focused and central typing infrastructure change, adding shared protocols/aliases used by multiple modules.", + "canonical_pr_reason": "PR 45425 is the more focused and coherent typing change: it introduces shared typing helpers and updates a small set of type-sensitive call sites. It is a better representative than the broad sweep in 45415.", "best_issue_reason": null, - "best_pr_reason": "pull_request:45425 is the strongest representative of the cluster because it introduces reusable typing abstractions rather than a broad mechanical sweep.", + "best_pr_reason": "PR 45425 is the best single PR in the cluster because it is narrower, more self-contained, and easier to treat as a distinct change. PR 45415 is a large multi-file type-checking cleanup rather than the same fix.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Both are about typing, but they address different concrete changes and code paths; 45415 is a broad type-checking sweep, while 45425 adds shared typing protocols and aliases. They do not look mergeable into one PR." + "reason": "Do not merge: 45415 is a broad type-checking sweep across many files, while 45425 adds typing aliases/protocols and a few targeted fixes. They share a theme but not the same concrete bug or change, and they would not naturally combine into one PR." } ] }, @@ -21350,7 +21314,7 @@ "soft_pairs": [ "pull_request:45513|pull_request:46084" ], - "prepared_review_unit_hash": "d0e0d7e00df06de76bfd46f900a79fc941f7a06ae694c48ad985b9d3fe2ea868", + "prepared_review_unit_hash": "a0ce8f836cf5ff9d64f7134dfb48590ee1000ada20d97c29ea62de7680dd1a80", "component_budget": { "node_count": 2, "item_count": 2, @@ -21372,18 +21336,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated pull requests about cache-handling bugs in different model families; they should not be clustered as duplicates.", - "confidence": 0.97, + "summary": "These are two unrelated pull requests fixing different cache-forward bugs in different model families: one in Qwen3.5/Qwen3 Next/Olmo Hybrid linear attention, the other in Mamba2 CUDA kernels. They should not be clustered as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46084 is the cleaner canonical representative: it has an explicit issue target, a narrow scope, and a concrete Mamba2 cache-path fix for seq_len > 1.", + "canonical_pr_reason": "PR #46084 is the better canonical PR for this cluster only because it is narrowly scoped, has an explicit issue target (#46032), and the diff clearly addresses one concrete seq_len>1 cache path in Mamba2.", "best_issue_reason": null, - "best_pr_reason": "PR 46084 is the best single PR to represent the cluster because it is tightly scoped and directly tied to a specific bug report; PR 45513 is broader and spans multiple Qwen/Olmo model files.", + "best_pr_reason": "PR #46084 is the strongest standalone PR: it targets a specific bug, has a clear issue reference, and the change is tightly limited to one code path with accompanying tests. PR #45513 is broader and spans multiple model files, making it less suitable as the canonical representative.", "soft_edge_verdicts": [ { "left": "pull_request:45513", "right": "pull_request:46084", "accept": false, - "reason": "Different bugs and code paths: 45513 fixes multi-token cached forward for GDN linear attention in OLMO/Qwen-family models, while 46084 fixes Mamba2 use_cache behavior when seq_len > 1. They are not the same underlying change and would not be merged into one PR." + "reason": "Different underlying bugs in different architectures. #45513 fixes multi-token cached forward behavior for GDN linear attention in Qwen/Olmo models, while #46084 fixes use_cache with seq_len > 1 in Mamba2 CUDA kernels. They do not look mergeable into one PR." } ] }, @@ -21413,12 +21377,12 @@ "soft_pairs": [ "pull_request:45548|pull_request:45662" ], - "prepared_review_unit_hash": "ed3c6ca78223631054b3e15c6cf3d476844816c97a6ff1df4da3a773f1cd5579", + "prepared_review_unit_hash": "b3d8e14d3def9024d270bb325546187b11e78a650b513d268c2db44ccda8a5a3", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4506, + "serialized_chars": 4508, "estimated_input_tokens": 1127, "estimated_eval_tokens": 2510 }, @@ -21426,7 +21390,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4506, + "serialized_chars": 4508, "estimated_input_tokens": 1127, "estimated_eval_tokens": 2510 }, @@ -21435,18 +21399,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate EP-related pull requests, but they target different failure modes and code paths: one fixes EP + DeepSpeed ZeRO-3 loading via accelerate launch, the other fixes EP + FSDP2 expert parameters being overwritten by rank-0 broadcast. They are related in subsystem, but not duplicates.", - "confidence": 0.93, + "summary": "These PRs are related through expert parallelism, but they address different bugs: #45548 fixes EP + DeepSpeed ZeRO-3 loading via accelerate launch by serializing distributed_config and adding an EP flag, while #45662 fixes EP + FSDP2 expert parameters being overwritten during distributed loading. They should not be merged as duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45662 is the stronger representative: it addresses a concrete expert-parameter overwrite bug in the loading path and includes a focused multi-file fix around distributed wrapping.", + "canonical_pr_reason": "PR #45662 is the better representative of the cluster because it fixes the more concrete corruption path in distributed loading and spans the broader EP/FSDP2 code path, whereas #45548 is a narrower DeepSpeed ZeRO-3/accelerate configuration serialization fix.", "best_issue_reason": null, - "best_pr_reason": "PR 45662 is the best standalone PR in this cluster because it targets a specific, reproducible EP/FSDP2 regression with an explicit code-path fix; PR 45548 is a different loading issue and should not be merged as the same change.", + "best_pr_reason": "PR #45662 is the strongest standalone PR here: it targets a specific runtime bug where expert parameters can be silently overwritten, and its changes are tightly coupled to that failure mode.", "soft_edge_verdicts": [ { "left": "pull_request:45548", "right": "pull_request:45662", "accept": false, - "reason": "Both concern expert parallelism loading, but they fix different bugs: 45548 adds distributed_config handling and has_ep for EP + DeepSpeed ZeRO-3 via accelerate launch, while 45662 prevents expert tensors from being overwritten in the FSDP2/rank-0 broadcast path. Different root causes and code paths." + "reason": "Different underlying bugs. #45548 is about preserving distributed_config and exposing EP state for DeepSpeed ZeRO-3 loading via accelerate; #45662 fixes expert weights being overwritten during FSDP2/distributed loading. Shared EP terminology and one common file are not enough to treat them as the same change." } ] }, @@ -21476,7 +21440,7 @@ "soft_pairs": [ "pull_request:45639|pull_request:45645" ], - "prepared_review_unit_hash": "ada0656ed9f8b0c5062ab02363b2efb4502f74400bc995839dd49835cf8e8b24", + "prepared_review_unit_hash": "3f948ad5bc91207759e6db2058f97662838f178d67970f698c0651fbe96dcf34", "component_budget": { "node_count": 2, "item_count": 2, @@ -21498,18 +21462,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs target the same xdist-related captured_info logging bug in transformers testing utilities. PR 45645 is the broader, more complete fix because it includes the worker-safe log file change plus CI workflow updates to collect multiple captured_info files.", - "confidence": 0.86, + "summary": "Both PRs address the same xdist-related captured_info collision in `testing_utils.py`; PR 45645 is the broader, more complete fix that also updates CI artifact collection, while 45639 is the narrower core-code change.", + "confidence": 0.89, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45645 is the more complete fix: it includes the xdist-safe file handling in testing_utils and the CI-side changes needed to preserve and aggregate captured_info logs.", + "canonical_pr_reason": "PR 45645 is the better canonical PR because it covers the core xdist-safe output-file change in `src/transformers/testing_utils.py` and also updates the CI workflow to collect all split `captured_info*.txt` artifacts, making it the more complete end-to-end fix.", "best_issue_reason": null, - "best_pr_reason": "PR 45645 best represents the cluster because it addresses the same underlying collision bug while also updating the workflow to handle the new multi-file log output.", + "best_pr_reason": "PR 45645 is the best representative of the cluster since it addresses the same underlying bug as 45639 but includes the surrounding CI/log-handling pieces needed for the fix to work in practice.", "soft_edge_verdicts": [ { "left": "pull_request:45639", "right": "pull_request:45645", "accept": true, - "reason": "Both PRs fix the same concrete xdist/captured_info collision problem in the same code path; 45645 is a superset of 45639 rather than a separate change." + "reason": "Accept: both PRs target the same captured_info/xdist collision problem in `testing_utils.py` and share the same issue target; 45645 extends the same fix rather than introducing a different change." } ] }, @@ -21541,12 +21505,12 @@ "pull_request:45670|pull_request:46320", "pull_request:46320|pull_request:46401" ], - "prepared_review_unit_hash": "22557b28790366643265e193b60966633240585e8ad82bc9d8ed7f0fcb991763", + "prepared_review_unit_hash": "e722c74e86b11169ba4f3b942d3e31d2ac7d4050b2a8ddbd59d813600f0d10bc", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 7029, + "serialized_chars": 7030, "estimated_input_tokens": 1758, "estimated_eval_tokens": 3772 }, @@ -21554,7 +21518,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 7029, + "serialized_chars": 7030, "estimated_input_tokens": 1758, "estimated_eval_tokens": 3772 }, @@ -21563,24 +21527,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The PRs are all separate feature fixes, not duplicates. Two are same broad theme (attn_implementation dispatch) but for different model families, so they should stay separate. The auto-model mapping change is unrelated.", + "summary": "Only PRs 46320 and 46401 are superficially similar, but they target different model families and code paths, so this cluster does not contain a true duplicate change set. PR 45670 is unrelated to the attention-dispatch work.", "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46320 is the clearest standalone feature change: it adds RecurrentGemma attn_implementation dispatch support with matching model, conversion, and test updates. It is more self-contained and representative than the other unrelated PRs.", + "canonical_pr_reason": "PR 46320 is the best representative of the cluster\u2019s only near-duplicate theme: a concrete attn_implementation dispatch implementation with code changes, conversion mapping updates, docs, and tests.", "best_issue_reason": null, - "best_pr_reason": "PR 46320 is the strongest overall candidate because it has a focused, concrete model-path change with supporting code and tests, unlike the other PRs which target different subsystems or model families.", + "best_pr_reason": "PR 46320 is the strongest standalone representative because it is the most complete dispatch-support change and is closest in shape to the other similar PR, while still being clearly distinct from the unrelated glmasr mapping tweak.", "soft_edge_verdicts": [ { "left": "pull_request:45670", "right": "pull_request:46320", "accept": false, - "reason": "Different features and code paths: AutoModel registration for glmasr vs RecurrentGemma attn dispatch support. No shared underlying bug/change." + "reason": "These are different changes: 45670 adds model mapping entries for glmasr/dia, while 46320 adds RecurrentGemma attention-dispatch support across model code, conversion mapping, docs, and tests. Not the same bug or change." }, { "left": "pull_request:46320", "right": "pull_request:46401", "accept": false, - "reason": "Same general pattern, but different concrete models (RecurrentGemma vs BLIP-2/InstructBLIP). These are separate model-specific dispatch implementations, not one mergeable fix." + "reason": "Both are attn_implementation dispatch work, but for different architectures and code paths (RecurrentGemma vs BLIP-2/InstructBLIP/InstructBLIPVideo). They are not the same underlying fix and would not plausibly merge into one PR." } ] }, @@ -21610,7 +21574,7 @@ "soft_pairs": [ "pull_request:45720|pull_request:46259" ], - "prepared_review_unit_hash": "39603f839c3f1c5b2c37475c4f94ed00318d9f4dac478d00b3d2b3cb522fb711", + "prepared_review_unit_hash": "113785978077212c586b4c67f049040fd602df1024789239f35a1be136efb0fb", "component_budget": { "node_count": 2, "item_count": 2, @@ -21632,18 +21596,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not duplicates: one implements TRF003 around return_dict/can_return_tuple handling in model forward methods, while the other adds TRF018 for dead do_* flags in image/video processor overrides. They target different code paths and different rule families.", + "summary": "These two pull requests are not duplicates: one implements a linter rule for model forward signatures, while the other adds a different modeling rule about consumed image/video processor flags. They overlap only at the level of shared transformer model files.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46259 is the more complete standalone change: it defines TRF018 in the rules documentation and applies the corresponding enforcement across image/video processor classes.", + "canonical_pr_reason": "PR 45720 is the cleaner representative: it has a focused, concrete change set around forward signatures and return handling, with title and diff aligned.", "best_issue_reason": null, - "best_pr_reason": "PR 46259 is the clearest representative of this cluster because it adds a new, well-scoped rule with explicit spec plus code updates; PR 45720 is a separate TRF003 fix in a different subsystem.", + "best_pr_reason": "PR 45720 is the best representative of the cluster because its scope is narrower and more internally consistent than the other PR\u2019s broader rule/docs update.", "soft_edge_verdicts": [ { "left": "pull_request:45720", "right": "pull_request:46259", "accept": false, - "reason": "Different underlying changes: TRF003 for model forward return handling vs TRF018 for image/video processor boolean flag usage. They are not the same bug and would not plausibly merge into one PR." + "reason": "Different underlying changes: TRF003 forward-signature/return_dict cleanup versus a separate TRF016 rule for processor boolean flags. Shared files are incidental, not the same bug or mergeable fix." } ] }, @@ -21739,40 +21703,40 @@ "soft_pairs": [ "pull_request:45932|pull_request:46100" ], - "prepared_review_unit_hash": "ac4f87decbea0ed572991f5314ca95b5eb3ce2ca95a81ea16109453ce1419c08", + "prepared_review_unit_hash": "71f34ae2890f2e2718cb4e1085855949a3a7ef770a2429d972641675a0f1227f", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4300, - "estimated_input_tokens": 1075, - "estimated_eval_tokens": 2406 + "serialized_chars": 4301, + "estimated_input_tokens": 1076, + "estimated_eval_tokens": 2408 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4300, - "estimated_input_tokens": 1075, - "estimated_eval_tokens": 2406 + "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 pull requests are not duplicates: one is a broad experimental/test-infra and model-code patch, while the other is a tiny test-only migration with deliberate failing assertions. The shared \"DO NOT MERGE\" vibe and testing context are superficial.", - "confidence": 0.83, + "summary": "The two PRs are unrelated: one is a broad testing/infra and model-maintenance change, while the other is a narrow ViT test edit that adds intentional failing assertions. They do not appear to fix the same bug or implement the same change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45932 is the more substantive representative of the cluster: it changes CI secrets and multiple model files, whereas PR 46100 is a small test-only edit with obvious placeholder failures.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 45932 is the best global representative because it has the larger, concrete code/config footprint; PR 46100 looks like a throwaway test migration and is less representative of meaningful work.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:45932", "right": "pull_request:46100", "accept": false, - "reason": "Reject: they do not appear to fix the same concrete code-path problem. PR 45932 touches workflow/config and several model implementations; PR 46100 only edits a ViT test file with intentional failing assertions. Shared testing context is not enough for a duplicate PR merge." + "reason": "Different underlying intent and code paths: PR 45932 touches workflow/secrets and multiple model files, while PR 46100 only changes one ViT test file and appears to be test-migration noise. They are not plausibly mergeable as one PR." } ] }, @@ -21805,7 +21769,7 @@ "pull_request:46073|pull_request:46196", "pull_request:46154|pull_request:46196" ], - "prepared_review_unit_hash": "92f6b6e7a47664aba208cd3f10e59b868fd6a3095e884c90cec82443ad61504c", + "prepared_review_unit_hash": "21ddcc6283bb2aa50e105116b3989f412b5194e7858a75114a4966bb8b1ec8b8", "component_budget": { "node_count": 3, "item_count": 3, @@ -21827,30 +21791,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate model-specific PRs that all add SDPA dispatch support, but they target different architectures and code paths. They look related by theme, not duplicates of the same underlying change.", + "summary": "These are related in theme only: all add SDPA dispatch support, but each PR targets a different model family and distinct code paths, so they are not duplicates.", "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: CTRL, T5, and wav2vec2_conformer/bert + SeamlessM4T/v2 each implement SDPA dispatch in different model implementations and tests.", + "canonical_pr_reason": "No canonical duplicate PR: CTRL, T5, and the wav2vec2/SeamlessM4T work are separate model-specific SDPA implementations.", "best_issue_reason": null, - "best_pr_reason": "No global best PR for duplication triage; each PR is a distinct model-family adaptation rather than one mergeable change.", + "best_pr_reason": "PR 46196 is the strongest representative of the theme because it covers multiple related speech models and includes broader code/test updates, making it the most complete SDPA-dispatch change.", "soft_edge_verdicts": [ { "left": "pull_request:46073", "right": "pull_request:46154", "accept": false, - "reason": "Both are SDPA-dispatch support PRs, but one is for CTRL and the other for T5; they touch different modeling files, tests, and model-specific attention paths, so they are not the same change." + "reason": "Different architectures and code paths: CTRL vs T5. Same feature theme, but not the same concrete bug/change and not mergeable into one PR." }, { "left": "pull_request:46073", "right": "pull_request:46196", "accept": false, - "reason": "Shared theme only. CTRL versus wav2vec2_conformer/bert and SeamlessM4T/v2 are distinct architectures with different code paths, so these are not duplicates." + "reason": "CTRL support is unrelated to wav2vec2/SeamlessM4T speech-model dispatch internals. Broad SDPA theme only, not a duplicate fix." }, { "left": "pull_request:46154", "right": "pull_request:46196", "accept": false, - "reason": "Although both add SDPA support, they apply to different model families and require different implementations and tests; they should not be merged as one PR." + "reason": "T5 seq2seq attention dispatch and speech-model dispatch are separate implementations in different model families; they are not the same underlying change." } ] }, @@ -21880,12 +21844,12 @@ "soft_pairs": [ "pull_request:46126|pull_request:46269" ], - "prepared_review_unit_hash": "137b5da2adcc04f65479358be44d7166720049c3e285a222a5b365eb7cfc9f62", + "prepared_review_unit_hash": "e29b30697c2e410af6697160a7ca276bc1d89064aee8659693586e28dfe98db5", "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 }, @@ -21893,7 +21857,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5571, + "serialized_chars": 5572, "estimated_input_tokens": 1393, "estimated_eval_tokens": 3042 }, @@ -21902,18 +21866,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs in the distributed stack overlap on a few filenames, but they are not the same change: one is a focused EP/Trainer integration patch, the other is a broad distributed branch base spanning FSDP, tensor parallel, integrations, and CI.", - "confidence": 0.9, + "summary": "Two open PRs overlap in distributed infrastructure files, but they target different levels of work: one is a focused EP + Trainer integration, the other is a broad distributed branch base sweep. They are not the same change and should not be merged as duplicates.", + "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #46269 is the broader umbrella change and the more representative cluster anchor; it subsumes many distributed-system updates beyond the focused EP/trainer work in #46126.", + "canonical_pr_reason": "PR #46269 is the broader umbrella change: it spans many distributed subsystems and looks like the base branch integration point rather than a narrow feature patch, making it the more canonical representative of this cluster.", "best_issue_reason": null, - "best_pr_reason": "PR #46269 is the best single representative for the cluster because it is the larger, more central distributed integration branch and covers the widest set of related changes.", + "best_pr_reason": "PR #46269 is the best single representative because it covers the larger distributed refactor/integration surface and subsumes the shared files without being narrowly scoped to one feature path.", "soft_edge_verdicts": [ { "left": "pull_request:46126", "right": "pull_request:46269", "accept": false, - "reason": "They overlap on distributed files, but the concrete changes are different in scope and purpose. #46126 targets EP routing and trainer integration, while #46269 is a broad distributed base branch with FSDP, tensor parallel, integration, and CI changes; they do not look mergeable as one PR." + "reason": "Overlap is only at the subsystem/file level. #46126 is a specific EP + Trainer integration, while #46269 is a broad distributed base PR touching many unrelated areas; they do not appear to fix the same concrete code-path problem and would not plausibly merge into one PR." } ] }, @@ -21943,7 +21907,7 @@ "soft_pairs": [ "pull_request:46138|pull_request:46432" ], - "prepared_review_unit_hash": "580340171a3a72d354e051de3bc4d833bfbb360efe02364f6cf8d364649c35b4", + "prepared_review_unit_hash": "15ef53debc7b97ac1fc6e9fe62711fe7a118a0499503fbc1270b7f193ce7e49e", "component_budget": { "node_count": 2, "item_count": 2, @@ -21965,18 +21929,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two unrelated workflow-only PRs: one changes self-comment CI to read the comment from stdin, the other hardens ssh-runner by pinning a downloaded script, verifying its checksum, and unsetting HF_TOKEN. They are not duplicates.", + "summary": "These two pull requests are not duplicates: one changes a workflow invocation style in self-comment CI, while the other hardens ssh-runner by pinning a downloaded script and clearing HF_TOKEN before SSH.", "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46432 is the more substantial and security-relevant workflow change, with a pinned download, checksum verification, and token cleanup, so it is the better canonical representative of this cluster.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 46432 is the strongest standalone PR here; PR 46138 is a separate minor CI plumbing change in a different workflow.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:46138", "right": "pull_request:46432", "accept": false, - "reason": "Different workflows and different changes: stdin handling for self-comment CI versus supply-chain/token hardening in ssh-runner. They do not appear to fix the same underlying bug or change the same code path." + "reason": "Different workflows and different underlying changes: stdin piping for a CI helper vs security/hardening updates for ssh-runner. They do not fix the same code-path problem and would not plausibly merge into one PR." } ] }, @@ -22006,7 +21970,7 @@ "soft_pairs": [ "pull_request:46152|pull_request:46207" ], - "prepared_review_unit_hash": "77e21db6fd0f5c9c867ae0321f24761fc6278fd5cd9b6046cd540b0c86648122", + "prepared_review_unit_hash": "05f4f61fa02ba2c1df9d614e5a7193150b18ccc39e4c25c02b4ca4df708cc6cf", "component_budget": { "node_count": 2, "item_count": 2, @@ -22028,18 +21992,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two open PRs with a similar theme of removing brittle file-open source inspection, but they fix different code paths in different modules and should not be deduplicated.", - "confidence": 0.89, + "summary": "Both PRs are about avoiding brittle file-open/source inspection, but they change different code paths in different modules and address different implementation details. They are related thematically, not duplicates.", + "confidence": 0.9, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 46207 is the cleaner representative: it makes a concrete, test-backed fix in `modeling_utils.py` for `_can_set_attn_implementation`, replacing `open()`-based source reading with `inspect.getsource`/module inspection.", + "canonical_pr_reason": "PR 46207 is the more concrete exemplar: it tightens one specific source-reading heuristic in modeling_utils and includes tests, making it the cleaner representative of the cluster theme.", "best_issue_reason": null, - "best_pr_reason": "PR 46207 is the stronger cluster exemplar because it addresses a specific runtime heuristic bug in one code path and includes tests, making it more clearly actionable than the more infrastructure-oriented import-utils change.", + "best_pr_reason": "PR 46207 is the best single PR in this set because it has the clearest narrowly-scoped fix and test coverage. PR 46152 is a separate utility-level refactor in import_utils, not the same concrete bug.", "soft_edge_verdicts": [ { "left": "pull_request:46152", "right": "pull_request:46207", "accept": false, - "reason": "Both PRs are about avoiding brittle source-file access, but they target different functions and bugs: `import_utils.py` path/traversable resolution versus `modeling_utils.py` attention-implementation detection. They are related in style, not the same underlying change." + "reason": "Different files, different helpers, and different behaviors: one adds Traversable resolution for importing package source, the other replaces open() with inspect.getsource for attention-implementation detection. Shared motivation is not enough to treat them as the same underlying change." } ] }, @@ -22061,71 +22025,63 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-46337-4", + "cluster_id": "cluster-46337-3", "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" + "pull_request:46362|pull_request:46368" ], - "prepared_review_unit_hash": "50d97a9cca2e92621a23bc103bd55f93cb604a428a54f8a75ff2e86744383cec", + "prepared_review_unit_hash": "991b266e7fb2dacce86caec08ac65d505141e33a2cd08976917fe0c7f27c5a43", "component_budget": { - "node_count": 4, - "item_count": 4, - "soft_pair_count": 4, - "serialized_chars": 8272, - "estimated_input_tokens": 2068, - "estimated_eval_tokens": 4392 + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5964, + "estimated_input_tokens": 1491, + "estimated_eval_tokens": 3238 }, "budget": { - "node_count": 4, - "item_count": 4, - "soft_pair_count": 4, - "serialized_chars": 8272, - "estimated_input_tokens": 2068, - "estimated_eval_tokens": 4392 + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5964, + "estimated_input_tokens": 1491, + "estimated_eval_tokens": 3238 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are only loosely similar at the test-infra level, but they target different model families and different failures: Cohere2Moe device-map/offload setup, DeepSeek V4 quant-cache skipping and XPU support, and NemotronH quant-cache skipping. They are not duplicates of the same underlying bug/change.", - "confidence": 0.96, + "summary": "These are three separate test-only pull requests for different model families and different failure modes. The only commonality is that they all touch model test code and mention quant/cache or device-map related behavior; they do not look like duplicates or one shared fix.", + "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR 46366 is the strongest standalone candidate because it contains an actual functional test/backend adjustment for DeepSeek V4, whereas the others are narrowly scoped model-specific skips or test fixture tweaks. It is still not a duplicate of the other PRs.", + "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "pull_request:46337", "right": "pull_request:46368", "accept": false, - "reason": "Different models and different failures: Cohere2Moe device-map/offload test setup versus NemotronH quant-cache incompatibility. Same broad area, but not the same code-path bug or mergeable fix." + "reason": "Different model targets and fixes: Cohere2Moe adjusts model-test settings for offload/device-map splitting, while NemotronH skips an invalid quantized-cache test. No shared code path or mergeable single fix." }, { "left": "pull_request:46337", "right": "pull_request:46362", "accept": false, - "reason": "Cohere2Moe test environment tuning for accelerate/offload is unrelated to DeepSeek V4 skipping a quant-cache test. They do not fix the same underlying problem." + "reason": "Cohere2Moe test infrastructure tweak vs. DeepSeek V4 quant-cache skip. They address unrelated model-specific test failures and do not appear to be the same underlying bug." }, { "left": "pull_request:46362", "right": "pull_request:46368", "accept": false, - "reason": "Both mention quant cache, but they are model-specific skips for different architectures with different cache implementations. Not the same bug and not something to merge into one PR." - }, - { - "left": "pull_request:46362", - "right": "pull_request:46366", - "accept": false, - "reason": "DeepSeek V4 quant-cache skip and DeepSeek V4 XPU extension are different changes: one avoids a failing test, the other broadens backend coverage. They are not one concrete fix." + "reason": "Both skip quant-cache tests, but for different architectures and different incompatibilities (DeepSeek V4 compressor cache vs. NemotronH hybrid cache). Same broad subsystem, not the same concrete change." } ] }, @@ -22155,7 +22111,7 @@ "soft_pairs": [ "pull_request:46376|pull_request:46417" ], - "prepared_review_unit_hash": "44ec209756fee8b1110062b71c2e8b91597343e9ae5a132ac3092f924e2ad8c4", + "prepared_review_unit_hash": "0c64f31b0f74f743ffcc2e14cef82e92a1c2a39e0183d50d37b17ef2b7e8219b", "component_budget": { "node_count": 2, "item_count": 2, @@ -22177,18 +22133,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two pull requests are not duplicates: one adds multimodal support to continuous batching, while the other adds a new HunYuan VL model package, docs, and tests. They target different code paths and could not reasonably be merged into one PR.", - "confidence": 0.97, + "summary": "Two unrelated open PRs: one is a draft for multimodal continuous batching infrastructure, the other adds a new HunYuan VL model and its docs/tests. They do not look like duplicate fixes or mergeable variants of the same change.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": "pull_request:46417 is the more self-contained and representative change in the cluster: a concrete new model integration with associated docs and tests.", + "canonical_pr_reason": "No canonical PR: the two PRs target different problems and codepaths, so there is no duplicate representative.", "best_issue_reason": null, - "best_pr_reason": "pull_request:46417 is the best representative PR because it is a complete, coherent feature addition rather than a broad draft touching many subsystems.", + "best_pr_reason": "PR #46417 is the better standalone representative because it is a complete, non-draft feature addition with docs, model code, and tests; #46376 is still a draft infrastructure refactor.", "soft_edge_verdicts": [ { "left": "pull_request:46376", "right": "pull_request:46417", "accept": false, - "reason": "Different underlying changes: continuous-batching multimodal support versus adding a HunYuan VL model and documentation. No shared concrete bug or code path, so they should not be merged as duplicates." + "reason": "Different changes in different areas: continuous batching multimodal support versus adding a new HunYuan VL model. They do not fix the same concrete bug or share a mergeable codepath." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index 05c24fef2ed9f4fd6431aa336567001a80aa3483..30b105d0e6628e0760819d7991077bfba1fe2bf9 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-model-20260605t120123z", + "analysis_id": "hybrid-model-20260605t180026z", "archived_artifacts": { - "hybrid": "snapshots/20260605T120123Z/analysis-runs/hybrid-model-20260605t120123z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260605T120123Z/analysis-runs/hybrid-model-20260605t120123z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-06-05T12:11:21Z", + "published_at": "2026-06-05T18:11:22Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260605T120123Z", + "snapshot_id": "20260605T180026Z", "variant": "hybrid" } diff --git a/snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.json b/snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..1781002beeae1eedc1c8058590f49a24c0386610 --- /dev/null +++ b/snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.json @@ -0,0 +1,2177 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260605T180026Z", + "generated_at": "2026-06-05T18:11:20Z", + "evidence_quality": "full", + "llm_enrichment": true, + "meta_bugs": [ + { + "cluster_id": "cluster-43979-11", + "summary": "Cluster of 1 issues and 10 PRs centered on issue #43979.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43979, + "canonical_pr_number": 44007, + "issue_numbers": [ + 43979 + ], + "pr_numbers": [ + 43996, + 44007, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43996, + "right_pr_number": 44007, + "code_similarity": 0.179, + "size_similarity": 0.576, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.429, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44013, + "code_similarity": 0.122, + "size_similarity": 0.318, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.392, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44044, + "code_similarity": 0.245, + "size_similarity": 0.864, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.479, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44066, + "code_similarity": 0.225, + "size_similarity": 0.818, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.408, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44072, + "code_similarity": 0.14, + "size_similarity": 0.303, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.528, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44085, + "code_similarity": 0.216, + "size_similarity": 0.783, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44129, + "code_similarity": 0.163, + "size_similarity": 0.643, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.229, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44154, + "code_similarity": 0.153, + "size_similarity": 0.535, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.31, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 43996, + "right_pr_number": 44722, + "code_similarity": 0.225, + "size_similarity": 0.848, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.368, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44013, + "code_similarity": 0.19, + "size_similarity": 0.553, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.531, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44044, + "code_similarity": 0.186, + "size_similarity": 0.667, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.354, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44066, + "code_similarity": 0.188, + "size_similarity": 0.704, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.315, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44072, + "code_similarity": 0.212, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.708, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44085, + "code_similarity": 0.195, + "size_similarity": 0.735, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.318, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44129, + "code_similarity": 0.103, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.191, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44154, + "code_similarity": 0.238, + "size_similarity": 0.93, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.344, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44007, + "right_pr_number": 44722, + "code_similarity": 0.178, + "size_similarity": 0.679, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.28, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44044, + "code_similarity": 0.126, + "size_similarity": 0.368, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.351, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44066, + "code_similarity": 0.127, + "size_similarity": 0.389, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.325, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44072, + "code_similarity": 0.29, + "size_similarity": 0.952, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.667, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44085, + "code_similarity": 0.131, + "size_similarity": 0.406, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.329, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44129, + "code_similarity": 0.07, + "size_similarity": 0.205, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.192, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44154, + "code_similarity": 0.177, + "size_similarity": 0.594, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.389, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44013, + "right_pr_number": 44722, + "code_similarity": 0.118, + "size_similarity": 0.375, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.287, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44066, + "code_similarity": 0.25, + "size_similarity": 0.947, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.404, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44072, + "code_similarity": 0.136, + "size_similarity": 0.351, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.442, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44085, + "code_similarity": 0.24, + "size_similarity": 0.906, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.394, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44129, + "code_similarity": 0.147, + "size_similarity": 0.555, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.243, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44154, + "code_similarity": 0.17, + "size_similarity": 0.62, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.306, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44044, + "right_pr_number": 44722, + "code_similarity": 0.257, + "size_similarity": 0.982, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.402, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44072, + "code_similarity": 0.133, + "size_similarity": 0.37, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.393, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44085, + "code_similarity": 0.763, + "size_similarity": 0.957, + "file_overlap": 0.5, + "area_overlap": 0.825, + "patch_similarity": 0.887, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ] + } + ] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44129, + "code_similarity": 0.145, + "size_similarity": 0.526, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.263, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44154, + "code_similarity": 0.174, + "size_similarity": 0.654, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.286, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44066, + "right_pr_number": 44722, + "code_similarity": 0.907, + "size_similarity": 0.964, + "file_overlap": 1.0, + "area_overlap": 0.808, + "patch_similarity": 0.874, + "shared_filenames": [ + "src/transformers/models/codegen/modeling_codegen.py", + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/codegen/modeling_codegen.py", + "left_ranges": [ + [ + 245, + 266 + ] + ], + "right_ranges": [ + [ + 228, + 234 + ] + ] + }, + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 41 + ], + [ + 174, + 181 + ], + [ + 250, + 257 + ], + [ + 398, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 466 + ], + [ + 471, + 488 + ], + [ + 518, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 606 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44085, + "code_similarity": 0.137, + "size_similarity": 0.387, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.398, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44129, + "code_similarity": 0.074, + "size_similarity": 0.195, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.231, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44154, + "code_similarity": 0.175, + "size_similarity": 0.566, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.414, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44072, + "right_pr_number": 44722, + "code_similarity": 0.124, + "size_similarity": 0.357, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.347, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44129, + "code_similarity": 0.141, + "size_similarity": 0.503, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.272, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44154, + "code_similarity": 0.18, + "size_similarity": 0.684, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.289, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44085, + "right_pr_number": 44722, + "code_similarity": 0.728, + "size_similarity": 0.923, + "file_overlap": 0.5, + "area_overlap": 0.791, + "patch_similarity": 0.78, + "shared_filenames": [ + "src/transformers/models/gptj/modeling_gptj.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gptj/modeling_gptj.py", + "left_ranges": [ + [ + 33, + 40 + ], + [ + 173, + 180 + ], + [ + 249, + 256 + ], + [ + 397, + 405 + ], + [ + 408, + 419 + ], + [ + 425, + 434 + ], + [ + 459, + 465 + ], + [ + 470, + 489 + ], + [ + 517, + 539 + ], + [ + 553, + 559 + ], + [ + 565, + 574 + ], + [ + 579, + 597 + ], + [ + 600, + 611 + ], + [ + 633, + 639 + ], + [ + 645, + 652 + ], + [ + 657, + 673 + ], + [ + 716, + 728 + ], + [ + 737, + 743 + ], + [ + 748, + 771 + ], + [ + 789, + 794 + ] + ], + "right_ranges": [ + [ + 33, + 42 + ], + [ + 175, + 182 + ], + [ + 251, + 258 + ], + [ + 399, + 420 + ], + [ + 426, + 435 + ], + [ + 460, + 467 + ], + [ + 472, + 489 + ], + [ + 519, + 540 + ], + [ + 554, + 560 + ], + [ + 566, + 575 + ], + [ + 580, + 598 + ], + [ + 601, + 612 + ], + [ + 634, + 640 + ], + [ + 646, + 653 + ], + [ + 658, + 663 + ], + [ + 666, + 674 + ], + [ + 717, + 722 + ], + [ + 738, + 744 + ], + [ + 749, + 772 + ], + [ + 790, + 795 + ] + ] + } + ] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44154, + "code_similarity": 0.099, + "size_similarity": 0.344, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.199, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44129, + "right_pr_number": 44722, + "code_similarity": 0.146, + "size_similarity": 0.545, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.247, + "shared_filenames": [], + "shared_file_areas": [] + }, + { + "left_pr_number": 44154, + "right_pr_number": 44722, + "code_similarity": 0.164, + "size_similarity": 0.631, + "file_overlap": 0.0, + "area_overlap": 0.0, + "patch_similarity": 0.255, + "shared_filenames": [], + "shared_file_areas": [] + } + ] + }, + { + "cluster_id": "cluster-41211-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41211.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.155, + "size_similarity": 0.094, + "file_overlap": 0.273, + "area_overlap": 0.078, + "patch_similarity": 0.183, + "shared_filenames": [ + "docs/source/en/model_doc/deimv2.md", + "src/transformers/models/deimv2/__init__.py", + "src/transformers/models/deimv2/configuration_deimv2.py", + "src/transformers/models/deimv2/modeling_deimv2.py", + "tests/models/deimv2/__init__.py", + "tests/models/deimv2/test_modeling_deimv2.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/deimv2.md", + "left_ranges": [ + [ + 1, + 132 + ] + ], + "right_ranges": [ + [ + 1, + 65 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/__init__.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/deimv2/configuration_deimv2.py", + "left_ranges": [ + [ + 1, + 74 + ] + ], + "right_ranges": [ + [ + 1, + 266 + ] + ] + }, + { + "filename": "tests/models/deimv2/test_modeling_deimv2.py", + "left_ranges": [ + [ + 1, + 15 + ] + ], + "right_ranges": [ + [ + 1, + 1734 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43240-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43240.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43240, + "canonical_pr_number": 43251, + "issue_numbers": [ + 43240 + ], + "pr_numbers": [ + 43251, + 43254 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43251, + "right_pr_number": 43254, + "code_similarity": 0.794, + "size_similarity": 0.64, + "file_overlap": 1.0, + "area_overlap": 0.667, + "patch_similarity": 0.882, + "shared_filenames": [ + "src/transformers/loss/loss_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/loss/loss_utils.py", + "left_ranges": [ + [ + 30, + 50 + ] + ], + "right_ranges": [ + [ + 30, + 43 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-41115-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #41115.", + "status": "open", + "confidence": 0.55, + "canonical_issue_number": 41115, + "canonical_pr_number": 45613, + "issue_numbers": [ + 41115 + ], + "pr_numbers": [ + 41116, + 45613 + ], + "evidence_types": [ + "closing_reference" + ], + "pr_comparisons": [ + { + "left_pr_number": 41116, + "right_pr_number": 45613, + "code_similarity": 0.797, + "size_similarity": 0.885, + "file_overlap": 0.75, + "area_overlap": 0.831, + "patch_similarity": 0.693, + "shared_filenames": [ + "docs/source/en/_toctree.yml", + "docs/source/en/model_doc/minicpm3.md", + "src/transformers/models/auto/modeling_auto.py", + "src/transformers/models/minicpm3/__init__.py", + "src/transformers/models/minicpm3/configuration_minicpm3.py", + "src/transformers/models/minicpm3/modeling_minicpm3.py", + "src/transformers/models/minicpm3/modular_minicpm3.py", + "tests/models/minicpm3/__init__.py", + "tests/models/minicpm3/test_modeling_minicpm3.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/_toctree.yml", + "left_ranges": [ + [ + 713, + 720 + ] + ], + "right_ranges": [ + [ + 730, + 737 + ] + ] + }, + { + "filename": "docs/source/en/model_doc/minicpm3.md", + "left_ranges": [ + [ + 1, + 67 + ] + ], + "right_ranges": [ + [ + 1, + 45 + ] + ] + }, + { + "filename": "src/transformers/models/auto/modeling_auto.py", + "left_ranges": [ + [ + 285, + 291 + ], + [ + 700, + 706 + ], + [ + 1298, + 1304 + ] + ], + "right_ranges": [ + [ + 720, + 726 + ], + [ + 1331, + 1337 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/__init__.py", + "left_ranges": [ + [ + 1, + 29 + ] + ], + "right_ranges": [ + [ + 1, + 29 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/configuration_minicpm3.py", + "left_ranges": [ + [ + 1, + 141 + ] + ], + "right_ranges": [ + [ + 1, + 126 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 544 + ] + ], + "right_ranges": [ + [ + 1, + 522 + ] + ] + }, + { + "filename": "src/transformers/models/minicpm3/modular_minicpm3.py", + "left_ranges": [ + [ + 1, + 444 + ] + ], + "right_ranges": [ + [ + 1, + 342 + ] + ] + }, + { + "filename": "tests/models/minicpm3/test_modeling_minicpm3.py", + "left_ranges": [ + [ + 1, + 109 + ] + ], + "right_ranges": [ + [ + 1, + 136 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45647-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45647.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 45647, + "canonical_pr_number": 45738, + "issue_numbers": [ + 45647 + ], + "pr_numbers": [ + 45737, + 45738 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 45737, + "right_pr_number": 45738, + "code_similarity": 0.526, + "size_similarity": 0.11, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.024, + "shared_filenames": [ + "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/musicgen_melody/modeling_musicgen_melody.py", + "left_ranges": [ + [ + 491, + 497 + ] + ], + "right_ranges": [ + [ + 491, + 497 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45561-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45561.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45561, + "canonical_pr_number": 45645, + "issue_numbers": [ + 45561 + ], + "pr_numbers": [ + 45639, + 45645 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45639, + "right_pr_number": 45645, + "code_similarity": 0.494, + "size_similarity": 0.619, + "file_overlap": 0.5, + "area_overlap": 0.434, + "patch_similarity": 0.451, + "shared_filenames": [ + "src/transformers/testing_utils.py", + "tests/utils/test_testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3595, + 3602 + ], + [ + 3819, + 3853 + ] + ], + "right_ranges": [ + [ + 3525, + 3558 + ], + [ + 3782, + 3788 + ] + ] + }, + { + "filename": "tests/utils/test_testing_utils.py", + "left_ranges": [ + [ + 1, + 86 + ] + ], + "right_ranges": [ + [ + 1, + 114 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45666-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 46339, + "issue_numbers": [], + "pr_numbers": [ + 45666, + 46339 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "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-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, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45647-3", + "canonical_pr_number": 45738, + "duplicate_pr_numbers": [ + 45737 + ], + "target_issue_number": 45647, + "reason": "PRs in cluster-45647-3 are treated as duplicates because they converge on issue #45647 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-45666-2", + "canonical_pr_number": 46339, + "duplicate_pr_numbers": [ + 45666 + ], + "target_issue_number": null, + "reason": "PRs in cluster-45666-2 are treated as duplicates because they share soft_similarity 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/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..9f1bc3651e13c60841b4b990ecfe822e22026c8f --- /dev/null +++ b/snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,22159 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260605T180026Z", + "generated_at": "2026-06-05T18:11:20Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43296", + "issue:43425", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43827", + "issue:43828", + "issue:43866", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44805", + "issue:44918", + "issue:45081", + "issue:45161", + "issue:45237", + "issue:45464", + "issue:45507", + "issue:45685", + "issue:45800", + "issue:45823", + "issue:45859", + "issue:45874", + "issue:45901", + "issue:45902", + "issue:46015", + "issue:46165", + "issue:46294", + "issue:46304", + "issue:46402" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:45237|issue:45685", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:45901|issue:46015", + "issue:44509|issue:44512", + "issue:35545|issue:36010", + "issue:44918|issue:46165", + "issue:44805|issue:45823", + "issue:43638|issue:44805", + "issue:43638|issue:46165", + "issue:43596|issue:44805", + "issue:44805|issue:46294", + "issue:44805|issue:45161", + "issue:44560|issue:46165", + "issue:43596|issue:46165", + "issue:43572|issue:46165", + "issue:34634|issue:35545", + "issue:30277|issue:43638", + "issue:43596|issue:43638", + "issue:43039|issue:45161", + "issue:43975|issue:44457", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:45859|issue:45902", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:45874|issue:46294", + "issue:39290|issue:45081", + "issue:45823|issue:45874", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:43828|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:45685", + "issue:44918|issue:45161", + "issue:44292|issue:44918", + "issue:45507|issue:45685", + "issue:43716|issue:44292", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43828|issue:45685", + "issue:44918|issue:45685", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:34928|issue:36010", + "issue:43039|issue:44805", + "issue:45237|issue:46015", + "issue:43866|issue:46402", + "issue:45800|issue:46304", + "issue:43296|issue:46304", + "issue:41762|issue:43596", + "issue:39290|issue:41720" + ], + "prepared_review_unit_hash": "e28d7c4fe4a737de77d8a55d006c8ddd05ec39aaf83dcdff5180fd4373b4b3af", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 64, + "serialized_chars": 46874, + "estimated_input_tokens": 11719, + "estimated_eval_tokens": 23694 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of unrelated transformer/model/runtime issues. The only clear duplicate-like pair is the two DeepSpeed ZeRO-3 Bert IndexError reports; the rest should stay separate. No PRs are present.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43596 is the cleaner anchor for the shared ZeRO-3/Bert size-0 IndexError, and 43638 reads like a narrower variant of the same failure.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43596 is the best representative issue in this set because it captures the recurring ZeRO-3/Bert loading failure in a general form.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: config attribute missing vs streaming chat/completions error." + }, + { + "left": "issue:45237", + "right": "issue:45685", + "accept": false, + "reason": "Different backends and symptoms: AMD GPU failure vs MPS int histogram limitation." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Different bug classes: image dtype mismatch vs AMD GPU runtime failure." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Different code paths: autocast dtype mismatch vs AMD GPU model failure." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Unrelated model/backend issues; no shared underlying defect." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Different model families and errors; not the same bug." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "Deepspeed collective mismatch is unrelated to ONNX export errors." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "ONNX export issue vs Qwen3Moe FP8 failure are different problems." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "Export error and CUDA assert on auto device map are unrelated." + }, + { + "left": "issue:45901", + "right": "issue:46015", + "accept": false, + "reason": "One is a task crash, the other is outdated RAG docs; not the same issue." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs problems, but they concern different removed references and would be fixed separately." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export error and GenerationMixin import error are different failures." + }, + { + "left": "issue:44918", + "right": "issue:46165", + "accept": false, + "reason": "TRL SFT embedding handling is unrelated to resuming training with trust_remote_code." + }, + { + "left": "issue:44805", + "right": "issue:45823", + "accept": false, + "reason": "Mask/index shape mismatch is not the same as Gemma4 PLE device placement." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both involve indexing errors, but the concrete failure modes and code paths differ." + }, + { + "left": "issue:43638", + "right": "issue:46165", + "accept": false, + "reason": "ZeRO-3 Bert loading failure is unrelated to training resume/trust_remote_code." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Similar IndexError wording, but the underlying tensor-shape issue is different." + }, + { + "left": "issue:44805", + "right": "issue:46294", + "accept": false, + "reason": "Different multimodal/video input bugs with distinct symptoms." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Different tensor shape/indexing bug vs GPT-OSS tensor-parallel issue." + }, + { + "left": "issue:44560", + "right": "issue:46165", + "accept": false, + "reason": "Video chat-template StopIteration is unrelated to resume-training behavior." + }, + { + "left": "issue:43596", + "right": "issue:46165", + "accept": false, + "reason": "Different failure modes and code paths." + }, + { + "left": "issue:43572", + "right": "issue:46165", + "accept": false, + "reason": "Config field regression is unrelated to training resume behavior." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor preset handling and ONNX export are unrelated." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Collective mismatch and ZeRO-3 Bert IndexError are different bugs." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same ZeRO-3/Bert size-0 IndexError family; 43638 is a narrower variant of the same underlying bug." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger cross_entropy being called is unrelated to GPT-OSS tensor parallelism." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Detokenization mismatch is unrelated to LoRA merge/save/load inconsistency." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen-3-8B NVFP4 runtime failure is unrelated to LoRA save/reload output divergence." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Activation checkpointing/FSDP recomputation mismatch is a different bug from mask indexing." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Different DeepSpeed failures: weight dimensionality vs collective mismatch." + }, + { + "left": "issue:45859", + "right": "issue:45902", + "accept": false, + "reason": "Same module name, but CPU offload incompatibility and uninitialized-memory-after-to_empty are different bugs." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "Zero-3 Bert loading failure is unrelated to GPT-OSS tensor-parallel behavior." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "FSDP checkpointing mismatch and custom trainer signature regression are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45161", + "accept": false, + "reason": "Different model systems and failure modes." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Image dtype mismatch and GraniteMoE invalid method call are unrelated." + }, + { + "left": "issue:45874", + "right": "issue:46294", + "accept": false, + "reason": "Gemma4 inputs_embeds reverse lookup is unrelated to Qwen3.5 video token mismatch." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3/vLLM attribute regression is unrelated to Mistral regex patch failure." + }, + { + "left": "issue:45823", + "right": "issue:45874", + "accept": false, + "reason": "Different Gemma4 issues: device placement vs expensive reverse embedding lookup." + }, + { + "left": "issue:43425", + "right": "issue:44292", + "accept": false, + "reason": "Torch 2.10 incompatibility and Qwen-3-8B NVFP4 runtime failure are different problems." + }, + { + "left": "issue:34928", + "right": "issue:41720", + "accept": false, + "reason": "Activation checkpointing/FSDP mismatch is unrelated to Qwen3 auto-device-map CUDA assert." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Autocast dtype mismatch is unrelated to GPT-OSS tensor parallel failure." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch and trainer signature mismatch are different bugs." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "FSDP recomputation mismatch is unrelated to GenerationMixin import failure." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger cross_entropy dispatch and tensor mask shape mismatch are different issues." + }, + { + "left": "issue:45237", + "right": "issue:46015", + "accept": false, + "reason": "AMD GPU runtime failure is unrelated to outdated RAG examples." + }, + { + "left": "issue:43866", + "right": "issue:46402", + "accept": false, + "reason": "Checkpoint corruption and missing Sam3Model weight loading are not the same bug." + }, + { + "left": "issue:45800", + "right": "issue:46304", + "accept": false, + "reason": "Torch/version incompatibility and USE_DISTRIBUTED=0 regression are separate compatibility issues." + }, + { + "left": "issue:43296", + "right": "issue:46304", + "accept": false, + "reason": "PaddleOCR-VL/vLLM loading failure is unrelated to the distributed-build regression." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": false, + "reason": "Both are ZeRO-3 IndexErrors, but on different models and code paths; too weak to merge." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues, but they concern different removed guidance and are separate edits." + }, + { + "left": "issue:43716", + "right": "issue:44560", + "accept": false, + "reason": "Different multimodal preprocessing failures." + }, + { + "left": "issue:43828", + "right": "issue:45685", + "accept": false, + "reason": "Different dtype/autocast issue vs MPS histogram integer limitation." + }, + { + "left": "issue:44918", + "right": "issue:45685", + "accept": false, + "reason": "TRl SFT embedding unpacking and MPS histogram failure are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different tensor kernel/autocast failures with no shared code path." + }, + { + "left": "issue:30277", + "right": "issue:36010", + "accept": false, + "reason": "Collective mismatch and import failure are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "Activation checkpointing/FSDP bug is unrelated to import regression." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Gemma3/vLLM attribute issue is unrelated to Qwen3 CUDA assert on device mapping." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted soft edge (43596/43638) is well supported by the shared ZeRO-3/Bert size-0 IndexError wording, and the rest of the rejections do not overstate similarity. No PRs are present, so there is no mergeability concern to flag." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 2, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36010", + "issue:39290", + "issue:41093", + "issue:41762", + "issue:43039", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43638", + "issue:43792", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43872", + "issue:43901", + "issue:43975", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44512", + "issue:44560", + "issue:44589", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45005", + "issue:45020", + "issue:45070", + "issue:45084", + "issue:45161", + "issue:45237", + "issue:45362", + "issue:45464", + "issue:45685", + "issue:45800", + "issue:45812", + "issue:45823", + "issue:45901", + "issue:45920", + "issue:46015", + "issue:46294", + "issue:46304" + ], + "soft_pairs": [ + "issue:44589|issue:46304", + "issue:45823|issue:46294", + "issue:45237|issue:45901", + "issue:43901|issue:44512", + "issue:44841|issue:45823", + "issue:43039|issue:44560", + "issue:43975|issue:46015", + "issue:45812|issue:45920", + "issue:44863|issue:46015", + "issue:45020|issue:46304", + "issue:43866|issue:44863", + "issue:43872|issue:46304", + "issue:44292|issue:46015", + "issue:39290|issue:43531", + "issue:41093|issue:43638", + "issue:43792|issue:45901", + "issue:43854|issue:43866", + "issue:43425|issue:45070", + "issue:43596|issue:45800", + "issue:43792|issue:45070", + "issue:43975|issue:45901", + "issue:45362|issue:45464", + "issue:44292|issue:45901", + "issue:39290|issue:45823", + "issue:45464|issue:45685", + "issue:43541|issue:45685", + "issue:43541|issue:43828", + "issue:43404|issue:43866", + "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": "93a1fc7acbcd7a2bd6e38f494def14a99b72bc19c39b969fbf0d9b5b588c439c", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39795, + "estimated_input_tokens": 9949, + "estimated_eval_tokens": 20154 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a mix of unrelated transformer regressions, but one duplicate family stands out: the v5 AutoTokenizer wrong-token-id regression across multiple model families. I only accept that soft duplicate; the rest are too model- or symptom-specific to merge safely.", + "confidence": 0.86, + "canonical_issue_reason": "issue:45920 is the broadest, clearest formulation of the shared AutoTokenizer token-ID regression and covers multiple impacted model families, making it the best representative of the only strong duplicate family here.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45920 is the best global issue representative because it generalizes the tokenizer regression beyond a single model and has the strongest evidence of being the shared underlying bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44589", + "right": "issue:46304", + "accept": false, + "reason": "Different failures: a Float8 storage lookup error vs a distributed-training compatibility issue." + }, + { + "left": "issue:45823", + "right": "issue:46294", + "accept": false, + "reason": "Both are model-loading/runtime bugs, but they involve different models and different code paths (device mismatch vs chat-template/video-token mismatch)." + }, + { + "left": "issue:45237", + "right": "issue:45901", + "accept": false, + "reason": "Completely different symptoms and subsystems: AMD GPU support vs table-question-answering crash." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues, but they reference different API changes and do not describe the same underlying change." + }, + { + "left": "issue:44841", + "right": "issue:45823", + "accept": false, + "reason": "Different models and different runtime failures; no evidence they share the same bug." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Unrelated features: Liger Kernel cross-entropy dispatch vs Qwen3-VL video input StopIteration." + }, + { + "left": "issue:43975", + "right": "issue:46015", + "accept": false, + "reason": "A decoding regression is unrelated to outdated RAG documentation." + }, + { + "left": "issue:45812", + "right": "issue:45920", + "accept": true, + "reason": "Same underlying AutoTokenizer v5 regression: wrong token IDs across model families, with 45920 just describing a broader set of affected models." + }, + { + "left": "issue:44863", + "right": "issue:46015", + "accept": false, + "reason": "Checkpoint loading failure vs documentation staleness; no shared code-path." + }, + { + "left": "issue:45020", + "right": "issue:46304", + "accept": false, + "reason": "Broad remote_code breakage is not the same as the USE_DISTRIBUTED=0 PyTorch compatibility issue." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Both mention loading problems, but the model families and failure modes differ." + }, + { + "left": "issue:43872", + "right": "issue:46304", + "accept": false, + "reason": "bitsandbytes constructor incompatibility is unrelated to distributed PyTorch support." + }, + { + "left": "issue:44292", + "right": "issue:46015", + "accept": false, + "reason": "Model execution/runtime error vs docs problem; unrelated." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Both mention sliding_window, but they are different models and different missing/incorrect config behavior." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Both are IndexErrors under ZeRO-3, but the symptoms and model contexts are different." + }, + { + "left": "issue:43792", + "right": "issue:45901", + "accept": false, + "reason": "Whisper model loading/running issue is unrelated to table-question-answering crashes." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Different models and distinct load failures; not the same bug." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch version incompatibility and a pydantic PretrainedConfig regression are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:45800", + "accept": false, + "reason": "Zero-3/BERT IndexError is a different failure from the torch 2.4.1 / transformers 5.8.0 incompatibility report." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper runtime failure and pydantic model breakage are unrelated." + }, + { + "left": "issue:43975", + "right": "issue:45901", + "accept": false, + "reason": "Detokenization regression is unrelated to table-question-answering crashes." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both involve Qwen3.5, but one is chat crashes and the other is streaming API failure; too different to merge." + }, + { + "left": "issue:44292", + "right": "issue:45901", + "accept": false, + "reason": "Unrelated model load/run bug vs table QA task crash." + }, + { + "left": "issue:39290", + "right": "issue:45823", + "accept": false, + "reason": "Different models and different symptom classes; no shared bug evidenced." + }, + { + "left": "issue:45464", + "right": "issue:45685", + "accept": false, + "reason": "Streaming inference failure and MPS histogram kernel limitation are unrelated." + }, + { + "left": "issue:43541", + "right": "issue:45685", + "accept": false, + "reason": "Different backend-specific MoE failures; not the same concrete bug." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Both are MoE-related, but one is torch-dynamo/grouped_mm tracing and the other is autocast dtype mismatch." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Tied-weight bug in Mistral3 is unrelated to checkpoint corruption/loading." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor failure and template compilation error are different components and failures." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "ImportError for GenerationMixin is unrelated to Gemma3 sliding_window config errors." + }, + { + "left": "issue:45800", + "right": "issue:45823", + "accept": false, + "reason": "Torch/transformers version incompatibility is not the same as the Gemma4 device-map forward mismatch." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Both are IndexErrors, but the model context and exact failure differ; not enough to treat as the same bug." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL loading failure and GGUF gpt-oss support are different model-specific issues." + }, + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Missing pad_token_idx after v5 update is unrelated to the pydantic PretrainedConfig regression." + }, + { + "left": "issue:44805", + "right": "issue:44918", + "accept": false, + "reason": "Similar shape-related wording, but the underlying failures are different (mask indexing vs embedding unpacking with TRL)." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Int4 OOM / CUDA memory growth is unrelated to tied-weights issues in translation models." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "MoE grouped_mm tracing failure is not the same as GPT-OSS tensor-parallel support problems." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Autocast dtype mismatch and streaming chat/completions failure are different bugs." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "_is_hf_initialized constructor error is unrelated to the PretrainedConfig pydantic regression." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GGUF architecture support and Qwen3.5 streaming failure are different issues." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Sliding-window config behavior and pydantic PretrainedConfig breakage are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The accepted duplicate pair, issue:45812 vs issue:45920, is well supported by the shared AutoTokenizer v5 wrong-token-ID regression. The remaining soft-edge rejections are appropriately cautious and do not overstate shared causality." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 3, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36331", + "issue:38175", + "issue:39186", + "issue:39692", + "issue:40990", + "issue:41093", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43404", + "issue:43541", + "issue:43596", + "issue:43644", + "issue:43645", + "issue:43716", + "issue:43828", + "issue:43866", + "issue:43872", + "issue:43873", + "issue:43950", + "issue:43975", + "issue:43994", + "issue:44292", + "issue:44360", + "issue:44403", + "issue:44479", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44589", + "issue:44863", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45020", + "issue:45399", + "issue:45507", + "issue:45542", + "issue:45632", + "issue:45698", + "issue:45701", + "issue:45759", + "issue:45800", + "issue:45902", + "issue:46294" + ], + "soft_pairs": [ + "issue:43716|issue:46294", + "issue:45020|issue:45800", + "issue:43541|issue:44589", + "issue:44479|issue:46294", + "issue:41093|issue:43596", + "issue:43054|issue:43994", + "issue:45759|issue:45902", + "issue:45542|issue:45800", + "issue:39186|issue:44918", + "issue:43716|issue:44918", + "issue:40990|issue:44960", + "issue:43645|issue:45399", + "issue:45632|issue:45698", + "issue:43404|issue:45005", + "issue:42915|issue:45399", + "issue:44530|issue:45005", + "issue:43873|issue:45005", + "issue:44863|issue:45005", + "issue:44360|issue:44512", + "issue:44918|issue:45507", + "issue:44492|issue:44512", + "issue:43975|issue:44292", + "issue:44292|issue:45507", + "issue:43950|issue:44534", + "issue:39692|issue:43054", + "issue:43644|issue:43950", + "issue: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": "a67f443e02065387555e888c91616393f8cae27e56125b798a5bf440c02c3d22", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36891, + "estimated_input_tokens": 9223, + "estimated_eval_tokens": 18702 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is best anchored on issue 43950, with 43644 and 44534 as the strongest near-duplicates around the same Transformers 5.x non-persistent-buffer corruption problem. The rest of the cluster is mostly adjacent but unrelated reports, so the grouping should stay fairly tentative overall.", + "confidence": 0.74, + "canonical_issue_reason": "Issue 43950 is the cleanest anchor for the shared non-persistent-buffer corruption topic: its title directly names the regression and it is the most specific representative of that bug family among the related reports.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43950 is the best single issue to represent the cluster because it most directly captures the common bug theme shared by the strongest nearby duplicates.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43716", + "right": "issue:46294", + "accept": false, + "reason": "Both are multimodal/video-related, but they describe different failures: image-preprocessor dtype mismatch versus chat-template video token/feature mismatch." + }, + { + "left": "issue:45020", + "right": "issue:45800", + "accept": false, + "reason": "These are different bug reports: a broad remote_code regression versus a specific torch/transformers version incompatibility." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Different code paths and failure modes: grouped_mm tracing errors versus missing Float8 storage." + }, + { + "left": "issue:44479", + "right": "issue:46294", + "accept": false, + "reason": "Both involve video, but the concrete regressions differ and the symptoms do not point to the same root cause." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are IndexErrors, but one is a mask/index shape mismatch and the other is a zero-sized tensor during ZeRO-3 init." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but different observed problems: degraded text embeddings versus nonsensical pipeline/AutoModel outputs." + }, + { + "left": "issue:45759", + "right": "issue:45902", + "accept": false, + "reason": "These are different issues: composite-config handling versus uninitialized-memory access after meta to_empty(cuda) materialization." + }, + { + "left": "issue:45542", + "right": "issue:45800", + "accept": false, + "reason": "TensorBoard/TensorFlow backend detection is unrelated to the torch/transformers compatibility issue." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "Both are model-loading/training errors, but the underlying failures are different: FSDP weight-shape handling versus TRL embedding unpacking." + }, + { + "left": "issue:43716", + "right": "issue:44918", + "accept": false, + "reason": "Different model families and different failure modes; these are not the same bug." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "Both are model-behavior complaints, but they concern unrelated models and problems." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Notebook custom-model initialization and flash-attn2 fallback gating are different bugs in different paths." + }, + { + "left": "issue:45632", + "right": "issue:45698", + "accept": false, + "reason": "Both involve custom code loading, but one is a cache-path collision and the other is loading the wrong module after save_pretrained." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but the concrete model families and failure modes differ too much to treat as one bug." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "FineGrainedFP8Config failure and flash-attn fallback gating are unrelated." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache group-type handling is unrelated to tied-weight loading regressions." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization offloading behavior is a different subsystem and symptom than tied-weight issues." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "NemotronH checkpoint loading failure is not the same bug as the translation-model tied-weights regression." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer discussion and a docs/runtime command issue are unrelated." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "TRl SFT embedding unpacking and GraniteMoEHybrid invalid method calls are different code paths and failures." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A typo in cache strategies is unrelated to the removed run-command docs issue." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Detokenization behavior for one model is unrelated to running a different quantized model." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Different model/runtime issues with no shared concrete failure path." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Both describe the same Transformers v5 non-persistent-buffer corruption regression." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "Related by model family only; the documentation/example issues are not the same bug as the text-embedding quality complaint." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "These reports point to the same non-persistent-buffer junk/regression; 43950 is the more specific instance." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Both mention dtype-related failures, but autocast mismatch and RLHF NaN/position-id collapse are different problems." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "CustomTrainer unexpected kwarg and ZeRO-3 zero-size index errors are unrelated." + }, + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "bitsandbytes Int8Params incompatibility and missing Float8 storage are different dependency/runtime issues." + }, + { + "left": "issue:43950", + "right": "issue:45902", + "accept": false, + "reason": "The titles are adjacent around buffer/materialization behavior, but 45902 is not clearly the same non-persistent-buffer corruption bug as 43950." + }, + { + "left": "issue:43054", + "right": "issue:45701", + "accept": false, + "reason": "Different symptom sets: SigLIP2 embedding quality versus version-dependent tokenization changes." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "A corrupted checkpoint and load-time noise are not the same underlying bug." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Same model family, but zero-probability outputs and weaker text embeddings are not enough to establish one root cause." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is a bit too strong for the packet. You can anchor the cluster on 43950, but the claims about an 'explicit repro' and 'clear regression framing' are not supported by the metadata shown here. More importantly, the accept on issue:45902 is not conservative enough from the title alone; it is adjacent but not clearly the same bug as the non-persistent-buffer corruption reports. The 43644/43950/44534 linkage is well grounded, but the rest should stay more tentative." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:36331", + "issue:41093", + "issue:41553", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43404", + "issue:43493", + "issue:43643", + "issue:43742", + "issue:43782", + "issue:43792", + "issue:43824", + "issue:43827", + "issue:43828", + "issue:43856", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44509", + "issue:44661", + "issue:44910", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue: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:44403|issue:44488", + "issue:43975|issue:45520", + "issue:43643|issue:45020", + "issue:42947|issue:43856", + "issue:41762|issue:42915", + "issue:45685|issue:45823", + "issue:43782|issue:44960", + "issue:43824|issue:43872", + "issue:44991|issue:45399", + "issue:43881|issue:43883", + "issue:45020|issue:45823", + "issue:44661|issue:45020", + "issue:43828|issue:45823", + "issue:42491|issue:44368", + "issue:45081|issue:45399", + "issue:45715|issue:46315", + "issue:45020|issue:45356", + "issue:43742|issue:45399", + "issue:43404|issue:43950", + "issue:43827|issue:44509", + "issue:43872|issue:44291", + "issue:42915|issue:43872", + "issue:43792|issue:44220", + "issue:36331|issue:41093" + ], + "prepared_review_unit_hash": "cfdd37fb995c472a1e762c6943bcaf8105a3be3e659de3f9c7932a921b99064d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36830, + "estimated_input_tokens": 9208, + "estimated_eval_tokens": 18672 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly over-broad: it mixes many unrelated model-loading, tokenizer, training, and docs regressions. Only two soft pairs look like true duplicates; the rest are similar at a high level but not the same concrete bug.", + "confidence": 0.91, + "canonical_issue_reason": "issue:43872 is the best anchor for the only clear duplicate pair: it names the concrete `_is_hf_initialized` / `init_empty_weights` TypeError, and issue:44291 reads like the same bug with a more specific wrapper title.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43872 is the most suitable global issue anchor because it is a crisp, reproducible library regression with a direct duplicate, whereas the other items are mostly model-specific or unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Different problems: Voxtral loading error message vs Molmo missing `all_tied_weights_keys`." + }, + { + "left": "issue:45663", + "right": "issue:45902", + "accept": false, + "reason": "Both are Gemma/Qwen MoE-related, but the bugs are different: FSDP2 `KeyError` vs uninitialized `inv_freq` after `to_empty(cuda)`." + }, + { + "left": "issue:44910", + "right": "issue:45859", + "accept": false, + "reason": "Different code paths: FlashAttention packed-sequence crash vs CPU offload incompatibility in a MoE rotary embedding." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "One is gradient-checkpointing not being effective with PEFT LoRA; the other is increased reserved memory under int4 quantization." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Same model family, but not enough to call it the same bug; one reports poor text embeddings, the other a HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Both are loading-related, but they concern different models and different failure modes." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "One is noisy logging on load; the other is an actual model-loading failure for ScandiBERT." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Same as above: docs/log noise vs a specific model load failure, not the same bug." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Detokenization regression and a Python 3.13 import-utils KeyError are unrelated." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Both mention remote code, but one is missing fields in `AutoConfig`, the other is a broad report of recent remote-code breakage." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Different regressions: ineffective checkpointing vs MoE training memory usage." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Different models and different failure classes: ZeRO-3 loading index error vs FineGrainedFP8Config failure." + }, + { + "left": "issue:45685", + "right": "issue:45823", + "accept": false, + "reason": "MPS integer support error and Gemma4 device mismatch are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3VL weight-only loading error vs a GLM5 issue; no same underlying bug is evident." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "ImportError for a missing class and a bitsandbytes `Int8Params` constructor regression are different bugs." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer load failure and FlashAttention fallback gating are unrelated." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Different model families and different symptoms: load failure vs missing tied-weights metadata." + }, + { + "left": "issue:45020", + "right": "issue:45823", + "accept": false, + "reason": "Remote-code breakage and Gemma4 device-map forward mismatch are distinct problems." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "Tokenizer mapping registration failure and broad remote-code regressions are not the same bug." + }, + { + "left": "issue:43828", + "right": "issue:45823", + "accept": false, + "reason": "Autocast dtype mismatch and device-placement mismatch are different runtime issues." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Version incompatibility for a trained LoRA model and a warning about `tie_word_embeddings` are not the same bug." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer regex patch crash vs FlashAttention fallback checks; different code paths." + }, + { + "left": "issue:45715", + "right": "issue:46315", + "accept": false, + "reason": "Both are tokenizer performance regressions, but they affect opposite conversion directions and different loops." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Broad remote-code regression vs Kimi-K2.5 tokenizer codec regression are not the same issue." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "Model load KeyError and FlashAttention availability gating are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "lm_head tying bug and silent corruption of non-persistent buffers are different regressions." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "These describe the same docs bug: pipeline tasks removed in v5 are still referenced in documentation." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying `_is_hf_initialized` / `init_empty_weights` TypeError during loading." + }, + { + "left": "issue:42915", + "right": "issue:43872", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and bitsandbytes `_is_hf_initialized` regression are different bugs." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": false, + "reason": "Whisper loading failure and `_torch_extract_fbank_features()` issue are too weakly connected to treat as duplicates." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Completely unrelated: trainer `compute_loss` API mismatch vs tensor mask shape mismatch." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The two accepted duplicate pairs are well supported by the packet titles: issue:43827 \u2194 issue:44509 (same docs/pipeline removal bug) and issue:43872 \u2194 issue:44291 (same `_is_hf_initialized` / `init_empty_weights` TypeError). The other soft-edge rejections are phrased cautiously and do not overclaim duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 5, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:38175", + "issue:41093", + "issue:41553", + "issue:42617", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43632", + "issue:43645", + "issue:43646", + "issue:43824", + "issue:43828", + "issue:43854", + "issue:43940", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44805", + "issue:44863", + "issue:44910", + "issue:44918", + "issue:44928", + "issue:44991", + "issue:45305", + "issue:45313", + "issue:45325", + "issue:45584", + "issue:45588", + "issue:45663", + "issue:45685", + "issue:45823", + "issue:45834", + "issue:45902", + "issue:45923", + "issue:46402" + ], + "soft_pairs": [ + "issue:43257|issue:45305", + "issue:42947|issue:45923", + "issue:41553|issue:44991", + "issue:43940|issue:45313", + "issue:44928|issue:45588", + "issue:43854|issue:44863", + "issue:41093|issue:44805", + "issue:44479|issue:44560", + "issue:41553|issue:44488", + "issue:24643|issue:30064", + "issue:44863|issue:46402", + "issue:44530|issue:44863", + "issue:43366|issue:43828", + "issue:45325|issue:45902", + "issue:45663|issue:45823", + "issue:43366|issue:43531", + "issue:44155|issue:45584", + "issue:43950|issue:45834", + "issue:44560|issue:45685", + "issue:43366|issue:44589", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:43645|issue:43646", + "issue:43550|issue:44910", + "issue:44560|issue:44918", + "issue:42617|issue:43366", + "issue:36010|issue:42915", + "issue:44661|issue:45823", + "issue:44560|issue:44805", + "issue:38175|issue:43994", + "issue:43632|issue:45834" + ], + "prepared_review_unit_hash": "4018b61d8a5c1c487d14b14caa0b57ac0af5f99bc2abbe2f0fd3bafba1ead3aa", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37144, + "estimated_input_tokens": 9286, + "estimated_eval_tokens": 18828 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly heterogeneous: the items span unrelated model-loading, attention, training, and tokenizer regressions. I only see two true duplicate-like pairs: the transformers v5 custom model initialization regression (43645/43646) and the siglip2-base-patch16-224 bad-output reports (38175/43994).", + "confidence": 0.83, + "canonical_issue_reason": "issue:43646 is the broader, cleaner statement of the v5 custom-model initialization regression; issue:43645 is the notebook-specific reproduction.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the best representative issue because it states the underlying regression in a general, actionable way without narrowing it to Jupyter.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43257", + "right": "issue:45305", + "accept": false, + "reason": "Both involve Qwen/DeepSpeed training, but one is weight conversion during load and the other is gradient averaging with GAS; different bugs and fixes." + }, + { + "left": "issue:42947", + "right": "issue:45923", + "accept": false, + "reason": "Gradient checkpointing appears in both, but one is LoRA checkpointing being ineffective and the other is a missing class flag on a trust_remote_code model." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading complaints, but they hit different models and failure modes." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 weight-loading failures, but one is Qwen3-Next with all params missing and the other is Qwen3.5 language_model weights; too model-specific to assume one fix." + }, + { + "left": "issue:44928", + "right": "issue:45588", + "accept": false, + "reason": "Different code paths: RLHF NaNs from 3D position_ids vs flash_attention crash on s_aux=None." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Both are model-loading problems, but they concern different models and errors." + }, + { + "left": "issue:41093", + "right": "issue:44805", + "accept": false, + "reason": "Same error class, but no evidence of the same root cause; likely separate model/path reproductions." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both mention video, but one is a broad v5.3.0 regression across Qwen VL variants and the other is a StopIteration in Qwen3-vl-embedding." + }, + { + "left": "issue:41553", + "right": "issue:44488", + "accept": false, + "reason": "Both are loading failures, but Voxtral error messaging and sleng-bert tokenizer loading are different issues." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "Unrelated: DeepSpeed training weight shape error vs image processor void segmentation maps." + }, + { + "left": "issue:44863", + "right": "issue:46402", + "accept": false, + "reason": "Different models and load paths; no shared underlying bug." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "PagedAttention/linear_attention crash vs checkpoint loading failure; unrelated." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF architecture support vs autocast dtype mismatch; different problems." + }, + { + "left": "issue:45325", + "right": "issue:45902", + "accept": false, + "reason": "Different regressions: Qwen2.5-VL rope scaling vs Qwen3_5Moe inv_freq materialization." + }, + { + "left": "issue:45663", + "right": "issue:45823", + "accept": false, + "reason": "Same model family, but one is an FSDP2 shared_kv_states KeyError and the other is a device-map mismatch." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "GGUF gpt-oss support and Qwen3-MoE sliding_window are unrelated." + }, + { + "left": "issue:44155", + "right": "issue:45584", + "accept": false, + "reason": "AudioFlamingo3 track leakage vs Whisper empty transcription handling; unrelated." + }, + { + "left": "issue:43950", + "right": "issue:45834", + "accept": false, + "reason": "Non-persistent buffer corruption vs OCR index error; different bugs." + }, + { + "left": "issue:44560", + "right": "issue:45685", + "accept": false, + "reason": "Video StopIteration vs MPS histogram limitation; unrelated." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF support vs Float8 storage lookup failure; unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Missing exported class vs meta-device loading regression; different failure modes." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "FineGrainedFP8Config failure vs meta-device loading regression; unrelated." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Same v5 custom model initialization regression; the notebook report is just a narrower reproduction of the broader custom-model init bug." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "torch.compile SDPA crash on Bamba vs flash attention illegal memory access on Qwen3.5; different code paths." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Video inference StopIteration vs TRL SFT input-embedding unpacking; unrelated." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py runtime issue vs GGUF gpt-oss support; unrelated." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "GenerationMixin import break vs Qwen3Moe FP8 config failure; unrelated." + }, + { + "left": "issue:44661", + "right": "issue:45823", + "accept": false, + "reason": "`add-new-model-like` tokenizer mapping bug vs Gemma4 device mismatch; different subsystems." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different models and symptoms; both are generic shape/runtime errors but no shared root cause." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same siglip2-base-patch16-224 output regression; one reports zero probabilities, the other nonsensical AutoModel/pipeline outputs." + }, + { + "left": "issue:43632", + "right": "issue:45834", + "accept": false, + "reason": "Transformers v5 `_is_hf_initialized` regression and Kosmos2.5 OCR index error are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted near-duplicates are the clearly related Transformers v5 custom-model initialization reports (43645/43646) and the siglip2-base-patch16-224 output regression reports (38175/43994); the other pair rejections are cautious and supported by the issue titles." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43425", + "issue:43596", + "issue:43638", + "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:43638|issue:44661", + "issue:43596|issue:44661", + "issue:44877|issue:45030", + "issue:44661|issue:44805", + "issue:43317|issue:44756", + "issue:38175|issue:39692", + "issue:43644|issue:44534", + "issue:42491|issue:43720", + "issue:41720|issue:42915", + "issue:44960|issue:44977", + "issue:42915|issue:44756", + "issue:44451|issue:44488", + "issue:42491|issue:45276", + "issue:45313|issue:45357", + "issue:44488|issue:44991", + "issue:39692|issue:43994", + "issue:43708|issue:44075", + "issue:43644|issue:43645", + "issue:45488|issue:45812", + "issue:43856|issue:43873", + "issue:43856|issue:44387", + "issue:43854|issue:43931", + "issue:43257|issue:43854", + "issue:45362|issue:45406", + "issue:44960|issue:46050", + "issue:42491|issue:45292", + "issue:42491|issue:43931", + "issue:44164|issue:44291", + "issue:44910|issue:45588", + "issue:45276|issue:45335", + "issue:43425|issue:43957" + ], + "prepared_review_unit_hash": "3d8af0154bbb4f7e9c879d43829d2ced4d9b9c23e24bf26cb2cafc6a1eef5b61", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "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": "This cluster is mostly a loose grab bag of distinct Transformers regressions, with only a few true duplicate pairs around the same concrete bug. The strongest duplicate groups are the non-persistent buffer initialization regression, the resize_token_embeddings propagation bug, and the repeated BERT loading failure reports.", + "confidence": 0.76, + "canonical_issue_reason": "issue:44488 is the best representative issue here: it is a concrete, user-facing loading regression, has a direct follow-up report (issue:44451), and is broader/more established than the earlier report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44488 is the strongest global issue candidate because it most clearly captures a real duplicate pair and has the best combination of breadth and discussion activity.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "Different failures in different code paths: ZeRO-3/BertModel indexing versus add-new-model-like tokenizer mapping behavior." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "Both mention BertModel-related failures, but the underlying bug reports are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both are config-validation complaints, but they target different models and different validation surfaces." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer mapping registration and mask shape IndexError are unrelated bugs." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Both involve loading/offload behavior, but one is dequantized model device mapping and the other is mmap/OOM on Strix Halo." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Same model family, but one is a zero-probability runtime issue and the other is a documentation/example error with additional quantization problems." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "These are the same non-persistent-buffer regression in Transformers 5, just reported under two titles." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Different model stacks and failures: Qwen3 MoE LoRA loading versus BitNet packed-weight unpacking." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Both are Qwen3/Qwen3-MoE related, but the concrete failures and triggers differ." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 flash-attention generation failures are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3-MoE FP8 loading and Strix Halo mmap/OOM are separate issues." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": true, + "reason": "Both report the same BERT loading regression on different models, with 44488 as the broader follow-up report." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Qwen3 MoE LoRA loading and resize_token_embeddings propagation are different problems." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is ZeRO-3 weight loading and the other is save_pretrained visual-encoder key regression." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer/load regressions, but they affect different model families and likely different root causes." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both involve SigLIP2, but the docs/example issue also includes quantization failures and is not clearly the same underlying bug as the nonsensical-output report." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer checkpoint resume math and SGD optimizer-arg handling are unrelated." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Shared Transformers 5 context is too broad; one is buffer initialization junk, the other is Jupyter custom-model initialization." + }, + { + "left": "issue:45488", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and different tokenizer code paths." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Training memory usage and quantized offloading behavior are different issues." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both mention memory pressure, but one is Qwen3 MoE training efficiency and the other is int4 quantization OOM in Transformers 5.x." + }, + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "GLM-4.7-Flash test loading and Qwen3-VL weight-shape mismatch are unrelated model-loading bugs." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MoE accelerate+deepspeed weight conversion and GLM-4.7-Flash unit-test loading do not share the same root cause." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Qwen3.5 chat crash and Gemma4Processor serve crash are distinct failures." + }, + { + "left": "issue:44960", + "right": "issue:46050", + "accept": false, + "reason": "GLM5 and sinq save/load quantization breakage are unrelated." + }, + { + "left": "issue:42491", + "right": "issue:45292", + "accept": false, + "reason": "LoRA model compatibility on Qwen3 MoE is not the same bug as resize_token_embeddings not updating output embeddings." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Different model families and failure modes." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "Both touch loading/saving, but extra_state handling and unexpected _is_hf_initialized arguments are different mechanics." + }, + { + "left": "issue:44910", + "right": "issue:45588", + "accept": false, + "reason": "Both are flash-attention crashes, but the triggering conditions and code-path failures are different." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Same resize_token_embeddings propagation bug, reported on two model variants." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version incompatibility and meta-device model loading are separate problems." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and the soft-edge decisions are conservative. The accepted pairs are well supported by title similarity and cluster metadata, and the rejected pairs avoid overclaiming when only broad thematic overlap is present. The canonical issue choice for issue:44488 is also reasonable given its discussion activity and the close 44451/44488 pair." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43404", + "issue:43425", + "issue:43454", + "issue:43645", + "issue:43646", + "issue:43708", + "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:43404|issue:43454", + "issue:41720|issue:44155", + "issue:39692|issue:44960", + "issue:44402|issue:44488", + "issue:42947|issue:45704", + "issue:36010|issue:39404", + "issue:43873|issue:44387", + "issue:45663|issue:45759", + "issue:44291|issue:44756", + "issue:45663|issue:45859", + "issue:44155|issue:45823", + "issue:44387|issue:44756", + "issue:44292|issue:44912", + "issue:43957|issue:44756", + "issue:43335|issue:45663", + "issue:42947|issue:43381", + "issue:45276|issue:45292", + "issue:44164|issue:44756", + "issue:44479|issue:44877", + "issue:43854|issue:45216", + "issue:43381|issue:44928", + "issue:45684|issue:45698", + "issue:45588|issue:45663", + "issue:43646|issue:43708", + "issue:43645|issue:44164", + "issue:45812|issue:46315", + "issue:43828|issue:43957", + "issue:45292|issue:45335", + "issue:43425|issue:43828", + "issue:44402|issue:44451", + "issue:44589|issue:44756", + "issue:39404|issue:45701", + "issue:44683|issue:45823", + "issue:43425|issue:45823" + ], + "prepared_review_unit_hash": "d297deeff14f51190a8914a58e4a854ddad4b743dbced83337cb8f5446355f79", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37960, + "estimated_input_tokens": 9490, + "estimated_eval_tokens": 19236 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous; nearly all soft pairs are related only by broad subsystem similarity, not the same concrete bug. The only clear duplicate-like subcluster is around `resize_token_embeddings` propagation to tied/output embeddings.", + "confidence": 0.36, + "canonical_issue_reason": "`issue:45292` is the clearest, most general representative of the resize-token-embedding regression; `issue:45276` and `issue:45335` look like narrower manifestations of the same fix area.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:45292` is the most succinct and broadly framed report of the only near-duplicate issue family in this set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both are weight-tying complaints, but they concern different models and likely different code paths." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "Different models and failures: CUDA assert during device mapping vs cross-track token/embedding leakage." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "Docs/example errors for SigLIP2 vs an unrelated GLM5 issue; no shared underlying bug." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Both mention model loading/tokenizer issues, but the affected models and failure modes differ." + }, + { + "left": "issue:42947", + "right": "issue:45704", + "accept": false, + "reason": "Gradient checkpointing/LoRA vs an Apex RMSNorm memory leak are unrelated bugs." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "Import path breakage for `GenerationMixin` is unrelated to Whisper pipeline `return_language` behavior." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "Both touch quantization/offloading, but one is offloading failure and the other is increased reserved memory/OOM." + }, + { + "left": "issue:45663", + "right": "issue:45759", + "accept": false, + "reason": "Different models and problems: Gemma4 FSDP attention state vs Qwen config unwrapping." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "`init_empty_weights` argument handling and Strix Halo mmap/OOM are unrelated." + }, + { + "left": "issue:45663", + "right": "issue:45859", + "accept": false, + "reason": "Unrelated model-specific issues: Gemma4 FSDP key error vs Qwen3.5-MoE CPU offload incompatibility." + }, + { + "left": "issue:44155", + "right": "issue:45823", + "accept": false, + "reason": "AudioFlamingo batched inference leakage and Gemma4 device mismatch are different bugs." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Both mention memory issues, but the causes and subsystems are different." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Different quantization/model-loading regressions with no concrete shared code-path." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading breakage is unrelated to the mmap/OOM report." + }, + { + "left": "issue:43335", + "right": "issue:45663", + "accept": false, + "reason": "SwitchTransformers sparse-layer construction and Gemma4 FSDP attention state are unrelated." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both are about gradient checkpointing, but one is ignored with PEFT LoRA and the other is an eval-mode restriction." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": true, + "reason": "Same `resize_token_embeddings` propagation bug; `45276` is a model-specific instance of the same missing update to output embeddings." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "Checkpoint save/load `extra_state` handling is unrelated to Strix Halo mmap OOM." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video-input regression for Qwen VL models and strict Granite config loading are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different model-loading regressions with no shared underlying failure." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Gradient checkpointing in eval mode is unrelated to RLHF gradient explosion from 3D `position_ids`." + }, + { + "left": "issue:45684", + "right": "issue:45698", + "accept": false, + "reason": "Both involve save/load of custom models, but one is a permissions propagation issue and the other is wrong-module loading." + }, + { + "left": "issue:45588", + "right": "issue:45663", + "accept": false, + "reason": "Flash-attention `s_aux=None` crash and Gemma4 FSDP attention-state rebuilding are unrelated." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization breakage and checkpoint resume batch-size math are different problems." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook custom-model initialization and `extra_state` save/load handling are not the same bug." + }, + { + "left": "issue:45812", + "right": "issue:46315", + "accept": false, + "reason": "Wrong token IDs and tokenizer performance regression are different failure modes." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Autocast dtype mismatch and meta-device loading failures are unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same resizing propagation bug across architectures: both report `resize_token_embeddings` failing to update tied embeddings." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version compatibility and Phi MoE autocast dtype mismatch are unrelated." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer vocab mismatch for one model and load failure for another do not look like the same defect." + }, + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Float8 storage lookup error and mmap/OOM are unrelated." + }, + { + "left": "issue:39404", + "right": "issue:45701", + "accept": false, + "reason": "Whisper pipeline `return_language` regression is unrelated to tokenizer output changing across versions." + }, + { + "left": "issue:44683", + "right": "issue:45823", + "accept": false, + "reason": "Compiled flex attention on new torch versions and Gemma4 device mismatch are unrelated." + }, + { + "left": "issue:43425", + "right": "issue:45823", + "accept": false, + "reason": "Torch 2.10 compatibility and Gemma4 `device_map=\"auto\"` forward mismatch are different issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The resize_token_embeddings cluster (45276/45292/45335) is the only clearly supported near-duplicate family, and the other pairwise rejections are generally cautious and aligned with the issue titles. I don\u2019t see an overstatement that would require rejection." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 8, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:43644", + "issue:43645", + "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:43611|issue:43645", + "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:44589|issue:45084", + "issue:43906|issue:46326", + "issue:43425|issue:44918", + "issue:43901|issue:44509", + "issue:46050|issue:46326", + "issue:43994|issue:44079", + "issue:36032|issue:43452", + "issue:43299|issue:43931", + "issue:43596|issue:44683", + "issue:44164|issue:45759", + "issue:44671|issue:44991", + "issue:44991|issue:45081", + "issue:43856|issue:44756", + "issue:39404|issue:44671", + "issue:44164|issue:44479", + "issue:44488|issue:45081", + "issue:44737|issue:45859", + "issue:44661|issue:44683", + "issue:44488|issue:44987", + "issue:43582|issue:43632", + "issue:32090|issue:39404", + "issue:43611|issue:43644", + "issue:43716|issue:45084", + "issue:46285|issue:46402", + "issue:40990|issue:43421", + "issue:36032|issue:42915" + ], + "prepared_review_unit_hash": "a47c9c1f9bf57fe409a20d3e11f6d4e5040e04dc9188671147834f4f230144d6", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37700, + "estimated_input_tokens": 9425, + "estimated_eval_tokens": 19106 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These issues are mostly distinct regressions or feature requests across different models and subsystems; the soft-linked pairs share broad themes but not the same concrete bug or fix. No duplicate cluster emerges here.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both relate to Transformers v5 model-loading regressions, but one is about `base_model_prefix` handling and the other about custom model init in Jupyter; different failure modes and fixes." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Unrelated problems: torch version incompatibility versus an image-preprocessor/model dtype mismatch for Mistral-3." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Both are crashes, but one is Qwen3.5 cache/group-type handling and the other is a template compilation error; not the same code path." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "One is a reproduction of another issue, but the topics differ: an isolated prior bug reproduction versus a Qwen3.5 PagedAttentionCache crash." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper `return_language` pipeline regression and MXFP4 quantization fallback are unrelated model/pipeline bugs." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "Different areas: `torch.export`/`torch_compilable_check` versus Perceiver image classification with non-default resolution." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Both involve runtime shape/path failures, but one is a video-input regression for Qwen VL models and the other is a generic mask/tensor index error." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Distinct failures: missing Float8 storage type versus template compilation on non-template nodes." + }, + { + "left": "issue:43906", + "right": "issue:46326", + "accept": false, + "reason": "An isolated reproduction of an earlier issue is not the same bug as the mistral_common `continue_final_message` tokenizer bug." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch incompatibility is not the same as Qwen3.5 embedding unpacking with TRL SFT trainer." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "One is a docs/behavior mismatch for text classification; the other is documentation still referencing removed pipeline tasks." + }, + { + "left": "issue:46050", + "right": "issue:46326", + "accept": false, + "reason": "Quantized-model save/load corruption and mistral_common tokenizer state handling are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "SigLIP2 model/pipeline outputs and `ModelOutput` key assignment with `None` are different bugs in different layers." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "Tokenizer special-method conflict and `gguf_file` breaking model/tokenizer loading are different loading regressions." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both affect Qwen3-VL/Qwen3VL loading, but one is a Moe-model version regression and the other is a weight-shape mismatch for a specific checkpoint." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "DeepSpeed ZeRO-3 init indexing error and compiled flex_attention failure on newer torch are unrelated." + }, + { + "left": "issue:44164", + "right": "issue:45759", + "accept": false, + "reason": "`extra_state` save/load handling is not the same as `from_config` not unwrapping `text_config` for composite multimodal configs." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT masked LM prediction regression and tokenizer loading failure for a specific model are different issues." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both involve tokenizer loading, but one is an EMBEDDIA tokenizer failure and the other is a Mistral regex patch crash; different causes." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE training memory inefficiency is not the same as disabling mmap on Strix Halo to avoid OOM." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline language-return regression and CamemBERT masked LM predictions are unrelated model-specific bugs." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "`extra_state` serialization and Qwen VL video-input regression are distinct code paths." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Loading `cjvt/sleng-bert` and Mistral tokenizer regex patching are different tokenizer/model compatibility issues." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error and Whisper pipeline `return_language` regression are unrelated." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 regressions, but one is model loading via `base_model_prefix` and the other is non-persistent buffer initialization; different underlying bugs." + }, + { + "left": "issue:43716", + "right": "issue:45084", + "accept": false, + "reason": "Image preprocessor dtype mismatch and template compilation errors do not share a concrete failing code path." + }, + { + "left": "issue:46285", + "right": "issue:46402", + "accept": false, + "reason": "CLIP structural inconsistency and Sam3 text encoder weights not loading are separate model-architecture regressions." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity anomaly on gpt-oss-20b and tokenizer runtime post-processor updates are unrelated." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer method conflict and Qwen3Moe fine-grained FP8 failure are distinct issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "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:43643", + "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: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:43643|issue:44530", + "issue:34567|issue:40444", + "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:42175|issue:46178", + "issue:40444|issue:46326", + "issue:42915|issue:44530", + "issue:44811|issue:45584", + "issue:43644|issue:43646", + "issue:43452|issue:43482", + "issue:44206|issue:44220", + "issue:43883|issue:44488", + "issue:43317|issue:43482", + "issue:36032|issue:43645", + "issue:45084|issue:45507", + "issue:44987|issue:44991", + "issue:44849|issue:45325", + "issue:44464|issue:45987", + "issue:43746|issue:46326", + "issue:43262|issue:43377", + "issue:43645|issue:44534", + "issue:43646|issue:44534", + "issue:44403|issue:45081", + "issue:42915|issue:45084", + "issue:43404|issue:44466", + "issue:43883|issue:44991", + "issue:43994|issue:44610", + "issue:43322|issue:43540", + "issue:43089|issue:45987", + "issue:43377|issue:43819", + "issue:44451|issue:45081", + "issue:44987|issue:45081", + "issue:44265|issue:44610", + "issue:44991|issue:46315", + "issue:42915|issue:44291" + ], + "prepared_review_unit_hash": "b82d389e8bc4bf26da74f810d7731125a8e5137c1423a00f52d66f0d67634bc4", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38378, + "estimated_input_tokens": 9595, + "estimated_eval_tokens": 19446 + }, + "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 and model-loading bugs. A few pairs are superficially similar (especially v5/model-loading complaints), but none are strong enough to treat as the same underlying bug with mergeable fixes.", + "confidence": 0.74, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "Different failures: AutoConfig fields missing vs Qwen3.5 PagedAttentionCache linear_attention crash." + }, + { + "left": "issue:34567", + "right": "issue:40444", + "accept": false, + "reason": "TrainerState token-count tracking bug vs Qwen2.5-VL IterableDataset multi-image finetuning failure." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are load failures, but for different models and likely different root causes." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Grouped_mm torchdynamo tracing error is unrelated to template compilation TypeError." + }, + { + "left": "issue:42175", + "right": "issue:46178", + "accept": false, + "reason": "Backend packaging omission vs BACKENDS_MAPPING error for tensorflow_text are not the same bug." + }, + { + "left": "issue:40444", + "right": "issue:46326", + "accept": false, + "reason": "Multi-image video/audio dataset bug vs mistral_common continue_final_message tokenizer bug." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "Qwen3Moe FP8 config failure is distinct from Qwen3.5 cache group-type crash." + }, + { + "left": "issue:44811", + "right": "issue:45584", + "accept": false, + "reason": "Both involve Whisper, but one is batch_decode token skipping and the other is empty-transcription generation after alignment." + }, + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Non-persistent buffer initialization corruption is not the same as custom model initialization breakage." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "GGUF loader breakage is plausible, but one is generic from_pretrained/gguf_file and the other is a specific model loading report." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Both audio-feature related, but the titles point to different failing layers and symptoms." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing all_tied_weights_keys attribute vs specific model loading failure are unrelated." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Dequantized offload device_map failure is unrelated to Qwen2.5-GGUF loading." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer method-name conflict vs notebook custom-model init regression are different bugs." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template-node compilation TypeError and GraniteMoEHybrid invalid-method call are different code paths." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are v5 loading failures, but on different models/tokenizers and with no shared concrete root cause." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden-states bug is unrelated to Qwen2.5-VL rope-index scaling." + }, + { + "left": "issue:44464", + "right": "issue:45987", + "accept": false, + "reason": "Chunked generation compiled-forward inconsistency is unrelated to StaticCache seq_length return-type bug." + }, + { + "left": "issue:43746", + "right": "issue:46326", + "accept": false, + "reason": "PEFT checkpoint loading issue and mistral_common continue_final_message bug are distinct." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio sampling-rate defaulting and MIMI padding-mask batching bug are different issues." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Notebook custom model init regression is not the same as non-persistent buffer junk values." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization breakage and buffer-junk regression are separate failure modes." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Loading noise issue is unrelated to the mistral regex patch crash." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and non-template node compilation failure do not match." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Both mention lm_head/tied weights, but one is incorrect tying and the other is inconsistent serialization by device." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Missing all_tied_weights_keys attribute is unrelated to tokenizer loading failure for a specific model." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 bad results and OmDet-Turbo processor size mismatch are different model/processor bugs." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Segmentation fault loading Llava Next and Qwen3OmniMoe video ValueError are unrelated multimodal failures." + }, + { + "left": "issue:43089", + "right": "issue:45987", + "accept": false, + "reason": "Generation performance overhead is unrelated to StaticCache return-type contract violation." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "Both are audio-codec correctness bugs, but they involve different models and different missing pieces." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "Specific model load failure vs Mistral tokenizer regex crash are not the same underlying bug." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Loading a specific model and crashing in fix_mistral_regex are different issues." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export/exportability failure is unrelated to OmDet-Turbo processor input-size mismatch." + }, + { + "left": "issue:44991", + "right": "issue:46315", + "accept": false, + "reason": "Tokenizer load failure and convert_tokens_to_ids performance regression are different problems." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "FP8 config failure and init_empty_weights unexpected-argument failure are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:44220", + "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:44220|issue:44683", + "issue:45216|issue:45357", + "issue:44062|issue:44987", + "issue:36032|issue:43742", + "issue:42915|issue:44589", + "issue:46209|issue:46326", + "issue:44479|issue:45325", + "issue:43632|issue:44756", + "issue:36032|issue:41720", + "issue:43531|issue:44877", + "issue:43296|issue:43531", + "issue:43404|issue:45127", + "issue:40444|issue:45325", + "issue:43425|issue:44403", + "issue:43611|issue:44534", + "issue:43792|issue:44589", + "issue:40990|issue:44811", + "issue:44403|issue:44991", + "issue:43632|issue:44534", + "issue:44079|issue:45663", + "issue:33666|issue:40444", + "issue:44683|issue:45020", + "issue:43421|issue:44568", + "issue:43381|issue:45923", + "issue:44291|issue:44589", + "issue:43482|issue:45538", + "issue:35707|issue:40444", + "issue:44829|issue:45479", + "issue:44462|issue:45812", + "issue:43482|issue:43525" + ], + "prepared_review_unit_hash": "f9546b8a7f01efe082c75d8100b87658dd3b00dc7bb66c822a65400a633236b4", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36644, + "estimated_input_tokens": 9161, + "estimated_eval_tokens": 18578 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a noisy mix of unrelated Transformers regressions. Only one soft pair looks like a true duplicate: `AutoTokenizer` ignoring `tokenizer.json` appears to explain the Granite wrong-token-ID regression.", + "confidence": 0.63, + "canonical_issue_reason": "`issue:44462` is the clearest root-cause report: it describes `AutoTokenizer` ignoring `tokenizer.json`, which is broad enough to cover the Granite token-ID failure in `issue:45812`.", + "canonical_pr_reason": null, + "best_issue_reason": "`issue:44462` is the best representative issue because it states the underlying loading bug directly and is broader than the downstream Granite-specific symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "Audio feature extraction vs compiled flex-attention; unrelated code paths." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": false, + "reason": "Both touch Qwen3.5 save_pretrained, but one is a broad checkpoint regression and the other is a specific visual-encoder key issue; not enough to prove the same bug." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer `AddedToken` argument conflict vs a model-loading failure for a specific repo; different failure modes." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "T5 tokenizer method-name conflict vs MobileLLM loading key error; unrelated." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "FP8 config failure on Qwen3MoE vs a missing Float8 storage type error; different underlying issues." + }, + { + "left": "issue:46209", + "right": "issue:46326", + "accept": false, + "reason": "FP8 inference on B200 vs tokenizer `continue_final_message`; unrelated." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "General video-input regression across Qwen VL models vs a still-image temporal \u0564\u056b\u0580position-id scaling bug; distinct multimodal paths." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "HF initialization flag regression vs mmap/OOM behavior on Strix Halo; unrelated." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Tokenizer load conflict vs CUDA assert during auto device mapping; unrelated." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior vs Granite speech config loading; unrelated." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL load failure in vLLM vs Qwen3-MoE sliding-window issue; unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Both involve tied embeddings, but one is a missing tie in Mistral3 image-text generation and the other is LoRA merge collapse after vocab extension; different bugs." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Multi-image IterableDataset finetuning failure vs still-image rope scaling; different bugs." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Torch version incompatibility vs noisy loading output; unrelated." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Base-model-prefix loading regression vs non-persistent buffer initialization junk; different code paths." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Whisper runtime failure vs Float8 storage lookup error; unrelated." + }, + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "High perplexity on GPT-OSS vs Whisper batch_decode skip_special_tokens behavior; unrelated." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Load-time noise vs tokenizer loading failure for EMBEDDIA/est-roberta; not the same bug." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but `_is_hf_initialized` and non-persistent buffer corruption are separate issues." + }, + { + "left": "issue:44079", + "right": "issue:45663", + "accept": false, + "reason": "ModelOutput key assignment bug vs Gemma4 FSDP2 shared_kv_states KeyError; unrelated." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Multi-GPU Qwen2-VL training vs Qwen2.5-VL multi-image IterableDataset failure; related model family, but different concrete problems." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Flex-attention compilation on torch>=2.9 vs remote_code loading regressions; unrelated." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Runtime special-token post-processor update vs add_special_tokens not adding BOS/EOS; same general area, but different concrete bugs." + }, + { + "left": "issue:43381", + "right": "issue:45923", + "accept": false, + "reason": "Eval-mode checkpointing restriction vs missing supports_gradient_checkpointing flag; different issues." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` arg with init_empty_weights vs Float8 storage lookup failure; unrelated." + }, + { + "left": "issue:43482", + "right": "issue:45538", + "accept": false, + "reason": "GGUF loading failure vs CLIPTokenizer max-length constant; unrelated." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values vs Qwen2.5-VL multi-image finetuning; unrelated." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Flash-attention-3 training degeneration vs single-label num_labels=1 zero-loss behavior; different failure causes." + }, + { + "left": "issue:44462", + "right": "issue:45812", + "accept": true, + "reason": "Same tokenizer-loading path: ignoring `tokenizer.json` can directly produce the wrong Granite token IDs, so these look like one underlying bug." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "GGUF loading failure vs missing `pad_token_id` on Llama4Config; unrelated." + }, + { + "left": "issue:44987", + "right": "issue:43525", + "accept": false, + "reason": "Specific model load failure under transformers>=5.1 vs missing `pad_token_id`; different issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays mostly conservative. The only positive soft pair, `issue:44462` / `issue:45812`, is a reasonable duplicate candidate based on the titles, and the rest of the soft-edge rejections avoid overclaiming. The canonical issue choice is also defensible as the broader tokenizer-loading report." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 11, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:40444", + "issue:41553", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43454", + "issue:43482", + "issue:43531", + "issue:43582", + "issue:43606", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43701", + "issue:43746", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43873", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44466", + "issue:44479", + "issue:44534", + "issue:44877", + "issue:44898", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45491", + "issue:45542", + "issue:45701", + "issue:45800", + "issue:45823", + "issue:45834", + "issue:46402" + ], + "soft_pairs": [ + "issue:43317|issue:43856", + "issue:42175|issue:45542", + "issue:41553|issue:43582", + "issue:43818|issue:46402", + "issue:44079|issue:44534", + "issue:43531|issue:44291", + "issue:43818|issue:45357", + "issue:41720|issue:45081", + "issue:43377|issue:45491", + "issue:40444|issue:45381", + "issue:44265|issue:45072", + "issue:43644|issue:44079", + "issue:43749|issue:44079", + "issue:42175|issue:43296", + "issue:43611|issue:43646", + "issue:45325|issue:45381", + "issue:43792|issue:43957", + "issue:43994|issue:44898", + "issue:43632|issue:45072", + "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:43296|issue:45800", + "issue:43873|issue:45701", + "issue:43606|issue:44265", + "issue:43262|issue:43746" + ], + "prepared_review_unit_hash": "4ebaabb18230cac542a3256b27a08e7f82e59e0a047c2f3553d9542686e16e2c", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35894, + "estimated_input_tokens": 8974, + "estimated_eval_tokens": 18204 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers bugs and regressions. Only the two Qwen2.5-VL position-id reports look like the same underlying issue; the rest are thematic but not true duplicates.", + "confidence": 0.68, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:45325 is the clearest concrete report of the Qwen2.5-VL rope/position-id regression and best represents the only likely duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Offloading/quantization load failure is unrelated to Qwen3 MoE training memory usage." + }, + { + "left": "issue:42175", + "right": "issue:45542", + "accept": false, + "reason": "Both mention TensorFlow/backend packaging, but one is missing TF in the install extra and the other is a tensorboard-only environment error." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Tokenizer loading error messaging and Apple Silicon allocator TypeError are different bugs." + }, + { + "left": "issue:43818", + "right": "issue:46402", + "accept": false, + "reason": "Video-LLaVA temporal-attention/weight-sharing bug is unrelated to Sam3 text encoder weight loading." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "ModelOutput key assignment and non-persistent buffer initialization are separate code paths." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior is unrelated to the init_empty_weights argument regression." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Video-LLaVA weight sharing and Qwen3.5 visual-key serialization are different model families and failures." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "CUDA assert in Qwen3 auto device mapping is unrelated to the Mistral tokenizer regex patch crash." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both are batching issues, but one is a missing padding-mask bug in MIMI and the other is a Gemma3 sliding-window NaN issue." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Qwen2.5-VL multi-image finetuning failure and video vision_position_ids regression are related only loosely, not the same bug." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "torch.export failures from torch_compilable_check are unrelated to bfloat16 dtype mismatches." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Non-persistent buffer junk and ModelOutput key assignment are different regressions." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and ModelOutput key handling do not share the same failure mode." + }, + { + "left": "issue:42175", + "right": "issue:43296", + "accept": false, + "reason": "TF backend packaging and PaddleOCR-VL loading in vLLM are separate dependency/model-loading problems." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Both are v5 loading regressions, but one is about base_model_prefix and the other about custom model initialization; not the same concrete bug." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": true, + "reason": "Both point to the same Qwen2.5-VL 5.3.0 position-id/rope-index regression affecting image and video inputs." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper-large-v2 runtime failure and meta-device loading regression are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Siglip2 nonsensical outputs and Perceiver non-default-resolution failure are different model-specific issues." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "_is_hf_initialized flag regression is unrelated to dtype mismatch during inference." + }, + { + "left": "issue:45325", + "right": "issue:45834", + "accept": false, + "reason": "Qwen2.5-VL position ids and Kosmos2.5 OCR index errors are different bugs." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Missing lm_head tie and inconsistent lm_head.weight serialization are related only broadly, not the same bug." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Qwen3MoE FP8 failure and Qwen2.5-GGUF loading failure are separate regressions." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint key mismatch and Qwen2.5-VL rope-index bug are unrelated." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument and strict granite_speech config loading are different compatibility issues." + }, + { + "left": "issue:44479", + "right": "issue:45823", + "accept": false, + "reason": "Qwen video-input regression and Gemma4 device mismatch are unrelated model-specific failures." + }, + { + "left": "issue:43296", + "right": "issue:45800", + "accept": false, + "reason": "PaddleOCR-VL loading in vLLM and torch 2.4.1 vs transformers 5.8.0 incompatibility are not the same bug." + }, + { + "left": "issue:43873", + "right": "issue:45701", + "accept": false, + "reason": "Quantization offloading behavior and tokenization changes across versions are separate concerns." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU-offload device mismatch for bark-small is unrelated to torch.export and torch_compilable_check." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and GraniteSpeech PEFT checkpoint loading are different issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The analysis is grounded in the packet and stays conservative. The only accepted duplicate, issue:45325 vs issue:45381, is plausibly the same Qwen2.5-VL position-id/rope regression based on the titles and version context. The remaining soft-edge rejections are generally cautious and do not overclaim duplicate status from thematic similarity alone." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:44387", + "issue:44466", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44610", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45325", + "issue:45357", + "issue:45584", + "issue:45704", + "issue:45759", + "issue:45823", + "issue:45920", + "issue:45923", + "issue:46050", + "issue:46291", + "issue:46305", + "issue:46315" + ], + "soft_pairs": [ + "issue:40990|issue:45584", + "issue:34634|issue:35707", + "issue:44568|issue:45920", + "issue:35707|issue:41720", + "issue:44186|issue:44265", + "issue:41720|issue:45823", + "issue:43606|issue:43819", + "issue:43454|issue:43550", + "issue:43881|issue:44488", + "issue:39692|issue:40444", + "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:44991|issue:46291", + "issue:43749|issue:44534", + "issue:42617|issue:42915", + "issue:41720|issue:44977", + "issue:44466|issue:45357", + "issue:43526|issue:43883", + "issue:43881|issue:44991", + "issue:44387|issue:45704" + ], + "prepared_review_unit_hash": "03963af0848c8adeca403a34eefeb2c47278f2b73007e3183c4aab7e226d6654", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36855, + "estimated_input_tokens": 9214, + "estimated_eval_tokens": 18684 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is thematically broad but not duplicate-tight: it mixes tokenizer regressions, model-loading/device-mapping failures, quantization/offloading bugs, and a few unrelated model-specific issues. None of the proposed soft edges look like the same underlying bug/change.", + "confidence": 0.82, + "canonical_issue_reason": "Issue 43873 is the broadest representative because it centers on offloading plus quantization, which is the nearest shared theme across several neighboring reports. It is still only a cluster representative, not a true duplicate root.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the best global issue pick for this cluster because it captures the widest common failure mode (offloading/quantization) without being overly specific to one model.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated symptoms: high perplexity on one model vs empty-transcription generation failure on Whisper." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "Different subsystems and failure modes: Bark voice preset vs progressive generation with inputs_embeds/past_key_values." + }, + { + "left": "issue:44568", + "right": "issue:45920", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different code paths and model families; not the same concrete bug." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Generation API issue vs device-mapping CUDA assert; no shared root cause." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "Tokenizer NER/padding crash and torch.export torch_compilable_check failure are distinct issues." + }, + { + "left": "issue:41720", + "right": "issue:45823", + "accept": false, + "reason": "Both involve device mapping, but the models and failure mechanisms differ too much to merge." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "CPU offload device mismatch is not the same as DAC missing STE behavior." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "Weight-tying serialization bug and torch.compile SDPA failure are unrelated." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Model load failure and tokenizer load failure are different bugs." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "Documentation example errors and training with multiple images per prompt are not the same underlying bug." + }, + { + "left": "issue:46305", + "right": "issue:46315", + "accept": false, + "reason": "Off-by-one sliding window masking and slow token conversion are unrelated regressions." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "MIMI padding-mask correctness and OmDet processor image-size mismatch are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "lm_head tying/serialization and dtype mismatch in bfloat16 inference are different code paths." + }, + { + "left": "issue:43257", + "right": "issue:45923", + "accept": false, + "reason": "Qwen MOE weight conversion and a missing gradient-checkpointing flag are unrelated." + }, + { + "left": "issue:43572", + "right": "issue:45759", + "accept": false, + "reason": "Config field missing in StableLM versus composite Qwen text_config unwrapping; not the same bug." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both are offload/quantization themed, but one is dequantized-model loading on gpu+cpu offload and the other is a broader quantization offloading report." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on a specific model vs gradient checkpointing not taking effect are unrelated." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "Batched-vs-single MIMI output mismatch and Perceiver non-default-resolution failure are different model bugs." + }, + { + "left": "issue:43322", + "right": "issue:46050", + "accept": false, + "reason": "Llava Next segfault on load and sinq save/load breakage are not the same code-path problem." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "lm_head tie failure and packed-weight unpacking during accelerate loading are different issues." + }, + { + "left": "issue:45325", + "right": "issue:46315", + "accept": false, + "reason": "Qwen2.5-VL rope/indexing bug and tokenizer conversion slowdown are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "CUDA assert during device-mapped Qwen3 loading and Llava Next segfault are different failures." + }, + { + "left": "issue:44991", + "right": "issue:46291", + "accept": false, + "reason": "Tokenizer loading regression and kernels version/revision import error are unrelated." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP RAM-efficient loading breakage and non-persistent buffer junk initialization are not the same bug." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "3d_parallel.py execution failure and Qwen3Moe fine-grained FP8 failure are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Device-mapping CUDA assert and flash-attention generation behavior are distinct model/runtime problems." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both touch serialization, but one is lm_head.weight inconsistency and the other is visual encoder key saving; not the same fix." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BEiT label reduction bug and Molmo tied-weights attribute error are unrelated." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "Loading a model versus loading a tokenizer; different failure points." + }, + { + "left": "issue:44387", + "right": "issue:45704", + "accept": false, + "reason": "Quantization memory/OOM regression and apex RMSNorm memory leak are separate memory problems with different causes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 13, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:36010", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42617", + "issue:42915", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43317", + "issue:43404", + "issue:43425", + "issue:43450", + "issue:43525", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43782", + "issue:43873", + "issue:44008", + "issue:44186", + "issue: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:45812", + "issue:45823", + "issue:46291", + "issue:46402" + ], + "soft_pairs": [ + "issue:40990|issue:45701", + "issue:36331|issue:42617", + "issue:43065|issue:46402", + "issue:44877|issue:44991", + "issue:44186|issue:45072", + "issue:36331|issue:42915", + "issue:43450|issue:45701", + "issue:43577|issue:43708", + "issue:44008|issue:45823", + "issue:43873|issue:44756", + "issue:43550|issue:43720", + "issue:42175|issue:46291", + "issue:43404|issue:45357", + "issue:41720|issue:42617", + "issue:44423|issue:45406", + "issue:43550|issue:43606", + "issue:44610|issue:44898", + "issue:44186|issue:44610", + "issue:43295|issue:45701", + "issue:42915|issue:43317", + "issue:45632|issue:45684", + "issue:43425|issue:46291", + "issue:36010|issue:42617", + "issue:41720|issue:45161", + "issue:43525|issue:46291", + "issue:43531|issue:43572", + "issue:44671|issue:44743", + "issue:45081|issue:45812", + "issue:43316|issue:44913", + "issue:43540|issue:43782" + ], + "prepared_review_unit_hash": "25216e03998ed2720acd6f50131de9c1a8059263a11a0044b3ebbf19111f5c99", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36780, + "estimated_input_tokens": 9195, + "estimated_eval_tokens": 18646 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly related by broad transformer-model regression themes, but the soft pairs are not true duplicates: they usually differ in model family, failure mode, or code path. No soft edge should be merged. The broadest representative issue is the generic tokenization-regression report.", + "confidence": 0.64, + "canonical_issue_reason": "issue:45701 is the broadest representative: it describes a version-wide tokenization regression and already has multiple inbound references, making it a better cluster anchor than the more model-specific reports.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45701 is the most generally applicable issue in this set; its scope is wider than the other model- or feature-specific bugs, so it best fits a canonical representative for the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:45701", + "accept": false, + "reason": "Perplexity on gpt-oss-20b is a concrete evaluation issue; the other is a broad tokenization regression. Different symptom and scope." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "CustomTrainer.compute_loss kwarg breakage is a trainer API mismatch; 3d_parallel.py is a separate execution failure." + }, + { + "left": "issue:43065", + "right": "issue:46402", + "accept": false, + "reason": "Both are Sam3-related, but one is a dummy Conv2d in the pixel decoder while the other is missing text encoder weights. Different components and fixes." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Different model families and different failure points: strict config loading vs tokenizer loading." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer crashes on NER/padding versus dtype mismatches in inference; unrelated bugs." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Trainer custom loss signature breakage is unrelated to Qwen3Moe FP8 loading failure." + }, + { + "left": "issue:43450", + "right": "issue:45701", + "accept": false, + "reason": "Video processor batch-shape bug is a multimodal preprocessing issue, not the tokenization regression described in 45701." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "One is dtype remaining float32 after load; the other is incorrect max_steps on resume. Different trainer/model-state bugs." + }, + { + "left": "issue:44008", + "right": "issue:45823", + "accept": false, + "reason": "Gemma 3n variable-name collision in forward is unrelated to Gemma4 device-map placement mismatch." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both touch loading/offload behavior, but one is quantization offloading and the other is mmap-induced OOM on Strix Halo. Not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "torch.compile with SDPA on Bamba and packed-weight unpacking during accelerate loading are distinct code paths." + }, + { + "left": "issue:42175", + "right": "issue:46291", + "accept": false, + "reason": "Packaging extras missing TensorFlow backend is different from kernels v0.15 requiring explicit version/revision metadata." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Weight tying in Mistral3 and incorrect visual encoder keys in Qwen3.5 save_pretrained are different regressions." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "cudaErrorAssert with auto device mapping is not the same as failing to run 3d_parallel.py." + }, + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Serve continuous-batching multimodal crash and Gemma4Processor missing _tokenizer are related to serving, but they are different concrete failures." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Both concern image sizing, but one is OmDet processor output size and the other is Perceiver non-default resolution handling. Different models and fixes." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Tokenizer/NER padding crash is unrelated to image processor resolution mismatch." + }, + { + "left": "issue:43295", + "right": "issue:45701", + "accept": false, + "reason": "Processor.tokenizer regression in v4.57.5 is not the same as a generic tokenization-change report, even if both mention tokenization." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "Qwen3Moe FP8 failure and dequantized model offload failure are different loading problems." + }, + { + "left": "issue:45632", + "right": "issue:45684", + "accept": false, + "reason": "Cache-path collision for local model leaf directories is unrelated to save_pretrained permission propagation." + }, + { + "left": "issue:43425", + "right": "issue:46291", + "accept": false, + "reason": "Torch 2.10 incompatibility is a version-compatibility issue, not the kernels dependency-revision error." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error is unrelated to the 3d_parallel.py runtime failure." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Both involve GPT-style model execution, but one is an auto device-map CUDA assert and the other is a tensor-parallel-only GPT-OSS MoE issue." + }, + { + "left": "issue:43525", + "right": "issue:46291", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is a model config bug; kernels version/revision requirements are a dependency-loading issue." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "These are different config-field regressions in different model families: sliding_window vs pad_token_idx." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM prediction errors and recurrent-state reset with cache are distinct behavior bugs." + }, + { + "left": "issue:45081", + "right": "issue:45812", + "accept": false, + "reason": "Mistral regex patch crash and Granite token-ID regression are both tokenizer-related but not the same underlying issue." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy and GPTNeoX rotary_pct not persisting are separate config consistency bugs." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and Qwen3VL weight_only load error are different multimodal loading/processing failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 14, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:29159", + "issue:30064", + "issue:33357", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:36032", + "issue:36331", + "issue:41720", + "issue:42490", + "issue:42915", + "issue:43322", + "issue:43329", + "issue:43388", + "issue:43421", + "issue:43526", + "issue:43531", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43720", + "issue:43746", + "issue:43873", + "issue:43881", + "issue:44186", + "issue:44361", + "issue:44387", + "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: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:30064|issue:34634", + "issue:43388|issue:45059", + "issue:43550|issue:43653", + "issue:43606|issue:43720", + "issue:44530|issue:46291", + "issue:43873|issue:45704", + "issue:45406|issue:45464", + "issue:41720|issue:43746", + "issue:44387|issue:45701", + "issue:45357|issue:46402", + "issue:43606|issue:43653", + "issue:43531|issue:45759", + "issue:44492|issue:44509", + "issue:34634|issue:41720", + "issue:42490|issue:43329", + "issue:34567|issue:35707", + "issue:44451|issue:44991", + "issue:44625|issue:45479", + "issue:43531|issue:44589", + "issue:44514|issue:44545" + ], + "prepared_review_unit_hash": "7a74483ff537125371ed7dc015b453c8f4ac9f5b2a95f8c3d5788d3775fbee41", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35386, + "estimated_input_tokens": 8847, + "estimated_eval_tokens": 17950 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose mix of unrelated bug reports connected only by weak textual similarity. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template crash report; all other soft edges should stay separate.", + "confidence": 0.97, + "canonical_issue_reason": "issue:44514 is the best canonical issue because it is the earlier report of the exact same Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash described in issue:44545.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44514 is the strongest global issue representative since it is the only true duplicate-match in the cluster and has the clearest, most specific reproduction target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizers, different failure modes, and no shared code path beyond generic tokenizer handling." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Different model families and errors; one is FP8 config loading, the other is a Llava Next segfault." + }, + { + "left": "issue:43421", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer-related, but one is runtime post-processor refresh and the other is a Granite token-ID regression; not the same bug." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Completely different subsystems: trainer API signature vs device-mapping CUDA assert." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "Unrelated: Beit image-label reduction bug versus a glm-4v loading failure." + }, + { + "left": "issue:29159", + "right": "issue:36032", + "accept": false, + "reason": "Both mention tokenizers, but the bugs are distinct: slow/fast inconsistency vs T5 method-name conflict." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS CLIP bus error and progressive generation with inputs_embeds/past_key_values are unrelated." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Image processor void-segmentation-map handling is unrelated to BarkProcessor voice_preset." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Different components and symptoms: metric gathering batch trimming vs SAM3 text/bbox behavior." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Different bugs in different areas: torch.compile/SDPA versus tokenizer special-token registration." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Both involve loading/offload-ish paths, but the concrete problems and models are different." + }, + { + "left": "issue:44530", + "right": "issue:46291", + "accept": false, + "reason": "PagedAttentionCache Qwen3.5 crash and kernels dependency import failure are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45704", + "accept": false, + "reason": "Quantization offload behavior and T5 apex RMSNorm memory leak are distinct issues." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Different crashes in different APIs/models: Gemma4Processor missing _tokenizer versus Qwen3.5 streaming inference failure." + }, + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Different model/load paths and different errors; no evidence of the same underlying bug." + }, + { + "left": "issue:44387", + "right": "issue:45701", + "accept": false, + "reason": "CUDA memory regression under int4 quantization is unrelated to a tokenization-version-change report." + }, + { + "left": "issue:45357", + "right": "issue:46402", + "accept": false, + "reason": "Both concern multimodal config/weight handling, but one is incorrect visual encoder keys and the other is missing text encoder weights; not the same bug." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "One is a CPU-offload device mismatch during model loading; the other is a tokenizer special-token decode issue." + }, + { + "left": "issue:43531", + "right": "issue:45759", + "accept": false, + "reason": "Both touch Qwen configs, but they address different problems: sliding_window behavior versus unwrapping text_config in from_config." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "Pure documentation/typo cleanup items, but they do not describe the same change or bug." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping failures are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43329", + "accept": false, + "reason": "Local-path import warning and multimodal token-counting undefined names are unrelated." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "TrainerState token-count tracking and progressive generation with embeddings/past KV are different code paths." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are model-loading regressions, but they affect different models and likely different tokenizer configs." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Both involve classification configuration, but one is Qwen3.5 num_labels propagation and the other is a generic zero-loss problem." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Sliding-window Qwen3-MoE issue is unrelated to a Float8 storage lookup TypeError." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Exact same Qwen2_5_VLProcessor.apply_chat_template batched-padding=False crash; clearly the same underlying bug and mergeable as one fix." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall: the cluster summary stays conservative, and the one accepted pair (issue:44514 / issue:44545) is clearly the same reported crash by title. The other soft-edge rejections are suitably cautious given the packet." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "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:44077|issue:44625", + "issue:45759|issue:45859", + "issue:45588|issue:45759", + "issue:44062|issue:44451", + "issue:43421|issue:43927", + "issue:44987|issue:45042", + "issue:44368|issue:45357", + "issue:39692|issue:44610", + "issue:44918|issue:44936", + "issue:43782|issue:45381", + "issue:43746|issue:46050", + "issue:42886|issue:44843", + "issue:43421|issue:43653", + "issue:45704|issue:45923", + "issue:43381|issue:43844", + "issue:45704|issue:45859", + "issue:45663|issue:45923", + "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": "a3cffb914d346bc00d2b68e6ed2fea068d4501d617417e189029c5de56575fb2", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "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 is mostly a heterogeneous cluster of unrelated bug reports across tokenizers, vision, multimodal, training, and config/loading paths. Only one pair looks like a plausible duplicate around offline tokenizer loading; the rest appear distinct.", + "confidence": 0.78, + "canonical_issue_reason": "No single issue is a clean canonical for the whole cluster; the items span unrelated bugs, with only the offline-tokenizer pair looking duplicate-like.", + "canonical_pr_reason": null, + "best_issue_reason": "No globally representative issue stands out for the full cluster. If anything, 44843 is the closest anchor for the one clear duplicate-like pair, but it does not cover the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Different models and failure modes; image resolution interpolation vs bfloat16 dtype mismatch." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "Unrelated bugs: optional post_init enforcement vs num_labels propagation in Qwen3.5." + }, + { + "left": "issue:45759", + "right": "issue:45859", + "accept": false, + "reason": "Different code paths: composite config unwrapping vs CPU offload incompatibility." + }, + { + "left": "issue:45588", + "right": "issue:45759", + "accept": false, + "reason": "Flash-attention s_aux crash is unrelated to the Qwen config unwrapping issue." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "Both tokenizer-related, but one is AddedToken argument handling and the other is a model-specific load failure." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both touch special tokens, but one is runtime post-processor updates and the other is save/load losing custom IDs." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Different regressions: model loading failure vs image processor requiring torchvision." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "Same model family, but warning noise vs incorrect saved visual encoder keys are different bugs." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Different models and symptoms; docs/example mismatch vs processor/model input-size mismatch." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Trainer SFT embedding unpacking and train/evaluate lifecycle failure are distinct issues." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Different Qwen VL problems: weight_only load failure vs video position-id bug." + }, + { + "left": "issue:43746", + "right": "issue:46050", + "accept": false, + "reason": "PEFT checkpoint loading and quantized save/load corruption are not the same bug." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": true, + "reason": "Both report offline tokenizer loading failing from cache; 44843 pinpoints the likely root cause in _patch_mistral_regex/model_info()." + }, + { + "left": "issue:45704", + "right": "issue:45923", + "accept": false, + "reason": "T5 RMSNorm memory leak and missing gradient-checkpointing flag are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Eval-mode checkpointing restriction is not the same as ZeRO-3 gradient explosion." + }, + { + "left": "issue:45704", + "right": "issue:45859", + "accept": false, + "reason": "T5 memory leak and Qwen rotary offload incompatibility are separate problems." + }, + { + "left": "issue:45663", + "right": "issue:45923", + "accept": false, + "reason": "FSDP2 shared_kv_states KeyError is unrelated to a missing supports_gradient_checkpointing flag." + }, + { + "left": "issue:41720", + "right": "issue:45704", + "accept": false, + "reason": "Qwen3 device-map assert and T5 norm leak have different root causes." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Qwen3-MoE sliding window bug and Jupyter custom-model initialization breakage are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different models and failure types; bad SigLIP2 outputs vs CamemBERT masked-LM regression." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "Runtime special-token updates are unrelated to LayoutLMv2 NER/padding crashes." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "PEFT checkpoint loading and quantization offloading do not appear to share a code-path bug." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve tokenizer patching, but one is a backend_tokenizer AttributeError and the other is a codec-handling regression." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 docs/example mismatch is unrelated to offloading not working with quantization." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Zero3/BertModel init error and audio fbank feature extraction are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:46294", + "accept": false, + "reason": "Both are video-related, but they describe different Qwen models and different failure points." + }, + { + "left": "issue:42915", + "right": "issue:43746", + "accept": false, + "reason": "Qwen3Moe FP8 failure is unrelated to GraniteSpeech local checkpoint loading." + }, + { + "left": "issue:45356", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer regression and list_repo_files revision forwarding are unrelated." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "Image segmentation processing bug and GenerationMixin import error are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative. The only accepted soft edge, 42886\u219444843, is supported by the shared offline tokenizer/cache failure and the explicit _patch_mistral_regex/model_info() root cause in 44843. The rest of the pairwise rejections are appropriately cautious and do not overstate duplication." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "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:43653|issue:43819", + "issue:43531|issue:43643", + "issue:43322|issue:45685", + "issue:43716|issue:44008", + "issue:43643|issue:44877", + "issue:44843|issue:45081", + "issue:43335|issue:43756", + "issue:43388|issue:46032", + "issue:30064|issue:36331", + "issue:43950|issue:45987", + "issue:41720|issue:44991", + "issue:44448|issue:44671", + "issue:44821|issue:45084", + "issue:43421|issue:44811", + "issue:43577|issue:43818", + "issue:39692|issue:45834", + "issue:43540|issue:45381", + "issue:45584|issue:45834" + ], + "prepared_review_unit_hash": "83ac19c87d1772fed70b3238c5119c803fc50da6020c237ec1cb3d771a1ee9ac", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36833, + "estimated_input_tokens": 9209, + "estimated_eval_tokens": 18674 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it spans unrelated tokenizer, vision, training, cache, and model-loading regressions. None of the soft-edge pairs look like true duplicates of the same underlying bug/change, so this should not be deduped as one cluster.", + "confidence": 0.96, + "canonical_issue_reason": "No clear canonical issue: the items describe many different failure modes across unrelated subsystems, so any single issue would be a poor representative.", + "canonical_pr_reason": null, + "best_issue_reason": "No single issue is suitable as a global representative because the cluster is not centered on one concrete bug or code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "Both are numerical/model bugs, but one is BitNet accelerate-loading and the other is DAC latent decoding; different code paths and fixes." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "Tokenizer special-token decoding bug vs BitNet packed-weight loading bug; unrelated subsystems." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio fbank feature extraction issue vs tokenizer mapping failure; no shared underlying defect." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype/SDPA-like behavior, but one is Bamba torch.compile with SDPA and the other is generic inference dtype mismatch; not the same bug." + }, + { + "left": "issue:43295", + "right": "issue:45834", + "accept": false, + "reason": "Processor/tokenizer regression vs long OCR input index error; different failure mechanisms." + }, + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Qwen3OmniMoe video-input processing error vs Gemma3n variable-name collision in forward(); unrelated code paths." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIP attentions field regression vs RTDetrV2 output drift in v5; different models and symptoms." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "AddedToken keyword conflict vs failure to load a specific tokenizer model; not the same defect." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Different model families and bugs: Bamba SDPA/compile issue vs DAC missing-STE mismatch." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "General input validation error messaging vs classification-loss degeneracy; only loosely related by model validation." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "LayoutLMv2 tokenizer crash vs Perceiver image classification resolution bug; distinct components and fixes." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBird mask-token registration issue vs DAC latent/forward mismatch; no shared root cause." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Qwen3-MoE sliding-window behavior vs AutoConfig trust_remote_code missing fields; unrelated areas." + }, + { + "left": "issue:43322", + "right": "issue:45685", + "accept": false, + "reason": "Llava Next loading segfault vs MPS histogram op not implemented for Int; completely different failure modes." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Mistral image-preprocessor dtype mismatch vs Gemma3n audio mask attribute collision; unrelated issues." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-loading regressions, but one is missing fields from trust_remote_code and the other is strict config blocking granite_speech; different causes." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Offline model_info call in tokenizer regex patch vs backend_tokenizer attribute crash in the same patch family, but the described failures are different bug manifestations and fixes." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "SwitchTransformers sparse-layer creation bug vs Smollm3 RoPE-layer mismatch; not the same model logic." + }, + { + "left": "issue:43388", + "right": "issue:46032", + "accept": false, + "reason": "Metric-gathering label truncation vs Mamba2 use_cache correctness; different runtime paths and symptoms." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Image processor segmentation-map handling vs CustomTrainer compute_loss signature mismatch; no overlap in underlying bug." + }, + { + "left": "issue:43950", + "right": "issue:45987", + "accept": false, + "reason": "from_pretrained buffer corruption vs StaticCache get_seq_length return-type bug; different APIs and fix scopes." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "Qwen3 auto device-mapping cuda assert vs tokenizer load failure for a specific model; unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Different-output regression for Pegasus vs incorrect masked LM predictions for CamemBERT; separate model-specific behaviors." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "Loading AutoImageProcessor from URL vs template compilation error; different subsystems." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime special-token/post-processor update feature vs Whisper batch_decode skip_special_tokens bug; not the same change." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "Blip2 dtype propagation issue vs Video-LLaVA temporal-attention/weight-sharing bug; distinct models and defects." + }, + { + "left": "issue:39692", + "right": "issue:45834", + "accept": false, + "reason": "SigLIP2 doc/example errors vs Kosmos2.5 long OCR index error; only broadly vision/OCR-related, not duplicate." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Qwen3OmniMoe video-processing ValueError vs qwen2.5-vl video vision_position_ids bug; same general family but different concrete failure points and fixes." + }, + { + "left": "issue:45584", + "right": "issue:45834", + "accept": false, + "reason": "Whisper empty transcription after token alignment vs Kosmos2.5 long OCR input index error; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:44186|issue:44442", + "issue:43782|issue:44977", + "issue:43644|issue:43749", + "issue:43606|issue:45072", + "issue:43825|issue:43994", + "issue:44479|issue:45834", + "issue:34634|issue:36010", + "issue:44805|issue:44936", + "issue:43577|issue:43582", + "issue:44485|issue:44509", + "issue:28218|issue:29159", + "issue:44561|issue:45020", + "issue:45987|issue:46192", + "issue:43720|issue:45072", + "issue:44610|issue:45072", + "issue:44671|issue:45701", + "issue:43825|issue:43827", + "issue:43782|issue:45325", + "issue:43749|issue:44291", + "issue:43582|issue:44857", + "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": "e4015acdf0211e2711709779979662ed417a05000e73072e40d3a9b92abf667d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36285, + "estimated_input_tokens": 9072, + "estimated_eval_tokens": 18400 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous tokenizer, model-loading, multimodal, and docs bugs. Only one soft edge looks like a plausible duplicate pair; the rest are too broad or clearly different code paths.", + "confidence": 0.71, + "canonical_issue_reason": "issue:28218 is the clearest concrete symptom report in the only plausible duplicate pair; issue:29159 is broader and reads like a more general description of the same tokenizer inconsistency.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:28218 is the best representative issue here because it is open, specific, and user-facing, while the rest of the cluster does not form a coherent duplicate set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "Different bugs: LayoutLMv2 tokenizer NER/padding crash vs AutoTokenizer failing to load a specific tokenizer." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Different model/runtime failures: weight_only loading error vs flash-attention generation issue." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Both involve loading, but one is non-persistent buffers after v5.0 and the other is FSDP CPU RAM efficient loading; not the same code-path." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "Both are CI/model inference mismatches, but they affect different models and different failure modes (CPU offload device mismatch vs bf16 dtype mismatch)." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "One is an error-message/docs issue about removed translation pipelines; the other is a nonsensical output bug for SigLIP2." + }, + { + "left": "issue:44479", + "right": "issue:45834", + "accept": false, + "reason": "Both are multimodal regressions, but they are distinct models and distinct input-processing failures." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "Unrelated: BarkProcessor voice_preset bug vs GenerationMixin import failure." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Different failures: mask shape IndexError vs trainer.evaluate() breaking after train()." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Different issues: BLIP2 dtype not propagating vs Apple Silicon warmup TypeError." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE implementation discussion vs docs mentioning removed pipeline tasks." + }, + { + "left": "issue:28218", + "right": "issue:29159", + "accept": true, + "reason": "Both point at tokenizer behavior inconsistencies around token insertion/spacing; the broader slow-vs-fast inconsistency report plausibly covers the specific extra-space symptom." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "Related to remote-code breakage, but one is is_torch_fx_available removal and the other is broader sys.modules/modeling_utils behavior changes." + }, + { + "left": "issue:45987", + "right": "issue:46192", + "accept": false, + "reason": "Unrelated API issues: StaticCache.get_seq_length return type vs return_tensors being ignored with text_kwargs." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "Different model-loading problems: packed-weight unpacking in BitNet vs bf16 dtype mismatches in unrelated models." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Different symptoms and paths: processor output size mismatch vs inference dtype mismatch." + }, + { + "left": "issue:44671", + "right": "issue:45701", + "accept": false, + "reason": "One is a CamemBERT masked-LM regression; the other is a broad complaint about tokenization changing across versions." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Both mention removed pipeline tasks, but one is an error message and the other is documentation; not the same underlying bug." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Different Qwen multimodal bugs with different failure points." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both concern model loading, but one is FSDP CPU RAM efficient loading and the other is init_empty_weights argument handling." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Different hardware/backend crashes with different stacks and failure causes." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Different AttributeErrors in different components: tokenizer regex patching vs Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Unrelated issues: pip extras dependency omission vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:43906", + "right": "issue:46209", + "accept": false, + "reason": "An isolated reproduction of a prior issue is not the same as an FP8 inference failure on B200." + }, + { + "left": "issue:45372", + "right": "issue:46326", + "accept": false, + "reason": "Both involve mistral_common, but one is a missing import for Gemma 4 loading and the other is a continue_final_message tokenizer bug." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Different audio/DAC code paths: chat-template sampling-rate default vs from_latents forward mismatch." + }, + { + "left": "issue:34634", + "right": "issue:36331", + "accept": false, + "reason": "Unrelated: BarkProcessor bug vs CustomTrainer compute_loss signature mismatch." + }, + { + "left": "issue:45003", + "right": "issue:46285", + "accept": false, + "reason": "Different regressions: unsafe sys.modules access vs CLIP text model structure change." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "Different bugs: BigBirdTokenizer special-token registration vs MXFP4 quantization falling back to bf16." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well-grounded in the packet and stays conservative overall. The accepted pair (issue:28218 / issue:29159) is a plausible tokenizer-behavior duplicate based on the titles, and the other soft-edge rejections are appropriately cautious rather than overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43742|issue:45020", + "issue:43526|issue:43697", + "issue:45084|issue:45538", + "issue:44060|issue:44368", + "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:43550|issue:44912", + "issue:43746|issue:44811", + "issue:43819|issue:43927", + "issue:44387|issue:44617", + "issue:36010|issue:43645", + "issue:43257|issue:43329", + "issue:43720|issue:44610", + "issue:43854|issue:44977", + "issue:38175|issue:44610", + "issue:43531|issue:45042", + "issue:36010|issue:36331", + "issue:46421|issue:46424", + "issue:39692|issue:44912", + "issue:45459|issue:45803", + "issue:42490|issue:43257", + "issue:43262|issue:44811", + "issue:43454|issue:43577", + "issue:43257|issue:43696" + ], + "prepared_review_unit_hash": "33da1165d3715326c22b1d98776812ec2498ed972f2a651f57dc99d3e1d3b2bd", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35718, + "estimated_input_tokens": 8930, + "estimated_eval_tokens": 18116 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose collection of mostly unrelated Transformers bug reports spanning loading, quantization, cache, processors, and model-specific regressions. I do not see a true duplicate/merge group among the soft pairs; most are only superficially similar by subsystem or symptom.", + "confidence": 0.34, + "canonical_issue_reason": "issue:45020 is the broadest, most representative regression report in the set: it describes a general breakage in recent Transformers versions affecting models that rely on `remote_code`.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best single anchor for the cluster because it captures a high-level, user-facing compatibility regression rather than a model-specific one-off.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "Both are loading regressions, but one is a MobileLLM key error and the other is a broad remote_code regression; different failure modes and likely different fixes." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "Unrelated bugs: one is a Beit label-reduction bug, the other is RTDetrV2 output divergence." + }, + { + "left": "issue:45084", + "right": "issue:45538", + "accept": false, + "reason": "Different issues in different layers: compilation error vs tokenizer max_length default." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention Qwen tied-weights warnings, but one is a specific incorrect parameter tying bug and the other is a generic warning during LoRA fine-tuning; not clearly the same fix." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "No overlap beyond being correctness bugs; label reduction and MPS attention correctness are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:45600", + "accept": false, + "reason": "Different models and code paths: video token counting bug vs stale auto_mappings reference." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Different concerns: float8 storage lookup vs unwanted torchvision dependency." + }, + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "One is a placeholder issue title, the other is a reproduction of a specific upstream bug; not the same underlying problem." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "Both affect loading, but one is gguf handling across AutoTokenizer/AutoModel and the other is a specific tokenizer load failure for EMBEDDIA/est-roberta." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "Different trainer failures: ZeRO-3 index error during training setup vs evaluate-after-train failure." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "Unrelated quantization/model bugs: SDPA compile failure in Bamba vs MXFP4 load fallback in git-oss-20b." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "Different processor/model behaviors: local PEFT checkpoint loading vs Whisper batch_decode option handling." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "Different bugs: DAC latent reconstruction mismatch vs DiaConfig losing custom token IDs on save/load." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both can lead to OOM, but the causes are different: int4 reserved-memory growth vs a specific video model consuming too much memory." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "One is an import path change for GenerationMixin, the other is a notebook custom-model initialization regression; not the same code-path." + }, + { + "left": "issue:43257", + "right": "issue:43329", + "accept": false, + "reason": "Same family, different defects: weight conversion under accelerate+deepspeed vs sliding-window handling in Qwen3-MoE." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Different subsystems entirely: BitNet packed-weight loading vs OmDet-Turbo image size mismatch." + }, + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Both are model-specific load/generation problems, but they concern different models and different failure mechanisms." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Unrelated model behavior: zero probabilities in SigLIP2 vs OmDet-Turbo processor/model size mismatch." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Different bugs in different components: Qwen3-MoE sliding-window handling vs PIL backend dependency detection." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "ImportError for GenerationMixin and Trainer compute_loss signature mismatch are separate API regressions." + }, + { + "left": "issue:46421", + "right": "issue:46424", + "accept": false, + "reason": "Both are cache-related, but one is about lazy init causing recompilation and the other about `max_cache_length` only applying to quantized cache; different fixes." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Different model/documentation/quantization issues; one is a SigLIP2 example bug, the other is MXFP4 loading behavior for git-oss-20b." + }, + { + "left": "issue:45459", + "right": "issue:45803", + "accept": false, + "reason": "Both involve exception handling, but in different functions and with different failure semantics; not the same bug." + }, + { + "left": "issue:42490", + "right": "issue:43257", + "accept": false, + "reason": "Local-path import warning with trailing slash is unrelated to Qwen3 MoE weight conversion under accelerate+deepspeed." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Different audio processor issues: sampling-rate defaulting vs batch_decode honoring `skip_special_tokens`." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "Different model-loading bugs in different architectures: AyaVision weight tying vs Blip2 dtype handling." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Completely different issues: Qwen3 MoE weight conversion vs GPT-oss-20b out-of-memory." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 19, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29159", + "issue:33357", + "issue:34567", + "issue:34689", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43317", + "issue:43329", + "issue:43450", + "issue:43452", + "issue: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:29159|issue:34567", + "issue:43749|issue:44756", + "issue:43257|issue:43931", + "issue:44568|issue:45356", + "issue:42915|issue:43742", + "issue:44291|issue:45042", + "issue:42915|issue:44821", + "issue:44987|issue:45538", + "issue:42915|issue:45538", + "issue:43927|issue:44912", + "issue:43452|issue:43582", + "issue:43697|issue:45357", + "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:45137|issue:45405", + "issue:43540|issue:44560", + "issue:43742|issue:44291", + "issue:24643|issue:29159", + "issue:44987|issue:45907" + ], + "prepared_review_unit_hash": "844fbe8fbcf65634585b0bdc91ddf2083fc26a77db94fad55f5a8f9bf37dcc6f", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "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": "This cluster is mostly a grab bag of unrelated Transformers bugs, with one clear subcluster around Qwen video-input regressions in v5.3.0. I only merged the pairs that appear to describe the same underlying video-processing bug family; the rest are false positives across different code paths and models.", + "confidence": 0.79, + "canonical_issue_reason": "issue:44479 is the best representative because it describes the broad Qwen video-input regression affecting multiple model families, and it matches the most convincing soft-edge neighbors in this set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the strongest cluster anchor: it is a concrete, versioned regression with the broadest scope and the clearest overlap to other video-input reports in the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43643", + "right": "issue:45759", + "accept": false, + "reason": "Different loading paths and failures: missing fields from trust_remote_code config vs. unwrapping text_config for composite multimodal CausalLM config." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "Both describe the same Qwen video-input regression class in the multimodal token/vision-video path, and the titles point to the same code area rather than separate bugs." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Both are video-related, but one is an internal token-counting bug and the other is a batched video-processor shape problem; these are different failure modes." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "Unrelated issues: trainer token accounting vs. unusually high perplexity on a specific model/dataset." + }, + { + "left": "issue:43606", + "right": "issue:45859", + "accept": false, + "reason": "Both involve CPU offload, but they concern different models and different concrete errors; not enough evidence of one underlying bug." + }, + { + "left": "issue:29159", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer consistency issue vs. trainer state token counting; different subsystems and symptoms." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Different problems: FSDP CPU RAM efficient loading vs. disabling mmap on Strix Halo to avoid OOM." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen loading issues, but one is accelerate+deepspeed MOE weight conversion and the other is a shape mismatch for a specific VL model." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Distinct tokenizer regressions affecting different models and different mechanisms (special token insertion vs codec handling/warning behavior)." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "Different loading failures: FineGrainedFP8Config vs. a key error for MobileLLM-125M." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "Different areas: init_empty_weights argument handling vs. PIL/image processor dependency requirements." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "No meaningful overlap: Qwen3Moe FP8 failure vs. loading AutoImageProcessor from URL." + }, + { + "left": "issue:44987", + "right": "issue:45538", + "accept": false, + "reason": "Different tokenizer/model-loading bugs with no shared concrete code path." + }, + { + "left": "issue:42915", + "right": "issue:45538", + "accept": false, + "reason": "Different models and failures; one is FP8 config handling, the other is an extreme model_max_length issue." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Custom token IDs being dropped on save/load is unrelated to MXFP4 quantization falling back to bf16." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "GGUF loading breakage and an AppleSilicon caching_allocator_warmup TypeError are unrelated." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "Different regression classes: RTDetr output differences vs. Qwen3.5 save_pretrained key serialization." + }, + { + "left": "issue:44849", + "right": "issue:45701", + "accept": false, + "reason": "Qwen3.5 hidden-state output bug and a broad tokenization-version-change report are not the same underlying issue." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different tokenizers and different behaviors; too broad to call the same bug." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "Bus error on MacOS CLIP model loading vs. trainer token counting; unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Different Qwen3 issues and different failure points: auto device mapping CUDA assert vs. video input processing ValueError." + }, + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Backend packaging issue vs. import failure for a specific model class; unrelated." + }, + { + "left": "issue:43317", + "right": "issue:45538", + "accept": false, + "reason": "Device-map/offload dequantized model loading is unrelated to CLIP tokenizer max length." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both are the same Qwen video-input regression family in v5.3.0, with one report broad across Qwen video models and the other a specific vision_position_ids manifestation." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Generic AddedToken argument conflict and a tokenizer loading failure for a specific model do not look like the same bug." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-related, but they concern different models and different errors; not enough evidence they share one fix." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "PyTorch-not-found import error and a Llama Vision model loading regression are unrelated." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque error vs. a PEFT version requirement bump are unrelated." + }, + { + "left": "issue:44987", + "right": "issue:45907", + "accept": false, + "reason": "Different CI/load failures with no shared code path: model loading regression vs. list_repo_files not forwarding revision." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The two accepted soft edges are plausible members of the same Qwen video-input regression family, while the rejections mostly avoid overclaiming shared root causes across different models and subsystems. The choice of issue:44479 as the canonical anchor is reasonable based on its broad v5.3.0 video regression title and its overlap with the accepted neighbors." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 20, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:42175|issue:44488", + "issue:38175|issue:44898", + "issue:44291|issue:44821", + "issue:43493|issue:43994", + "issue:43257|issue:43756", + "issue:44737|issue:45588", + "issue:34567|issue:36331", + "issue:43582|issue:43782", + "issue:43526|issue:45479", + "issue:44062|issue:45081", + "issue:43404|issue:45987", + "issue:27957|issue:28282", + "issue:44913|issue:46121", + "issue:44792|issue:44964", + "issue:43906|issue:45381", + "issue:45042|issue:45081", + "issue:43756|issue:46399", + "issue:43756|issue:45325", + "issue:43742|issue:44589", + "issue:43452|issue:44488", + "issue:43531|issue:44821", + "issue:45357|issue:45920", + "issue:45405|issue:45600", + "issue:43643|issue:45375" + ], + "prepared_review_unit_hash": "cf314d89686c5bc8a1ea099f7c31b90254424e2543c0980464b1abe6df12b91d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 36190, + "estimated_input_tokens": 9048, + "estimated_eval_tokens": 18352 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a grab-bag of mostly unrelated Transformers bug reports across different models and subsystems. None of the soft pairs look like the same underlying defect, so they should stay split. If one issue must serve as the anchor, issue 43749 is the strongest representative because it has the broadest impact and the most discussion/references.", + "confidence": 0.82, + "canonical_issue_reason": "issue:43749 is the most central and well-connected item in this mixed cluster, with multiple inbound references and substantial discussion.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43749 is the best single issue to anchor triage because it describes a cross-cutting loading failure and has the strongest evidence of broader impact.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "Different problems: PyTorch import/dependency detection vs a LayoutLMv3 error-message clarity issue." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Both mention dtype, but they affect different models and code paths; not the same concrete bug." + }, + { + "left": "issue:42915", + "right": "issue:43749", + "accept": false, + "reason": "FineGrainedFP8/Qwen3MoE loading failure is unrelated to FSDP CPU RAM efficient loading." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Backend package dependency issue vs a model-loading failure for cjvt/sleng-bert." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Separate vision-model bugs: unexpected zero probabilities in SigLIP2 vs Perceiver non-default resolution failure." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights/_is_hf_initialized TypeError is unrelated to AutoImageProcessor URL loading." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": false, + "reason": "Same model family, but one is HF-vs-JAX parity and the other is bad AutoModel/pipeline outputs; not clearly one fix." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Different Qwen/Smollm issues: MOE weight conversion under accelerate+deepspeed vs RoPE-layer dropping." + }, + { + "left": "issue:44737", + "right": "issue:45588", + "accept": false, + "reason": "XLNet CPU placement bug is unrelated to flash_attention.py crashing on sink-less models." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token accounting and CustomTrainer compute_loss signature are different training bugs." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Apple Silicon allocator TypeError is unrelated to Qwen3VL weight_only loading failure." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug is separate from single-label-classification zero-loss behavior." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer AddedToken argument collision and Mistral regex patch backend attribute crash are different failures." + }, + { + "left": "issue:43404", + "right": "issue:45987", + "accept": false, + "reason": "Mistral3 lm_head tying issue is unrelated to StaticCache.get_seq_length returning the wrong type." + }, + { + "left": "issue:27957", + "right": "issue:28282", + "accept": false, + "reason": "Flash Attention 2 support for XLM-Roberta is unrelated to a PyTorch ImportError on AutoModel." + }, + { + "left": "issue:44913", + "right": "issue:46121", + "accept": false, + "reason": "Both involve RoPE/config handling, but one is reload losing rotary_pct and the other is JSON list type handling during validation." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "Janus test failure and Phi-4 multimodal loading failure are different multimodal issues." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "Generic reproduction of another issue is unrelated to Qwen2.5-VL video vision_position_ids miscalculation." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "Image-processor torchvision dependency regression is unrelated to tokenizer regex patching." + }, + { + "left": "issue:43756", + "right": "issue:46399", + "accept": false, + "reason": "Both touch RoPE/Qwen3.5-adjacent areas, but one is missing RoPE layers and the other is a registration/decorator crash." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Smollm3 RoPE-layer dropping and Qwen2.5-VL temporal position scaling are different models and different bugs." + }, + { + "left": "issue:43742", + "right": "issue:44589", + "accept": false, + "reason": "MobileLLM loading key error is unrelated to Float8 storage lookup failure." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "GGUF file handling for tokenizer/CausalLM is unrelated to sleng-bert loading failure." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior is unrelated to loading AutoImageProcessor from URL." + }, + { + "left": "issue:45357", + "right": "issue:45920", + "accept": false, + "reason": "Qwen3.5 visual encoder save_pretrained bug is unrelated to AutoTokenizer producing wrong token IDs across multiple models." + }, + { + "left": "issue:45405", + "right": "issue:45600", + "accept": false, + "reason": "Unreleased PEFT version bump is unrelated to removed Sam3LiteText config references breaking CI." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "AutoConfig trust_remote_code missing fields is a different config-loading problem than Qwen3_5MoeVisionConfig strict-field dropping." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43720|issue:44898", + "issue:44625|issue:44877", + "issue:24643|issue:29127", + "issue:43697|issue:45216", + "issue:41762|issue:44936", + "issue:44493|issue:46050", + "issue:44514|issue:45381", + "issue:44466|issue:45005", + "issue:30064|issue:33666", + "issue:43452|issue:44843", + "issue:28282|issue:32090", + "issue:43819|issue:44811", + "issue:43688|issue:43749", + "issue:43452|issue:43577", + "issue:45372|issue:46170", + "issue:28282|issue:33357", + "issue:44936|issue:45823", + "issue:43906|issue:46202", + "issue:44361|issue:44442", + "issue:43653|issue:44361", + "issue:43295|issue:45356", + "issue:44484|issue:44704", + "issue:38175|issue:43295", + "issue:45542|issue:46178", + "issue:24643|issue:27957", + "issue:43653|issue:44568", + "issue:43577|issue:43697", + "issue:29159|issue:33357", + "issue:44464|issue:44743" + ], + "prepared_review_unit_hash": "06578af5df1362ab1ec15925683c06397b5855fcc12b5cb1c99a58221cf04572", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36663, + "estimated_input_tokens": 9166, + "estimated_eval_tokens": 18588 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The items are mostly heterogeneous, with a loose cluster around v5/tokenizer/save_pretrained regressions, but none of the soft pairs look like the same concrete bug. No PRs are present.", + "confidence": 0.61, + "canonical_issue_reason": "45005 is the broadest representative issue in the list, centered on a v5 tied-weights regression; the others are either more specific manifestations or unrelated failures.", + "canonical_pr_reason": null, + "best_issue_reason": "45005 is the most generally framed and likely to subsume related tied-weights regressions if any duplicate exists.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "Different models and failure modes: BitNet packed-weight loading vs Perceiver interpolation/resolution handling." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Both involve config/loading, but the bugs are different: num_labels propagation vs strict config rejection for granite_speech." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "Completely different problems: DeepSpeed training weight-shape error vs LayoutLMv3 box validation messaging." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "Different regressions: RTDetrV2 output changes vs Qwen3.5 checkpoint save_pretrained correctness." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Different paths: Gemma3 ZeRO-3 loading crash vs trainer evaluate-after-train behavior." + }, + { + "left": "issue:44493", + "right": "issue:46050", + "accept": false, + "reason": "Unexpected key/position-id serialization issue is not the same as sinq quantized save/load breakage." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Both are Qwen2.5-VL related, but one is batched chat template padding and the other is video vision_position_ids." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": false, + "reason": "Related theme of tied weights, but not clearly the same concrete bug; one is device-dependent lm_head serialization, the other is broader translation-model tied-weight issues." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Unrelated: void segmentation map processing vs Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "Different loader failures: gguf_file handling vs unconditional model_info() call in mistral regex patch." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Different errors and layers: missing PyTorch dependency vs Trainer GPU broadcast NoneType failure." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "Different functionality: DAC latent reconstruction vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Distinct bugs: auxiliary loss normalization vs FSDP CPU RAM efficient loading." + }, + { + "left": "issue:43452", + "right": "issue:43577", + "accept": false, + "reason": "Both concern model loading, but one is gguf/AutoTokenizer and the other is BLIP2 dtype propagation." + }, + { + "left": "issue:45372", + "right": "issue:46170", + "accept": false, + "reason": "Different dependency failures: mistral_common ReasoningEffort import vs Windows/ROCm torch.distributed dependency." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "Unrelated: PyTorch missing ImportError vs MacOS bus error with CLIP." + }, + { + "left": "issue:44936", + "right": "issue:45823", + "accept": false, + "reason": "Different runtime contexts and symptoms: trainer lifecycle bug vs Gemma4 PLE device mismatch." + }, + { + "left": "issue:43906", + "right": "issue:46202", + "accept": false, + "reason": "Both mention Gemma/pipeline behavior, but there is no clear evidence they are the same underlying roundtrip bug." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "Different tokenizer failures: MLuke task AttributeError vs FastSpeech2ConformerTokenizer loading." + }, + { + "left": "issue:43653", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizer issues: BigBird special-token registration vs MLuke task AttributeError." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Both are regressions involving processor/tokenizer behavior, but the concrete breakages differ." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "Different topics: save_pretrained shard size question vs missing kwargs passed to cached_file." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Unrelated: zero probabilities in SigLIP2 vs processor.tokenizer regression." + }, + { + "left": "issue:45542", + "right": "issue:46178", + "accept": false, + "reason": "Both concern backend availability, but they are different missing-backend errors with different libraries." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both involve tokenizer special-token behavior, but the specific bugs and models differ." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "Different issues: BLIP2 dtype staying float32 vs RTDetrV2 output changes." + }, + { + "left": "issue:29159", + "right": "issue:33357", + "accept": false, + "reason": "Unrelated: slow/fast tokenizer inconsistency vs MacOS bus error." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Both touch generation/cache behavior, but one is compiled forward chunking and the other is Qwen3.5 recurrent-state reset." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 22, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:34567", + "issue:36010", + "issue:41720", + "issue:43262", + "issue:43295", + "issue:43452", + "issue:43504", + "issue:43531", + "issue:43550", + "issue:43577", + "issue:43606", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43927", + "issue:44361", + "issue:44451", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44514", + "issue:44554", + "issue:44561", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44843", + "issue:44977", + "issue:44991", + "issue:45290", + "issue:45440", + "issue:45491", + "issue:45588", + "issue:45701", + "issue:45704", + "issue:45823", + "issue:46002", + "issue:46050", + "issue:46209", + "issue:46365", + "issue:46402" + ], + "soft_pairs": [ + "issue:44843|issue:44991", + "issue:43531|issue:43742", + "issue:44514|issue:45290", + "issue:43646|issue:43749", + "issue:43688|issue:45440", + "issue:43653|issue:43927", + "issue:43688|issue:45704", + "issue:43756|issue:46402", + "issue:45491|issue:46002", + "issue:43550|issue:43927", + "issue:43688|issue:45701", + "issue:43550|issue:44361", + "issue:44479|issue:44625", + "issue:44464|issue:44977", + "issue:43643|issue:44561", + "issue:41720|issue:44464", + "issue:32090|issue:34567", + "issue:44977|issue:46209", + "issue:43742|issue:44821", + "issue:43504|issue:43720", + "issue:43577|issue:44554", + "issue:33666|issue:36010", + "issue:43577|issue:45588", + "issue:44466|issue:46050", + "issue:43504|issue:43550", + "issue:32090|issue:33666", + "issue:43504|issue:44610", + "issue:43452|issue:44451", + "issue:43720|issue:43927", + "issue:44554|issue:46365", + "issue:43262|issue:43295", + "issue:43606|issue:43927", + "issue:45823|issue:46050" + ], + "prepared_review_unit_hash": "abe7bd906f64f1cebcf2efc5226a862020c0b98e69a1155d4083ef10de364f64", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37598, + "estimated_input_tokens": 9400, + "estimated_eval_tokens": 19056 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: tokenizer/config loading regressions, multimodal processor crashes, training bugs, and model-specific inference issues. None of the soft pairs looks like the same underlying defect with a single mergeable fix, so there is no safe duplicate canonicalization here.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but they fail for different reasons and code paths." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Different models and symptoms: Qwen3-MoE sliding-window behavior vs MobileLLM key error." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "Same `apply_chat_template` area, but distinct crash triggers and inputs." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization breakage and FSDP CPU RAM-efficient loading are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Separate MoE correctness bugs in different architectures and computations." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Tokenizer special-token registration bug vs config field persistence bug." + }, + { + "left": "issue:43688", + "right": "issue:45704", + "accept": false, + "reason": "Aux-loss normalization and RMSNorm memory leak are unrelated failures." + }, + { + "left": "issue:43756", + "right": "issue:46402", + "accept": false, + "reason": "Different models and load-time behaviors; no shared underlying defect." + }, + { + "left": "issue:45491", + "right": "issue:46002", + "accept": false, + "reason": "Both involve NaNs, but in different components and with different triggers." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba compile/SDPA failure is unrelated to DiaConfig token-ID loss." + }, + { + "left": "issue:43688", + "right": "issue:45701", + "accept": false, + "reason": "Related broadly to tokenization/model behavior, but not the same bug." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "Different model/runtime issues with no common code path." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "Video-input regression and num_labels propagation are separate defects." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Both affect generation, but compiled-forward inconsistency and flash-attention failure differ." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both hit `trust_remote_code` users, but they are different v5 breakages." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "Auto device-map assert on A800 is unrelated to chunked-generation compilation issues." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer broadcast TypeError and token-count tracking are separate training bugs." + }, + { + "left": "issue:44977", + "right": "issue:46209", + "accept": false, + "reason": "Flash-attention generation bug vs FP8/B200 inference failure are distinct." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "Model loading key error and AutoImageProcessor URL loading are different problems." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy-field load failure vs BitNet packed-weight handling are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "Blip2 dtype propagation and MPS attention correctness are different issues." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import error are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:45588", + "accept": false, + "reason": "Dtype-loading bug vs flash_attention aux-state crash are different code paths." + }, + { + "left": "issue:44466", + "right": "issue:46050", + "accept": false, + "reason": "Both involve save/load, but tied-weight serialization and sinq quantization are different defects." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Semantic-segmentation preset loading and Bamba compile/SDPA failures are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Trainer GPU broadcast error and Qwen2-VL multi-GPU training request are not the same bug." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Beit legacy-field loading and OmDet-Turbo processor size mismatch are separate issues." + }, + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "GGUF/pretrained loading breakage and ScandiBERT tokenizer/model loading failure are not the same defect." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "BitNet weight unpacking and DiaConfig token-ID persistence are unrelated." + }, + { + "left": "issue:44554", + "right": "issue:46365", + "accept": false, + "reason": "MPS attention correctness and Trainer import shadowing are different problems." + }, + { + "left": "issue:43262", + "right": "issue:43295", + "accept": false, + "reason": "Audio chat-template sampling-rate handling and processor.tokenizer regression are separate bugs." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU-offload device mismatch in bark-small is unrelated to DiaConfig token-ID loss." + }, + { + "left": "issue:45823", + "right": "issue:46050", + "accept": false, + "reason": "Gemma4 device mismatch and sinq save/load breakage are different issues." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 23, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:43994|issue:46192", + "issue:33666|issue:36331", + "issue:43525|issue:43572", + "issue:44849|issue:46209", + "issue:43278|issue:43688", + "issue:43720|issue:44361", + "issue:43493|issue:43825", + "issue:44625|issue:46002", + "issue:28218|issue:33357", + "issue:39692|issue:43493", + "issue:44964|issue:45405", + "issue:44849|issue:45357", + "issue:44568|issue:45812", + "issue:43493|issue:43844", + "issue:39692|issue:43295", + "issue:44488|issue:44843", + "issue:45200|issue:45229", + "issue:43540|issue:44977", + "issue:43278|issue:43844", + "issue:43646|issue:44466", + "issue:44464|issue:46424", + "issue:43334|issue:43572", + "issue:33357|issue:33666", + "issue:44589|issue:45520", + "issue:43257|issue:43299", + "issue:43334|issue:45759", + "issue:43824|issue:44589", + "issue:43927|issue:44361", + "issue:43504|issue:43606", + "issue:43572|issue:44291", + "issue:45229|issue:45823" + ], + "prepared_review_unit_hash": "f0f03347cf8c4bbd0b278d1da4f5ce59c1410e47420af8a8bc3dbb5133d681d2", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37116, + "estimated_input_tokens": 9279, + "estimated_eval_tokens": 18814 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the items span many unrelated model, tokenizer, config, and loading regressions. None of the soft pairs looks like a safe duplicate merge, though a few share broad themes (for example, v5 config regressions or Qwen/SigLIP2 family issues).", + "confidence": 0.77, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45812 is the broadest single regression here: it describes a silent tokenizer corruption affecting all Granite models and has clear user impact, so it is the strongest representative issue if one must be chosen.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43994", + "right": "issue:46192", + "accept": false, + "reason": "Different bugs: SigLIP2 wrong outputs versus a tokenizer kwargs/return_tensors handling issue." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Different areas and failures: Qwen2-VL multi-GPU training versus CustomTrainer.compute_loss signature mismatch." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are v5 padding-token config regressions, but they affect different models and different missing fields; not the same concrete bug." + }, + { + "left": "issue:44849", + "right": "issue:46209", + "accept": false, + "reason": "Unrelated issues: Qwen3.5 hidden-state behavior versus FP8 inference failure on B200." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Different failure modes: dtype drift in evaluation versus auxiliary-loss normalization." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Different subsystems: accelerate loading of packed BitNet weights versus tokenizer/task AttributeError." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "Both mention model/pipeline behavior, but SigLIP2 implementation mismatch and pipeline translation messaging are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:46002", + "accept": false, + "reason": "Different bugs: missing num_labels propagation versus NaNs in classification heads after checkout/init." + }, + { + "left": "issue:28218", + "right": "issue:33357", + "accept": false, + "reason": "Tokenizer spacing bug versus a MacOS bus error when loading a CLIP model; no shared underlying code path is evident." + }, + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both are SigLIP2-related, but one is documentation/example errors while the other is an implementation-vs-JAX discrepancy." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Different topics: Phi-4 multimodal loading versus an unreleased PEFT version bump." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Different Qwen3.5 regressions: hidden states output versus incorrect save_pretrained visual encoder keys." + }, + { + "left": "issue:44568", + "right": "issue:45812", + "accept": false, + "reason": "Tokenizer special-token insertion bug versus Granite token-id regression; same broad area, but not the same concrete issue." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "Different bugs: SigLIP2 discrepancy versus gradient blow-up with HfDeepSpeedConfig + ZeRO-3." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Different issue types: SigLIP2 doc-example problems versus processor/tokenizer regression in v4.57.5." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Both concern loading/tokenizer behavior, but one is a specific model load failure and the other is an offline-mode regression in _patch_mistral_regex." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Different Gemma4 problems: missing mm_token_type_ids default versus multi-GPU CUDA OOM." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Different Qwen variants and different bugs: video-input ValueError versus flash-attention generation issue." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Different regressions: embedding dtype drift versus gradient instability under ZeRO-3." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5-related, but one is custom model initialization and the other is serialization of tied weights; not the same code-path bug." + }, + { + "left": "issue:44464", + "right": "issue:46424", + "accept": false, + "reason": "Different topics: compiled-forward chunked generation versus cache-length handling for quantized cache." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing config attributes after v5-style changes, but they concern different models and different fields." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "Completely unrelated: MacOS bus error on CLIP versus Qwen2-VL training." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different runtime failures: missing Float8 storage type versus flash_attn import_utils KeyError on Python 3.13." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both are Qwen MoE loading issues, but one is accelerate+deepspeed conversion and the other is generic Qwen3VL loading; not a single duplicate bug." + }, + { + "left": "issue:43334", + "right": "issue:45759", + "accept": false, + "reason": "Different config/model-init regressions: missing pad_token_id versus AutoModelForCausalLM not unwrapping text_config." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "Unrelated: import failure for a Qwen2.5 VL class versus a float8 storage lookup error." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Different model components and symptoms: DiaConfig save/load corruption versus MLukeTokenizer AttributeError." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Different bugs: legacy-field loading failure versus CPU-offload device mismatch." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "Both are initialization/config regressions, but one is missing pad_token_idx and the other is an unexpected _is_hf_initialized argument; separate fixes." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Different compatibility problems in different modules; no shared underlying bug is apparent." + }, + { + "left": "issue:45229", + "right": "issue:45823", + "accept": false, + "reason": "Both involve Gemma4 inference, but one is OOM in multi-GPU inference and the other is a PLE device mismatch during forward." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 24, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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: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:43994", + "issue:44112", + "issue:44423", + "issue:44464", + "issue:44493", + "issue:44521", + "issue:44554", + "issue:44734", + "issue:44792", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45405", + "issue:45406", + "issue:45446", + "issue:45520", + "issue:45584", + "issue:45588", + "issue:45823", + "issue:46050", + "issue:46121", + "issue:46285", + "issue:46369" + ], + "soft_pairs": [ + "issue:43278|issue:43701", + "issue:43575|issue:45405", + "issue:44792|issue:45405", + "issue:44493|issue:46285", + "issue:43504|issue:43653", + "issue:44464|issue:45584", + "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:28218|issue:34567", + "issue:43844|issue:44964", + "issue:45406|issue:46121", + "issue:29159|issue:32090", + "issue:45084|issue:45520", + "issue:44554|issue:45446", + "issue:15354|issue:38175", + "issue:43582|issue:44977", + "issue:43540|issue:45200", + "issue:45072|issue:45198", + "issue:45200|issue:45823", + "issue:44521|issue:45290" + ], + "prepared_review_unit_hash": "c50cc905705dafa473a8007346fa398868ac8aa9bee0e2dd0d8077408ecea836", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 26, + "serialized_chars": 35505, + "estimated_input_tokens": 8877, + "estimated_eval_tokens": 18010 + }, + "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 look like true duplicates under a strict same-bug standard. The most representative issue is the multimodal chat-template bug (#44521), but it is not a duplicate of the other candidates.", + "confidence": 0.93, + "canonical_issue_reason": "#44521 is the broadest, most actively discussed issue here and has the clearest ongoing user impact (multimodal `apply_chat_template` behavior). It is the best representative issue for this otherwise mixed cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "#44521 is the strongest cluster head candidate because it is open, has multiple inbound references, and concerns a concrete, user-facing multimodal template failure rather than a niche or one-off regression.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Different problems: dtype drift in evaluation vs a checkpoint-resume key mismatch." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "OOM during tensor-parallel model load is unrelated to a PEFT version bump / release-availability issue." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "A failing janus test is not the same bug as the PEFT minimum-version mismatch." + }, + { + "left": "issue:44493", + "right": "issue:46285", + "accept": false, + "reason": "Unexpected position-id keys across many models is not the same as CLIP text-model structure inconsistency." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Legacy-field loading bug in BEiT segmentation is unrelated to BigBird tokenizer special-token registration." + }, + { + "left": "issue:44464", + "right": "issue:45584", + "accept": false, + "reason": "Compiled-forward chunked generation mismatch is a different code path from empty-transcription failure after `align_special_tokens`." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "CPU offload device mismatch in Bark is not the same as a stale CI test in GraniteSpeech." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "Nonsensical SigLIP outputs are unrelated to unexpected position-id keys." + }, + { + "left": "issue:43873", + "right": "issue:46050", + "accept": false, + "reason": "Quantization/offloading behavior is different from broken save/load of a sinq-quantized model." + }, + { + "left": "issue:45588", + "right": "issue:46050", + "accept": false, + "reason": "Flash-attention sinkless-model crash is not the same as quantized model persistence failure." + }, + { + "left": "issue:43065", + "right": "issue:46369", + "accept": false, + "reason": "A dummy Conv2d implementation in Sam3PixelDecoder is different from a Sam3 LiteText memory overflow." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both involve serve, but one is a multimodal `.to` crash and the other is KV-cache indexing; different bugs." + }, + { + "left": "issue:29942", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to MLFlow reporting after failed training jobs." + }, + { + "left": "issue:42490", + "right": "issue:43696", + "accept": false, + "reason": "Path-with-slash import warning is unrelated to GPU OOM for GPT-oss-20b." + }, + { + "left": "issue:28218", + "right": "issue:34567", + "accept": false, + "reason": "Tokenizer added-space behavior is not the same as trainer input-token accounting." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "DeepSpeed ZeRO-3 gradient growth and Phi-4 multimodal loading failure are different issues." + }, + { + "left": "issue:45406", + "right": "issue:46121", + "accept": false, + "reason": "Gemma4Processor missing `_tokenizer` is unrelated to rope-parameter list type handling." + }, + { + "left": "issue:29159", + "right": "issue:32090", + "accept": false, + "reason": "Slow-vs-fast tokenizer inconsistency is not the same as a trainer broadcast TypeError." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation failure is unrelated to a Python 3.13 flash_attn import KeyError." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "MPS attention correctness when head dims differ is not the same as a PyTorch version check bug in flex attention import." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript export support for GeneratorExp is unrelated to SigLIP zero-probability outputs." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon allocator TypeError is unrelated to Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Qwen3OmniMoe video-processing ValueError is unrelated to Gemma4 token-type defaults for text-only fine-tuning." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "bfloat16 dtype mismatches in inference are unrelated to Wav2Vec2 save/tokenization failures." + }, + { + "left": "issue:45200", + "right": "issue:45823", + "accept": false, + "reason": "Missing default `mm_token_type_ids` for text-only fine-tuning is not the same as a PLE device mismatch during forward." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both involve `apply_chat_template`, but one is zero assistant masks for multimodal inputs and the other is a crash on tool-call messages without content; not the same bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:41720", + "issue:42490", + "issue:42915", + "issue:43232", + "issue:43295", + "issue:43334", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43572", + "issue:43575", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43873", + "issue:43950", + "issue:43957", + "issue:43975", + "issue:44186", + "issue:44466", + "issue:44554", + "issue:44610", + "issue:44792", + "issue:44849", + "issue:44898", + "issue:44987", + "issue:45127", + "issue:45200", + "issue:45325", + "issue:45356", + "issue:45381", + "issue:45412", + "issue:45520", + "issue:45563", + "issue:45823", + "issue:46178", + "issue:46192", + "issue:46209", + "issue:46285" + ], + "soft_pairs": [ + "issue:37428|issue:43957", + "issue:43493|issue:43701", + "issue:44186|issue:46192", + "issue:43232|issue:44849", + "issue:45823|issue:46192", + "issue:43295|issue:43450", + "issue:45381|issue:46209", + "issue:43295|issue:44610", + "issue:43572|issue:44987", + "issue:43295|issue:43873", + "issue:39692|issue:46192", + "issue:44554|issue:45356", + "issue:43572|issue:43957", + "issue:37428|issue:43824", + "issue:45520|issue:46178", + "issue:44466|issue:45127", + "issue:43950|issue:46285", + "issue:44898|issue:45412", + "issue:43550|issue:43844", + "issue:43450|issue:43575", + "issue:45200|issue:45325", + "issue:43334|issue:43525", + "issue:42490|issue:43493", + "issue:42915|issue:43824", + "issue:43824|issue:43975", + "issue:43388|issue:43526", + "issue:44792|issue:45563", + "issue:38175|issue:43493", + "issue:28218|issue:35141", + "issue:41720|issue:46209" + ], + "prepared_review_unit_hash": "814514b7bc1976e7083f8c4dd54813b4935ce5fc6863026158b00d255fe58c3b", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36462, + "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 largely a false-positive similarity group: most items are unrelated regressions, import errors, tokenizer/processor bugs, or model-specific failures. I did not find any soft pair that clearly represents the same underlying issue, so no duplicate merge is warranted.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 43295 is the closest thing to a representative bug report here because it is a concrete, well-scoped regression with decent detail/activity, but the overall set is too heterogeneous for a true canonical duplicate.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43295 is the most representative high-signal issue in the list, but it should be treated only as a loose exemplar rather than an umbrella for the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "Different failures: a missing internal flash-attention symbol import versus a meta-device model loading regression." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation/discrepancy bug and checkpoint resume key mismatch are unrelated code paths." + }, + { + "left": "issue:44186", + "right": "issue:46192", + "accept": false, + "reason": "LayoutLMv2 NER/batched padding crash is not the same as return_tensors being ignored when text_kwargs is passed." + }, + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Generation state update after sync_gpus is unrelated to the Qwen3.5 output_hidden_states regression." + }, + { + "left": "issue:45823", + "right": "issue:46192", + "accept": false, + "reason": "Gemma4 device_map=\"auto\" device mismatch is a different bug from text processor kwargs handling." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor/tokenizer regression with images is not the same as batched video shape handling." + }, + { + "left": "issue:45381", + "right": "issue:46209", + "accept": false, + "reason": "Qwen2.5-VL video position_ids bug and FP8 inference failure on B200 are distinct model/hardware issues." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "These affect different multimodal paths: custom processor/tokenizer regression versus OmDet-Turbo processor input size mismatch." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "Missing pad_token_idx in StableLmConfig is unrelated to loading physical-intelligence/fast." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Processor/tokenizer regression and offloading/quantization behavior are different underlying problems." + }, + { + "left": "issue:39692", + "right": "issue:46192", + "accept": false, + "reason": "Documentation example/model-processor mismatch is not the same as return_tensors precedence being ignored." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness on mismatched head dims is unrelated to Kimi-K2.5 tokenizer codec/regex regression." + }, + { + "left": "issue:43572", + "right": "issue:43957", + "accept": false, + "reason": "Both mention loading-related breakage, but one is a config field regression and the other is a meta-device model-loading bug." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different missing symbols/classes and different models." + }, + { + "left": "issue:45520", + "right": "issue:46178", + "accept": false, + "reason": "flash_attn backend import handling on Python 3.13 is unrelated to the tensorflow_text BACKENDS_MAPPING error." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Tied-weight serialization inconsistency and LoRA merge collapse after vocab extension are different symptoms and fixes." + }, + { + "left": "issue:43950", + "right": "issue:46285", + "accept": false, + "reason": "Non-persistent buffer corruption in from_pretrained is unrelated to the CLIPTextModel structural change." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver interpolation failure and RT-DETR memory release are unrelated model/runtime issues." + }, + { + "left": "issue:43550", + "right": "issue:43844", + "accept": false, + "reason": "torch.compile+SDPA failure and DeepSpeed ZeRO-3 gradient inflation are different training/runtime bugs." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video processor shape bug is unrelated to tensor-parallel OOM during Qwen2-57B-A14B-Instruct loading." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Gemma 4 mm_token_type_ids defaults and Qwen2.5-VL still-image rope scaling are distinct multimodal bookkeeping issues." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "They share a missing pad_token_id symptom, but the affected model/config paths are different enough that this is not clearly one duplicate bug." + }, + { + "left": "issue:42490", + "right": "issue:43493", + "accept": false, + "reason": "Trailing-slash local path import warning and SigLIP2 JAX-vs-HF discrepancy are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "FineGrainedFP8Config failure on Qwen3Moe is not the same as the Qwen2.5-VL import error." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Qwen2.5-VL import failure and DeepSeek tokenizer detokenization regression do not share the same underlying bug." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "gather_for_metrics label dropping and BeitImageProcessorFast reduce_labels truncation are different label-handling bugs." + }, + { + "left": "issue:44792", + "right": "issue:45563", + "accept": false, + "reason": "Janus image-generation test failure is unrelated to the stale num_return_sequences warning." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Both involve SigLIP2, but one is a specific zero-probability symptom and the other is a broader implementation discrepancy; not clearly the same fix." + }, + { + "left": "issue:28218", + "right": "issue:35141", + "accept": false, + "reason": "Tokenizer spacing after added tokens is unrelated to embedding reinitialization after resizing with tie_word_embedding=False." + }, + { + "left": "issue:41720", + "right": "issue:46209", + "accept": false, + "reason": "cudaErrorAssert on A800 with auto device mapping is a different problem from FP8 inference failure on B200." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 26, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:43262", + "issue:43278", + "issue:43335", + "issue:43425", + "issue:43450", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43827", + "issue:43856", + "issue:43901", + "issue:44112", + "issue:44206", + "issue: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:46234", + "issue:46285", + "issue:46326", + "issue:46402" + ], + "soft_pairs": [ + "issue:43278|issue:43856", + "issue:43526|issue:46192", + "issue:43653|issue:44112", + "issue:44442|issue:45198", + "issue:43262|issue:43450", + "issue:38175|issue:45412", + "issue:45290|issue:46326", + "issue:45520|issue:46170", + "issue:44610|issue:44857", + "issue:45759|issue:46285", + "issue:44514|issue:44521", + "issue:44561|issue:46234", + "issue:43827|issue:43901", + "issue:43819|issue:44112", + "issue:44206|issue:45584", + "issue:43819|issue:45083", + "issue:44857|issue:45593", + "issue:43335|issue:43550", + "issue:37428|issue:43425", + "issue:43611|issue:46402", + "issue:43606|issue:45083", + "issue:29942|issue:32090", + "issue:44448|issue:44743", + "issue:45200|issue:45588", + "issue:43335|issue:43653", + "issue:45127|issue:45245", + "issue:37428|issue:44351", + "issue:44855|issue:45083", + "issue:43550|issue:44112" + ], + "prepared_review_unit_hash": "daad55b14c05b72b690c9abb60a170c859b483ef4fabd1f06bbad275023545bf", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36632, + "estimated_input_tokens": 9158, + "estimated_eval_tokens": 18572 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a heterogeneous cluster of distinct bugs and docs issues, not a true duplicate set. All soft pairs look superficially related by subsystem or API name, but they point to different code paths, models, or failure modes.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue fits: the items span unrelated bugs across import utils, chat templates, processors, model loading, training, and CI, so there is no real duplicate cluster to collapse.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue: any one item would be an arbitrary representative rather than a true canonical duplicate target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Different bugs: dtype drift in evaluation vs Qwen3 MoE memory usage." + }, + { + "left": "issue:43526", + "right": "issue:46192", + "accept": false, + "reason": "Different processor behaviors: label reduction vs return_tensors handling with text_kwargs." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "Tokenizer special-token registration vs a GraniteSpeech CI device-override test failure." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "Both touch speech tokenization, but they describe different failures in different model classes." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Different modalities and paths: audio chat-template sampling rate vs batched video output shape." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "Unrelated issues: zero probabilities in SigLIP2 vs RT-DETR memory not released." + }, + { + "left": "issue:45290", + "right": "issue:46326", + "accept": false, + "reason": "Both involve chat templates, but one is a tokenize=True crash with tool calls and the other is continue_final_message on mistral_common tokenizers." + }, + { + "left": "issue:45520", + "right": "issue:46170", + "accept": false, + "reason": "Different import/platform problems: flash_attn key lookup on Python 3.13 vs Windows ROCm/torch.distributed dependency failure." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "Different vision bugs: wrong processor resize output vs float16 AMP crash in a loss function." + }, + { + "left": "issue:45759", + "right": "issue:46285", + "accept": false, + "reason": "Different model-loading regressions: unwrapping text_config vs CLIP text-model structure mismatch." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both are multimodal chat-template bugs, but one is a batched padding=False crash and the other is all-zero assistant masks; not the same defect." + }, + { + "left": "issue:44561", + "right": "issue:46234", + "accept": false, + "reason": "Different missing utilities in import_utils; they are not the same removed symbol or code path." + }, + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are docs issues around pipeline API changes, but they affect different docs and different deprecations." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC forward/from_latents mismatch is unrelated to a GraniteSpeech CI stale device override test." + }, + { + "left": "issue:44206", + "right": "issue:45584", + "accept": false, + "reason": "Different audio bugs: unsupported center arg vs empty-transcription generation after align_special_tokens." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "Different model-specific helpers and failure modes; no shared underlying bug." + }, + { + "left": "issue:44857", + "right": "issue:45593", + "accept": false, + "reason": "Both are training-related vision issues, but one is an AMP/CUDA crash and the other is missing auxiliary losses." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "Unrelated: SwitchTransformers config creates the wrong layer structure; Bamba fails under torch.compile with SDPA." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "One is a missing flash-attention import symbol; the other is a broad Torch 2.10 compatibility issue." + }, + { + "left": "issue:43611", + "right": "issue:46402", + "accept": false, + "reason": "Both are loading regressions, but they concern different mechanisms and models." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "Device-mismatch on CPU offload is unrelated to qwen3_omni_moe feature-length calculations." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and Trainer _gpu_broadcast_one NoneType errors are separate problems." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "A v4/v5 output regression in Pegasus is unrelated to recurrent-state reset logic in qwen3_5." + }, + { + "left": "issue:45200", + "right": "issue:45588", + "accept": false, + "reason": "Different default/None-handling bugs in separate multimodal and flash-attention code paths." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "Sparse-layer creation and BigBird tokenizer special-token registration are unrelated." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA/tied-embeddings collapse and category-count overflow are unrelated failures." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they involve different missing names and different modules." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 torch.jit parsing error is unrelated to qwen3_omni_moe output-length helper behavior." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Different areas entirely: torch.compile/SDPA model bug vs GraniteSpeech CI test flake." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 27, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:43550", + "issue:43606", + "issue:43701", + "issue:43720", + "issue:43746", + "issue:43819", + "issue:43825", + "issue:43854", + "issue:44112", + "issue:44206", + "issue:44360", + "issue:44448", + "issue:44493", + "issue:44561", + "issue:44779", + "issue: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:29127|issue:32090", + "issue:45588|issue:45859", + "issue:45290|issue:45381", + "issue:39692|issue:43526", + "issue:45310|issue:45478", + "issue:43335|issue:45083", + "issue:39692|issue:44493", + "issue:43720|issue:44112", + "issue:44448|issue:44779", + "issue:45459|issue:46234", + "issue:43335|issue:44112", + "issue:43335|issue:45593", + "issue:34689|issue:35141", + "issue:43606|issue:44857", + "issue:42490|issue:43854", + "issue:44360|issue:44561", + "issue:38175|issue:44493", + "issue:44857|issue:45072", + "issue:45920|issue:46192", + "issue:44448|issue:44977", + "issue:28218|issue:32090", + "issue:40444|issue:43526", + "issue:44977|issue:45229", + "issue:43701|issue:45083", + "issue:43746|issue:46209", + "issue:32090|issue:33357", + "issue:29942|issue:35141", + "issue:43550|issue:45083" + ], + "prepared_review_unit_hash": "9c4814659d41d6967dc3b91c56261becada391f2adbd752a64ea2600be81be40", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37329, + "estimated_input_tokens": 9333, + "estimated_eval_tokens": 18922 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly heterogeneous and does not form a coherent duplicate cluster. Only one soft pair looks like an exact duplicate: the two Qwen3.5 MoE from_pretrained regression reports.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45310 is the best canonical issue: it is an exact duplicate of 45478 and the earlier report of the same Qwen3.5 MoE from_pretrained failure.", + "canonical_pr_reason": null, + "best_issue_reason": "45310 is the strongest representative for the only true duplicate pair in this set; the rest of the issues are unrelated and do not support a broader canonical cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "Different bugs: DAC.from_latents/STE mismatch vs position-id key warnings." + }, + { + "left": "issue:44206", + "right": "issue:45823", + "accept": false, + "reason": "Both are device/dtype issues, but they affect different models and code paths." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Both involve Beit, but one is sparse-layer config creation and the other is legacy preset loading." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Training/eval dtype drift is unrelated to the pipeline translation error message." + }, + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "Unrelated problems: LayoutLMv3 error clarity vs Trainer _gpu_broadcast_one NoneType handling." + }, + { + "left": "issue:45588", + "right": "issue:45859", + "accept": false, + "reason": "Both mention offload/flash-attention, but the failing code paths and symptoms differ." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat template crash with tool calls is unrelated to Qwen2.5-VL vision_position_ids." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "Documentation example errors vs BeitImageProcessorFast reduce_labels behavior are different issues." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "Same bug report: identical Qwen3.5 Moe from_pretrained regression 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:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 docs example issues are unrelated to generic position-id key warnings." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "BitNet accelerate-loading failure and GraniteSpeech CI device-override flake are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Different regressions: Pegasus output changes vs Deepseek tokenizer correctness." + }, + { + "left": "issue:45459", + "right": "issue:46234", + "accept": false, + "reason": "Tokenizer error masking via protobuf import is unrelated to a missing import-utils symbol." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "SwitchTransformers config bug is unrelated to GraniteSpeech CI flakiness." + }, + { + "left": "issue:43335", + "right": "issue:45593", + "accept": false, + "reason": "Different model components and symptoms; not the same underlying defect." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Both are loading/initialization issues, but they concern different models and causes." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to LwDetrImageLoss AMP crashes." + }, + { + "left": "issue:42490", + "right": "issue:43854", + "accept": false, + "reason": "Local-path warning and GLM-4.7 model-load test failure are not the same bug." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexer/ReLU discussion is unrelated to removal of is_torch_fx_available." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Unexpected zero probabilities in SigLIP2 is not the same as position-id key warnings." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different failure modes: AMP crash vs bfloat16 dtype mismatch in inference." + }, + { + "left": "issue:45920", + "right": "issue:46192", + "accept": false, + "reason": "Wrong token IDs regression is unrelated to return_tensors being ignored with text_kwargs." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus v4/v5 output drift and Qwen3.5 flash-attention generation failure are different bugs." + }, + { + "left": "issue:28218", + "right": "issue:32090", + "accept": false, + "reason": "Tokenizer spacing after added token is unrelated to Trainer GPU broadcast NoneType errors." + }, + { + "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 issue and multi-GPU CUDA OOM are different failure classes." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is unrelated to qwen3_omni_moe feature-length helper behavior." + }, + { + "left": "issue:43746", + "right": "issue:46209", + "accept": false, + "reason": "Local checkpoint loading with PEFT adapters is unrelated to FP8 inference on B200." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer broadcast NoneType and MacOS bus error on CLIP are unrelated." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to output embedding reinitialization in post_init." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba is unrelated to qwen3_omni_moe helper behavior." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall: the exact duplicate pair 45310/45478 is clearly supported by the packet (same title and same Qwen3.5 MoE from_pretrained regression), and the summary does not over-claim a broader cluster. The false soft-edge verdicts are phrased cautiously as different bugs/code paths, which fits 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": 21, + "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:33357", + "issue:34689", + "issue:37428", + "issue:39692", + "issue:41720", + "issue:42490", + "issue:43278", + "issue:43329", + "issue:43381", + "issue:43412", + "issue:43575", + "issue:43653", + "issue:43873", + "issue:43927", + "issue:43976", + "issue:44186", + "issue:44206", + "issue:44361", + "issue:44448", + "issue:44466", + "issue:44479", + "issue:44561", + "issue:44568", + "issue:44743", + "issue:44779", + "issue:44855", + "issue:44857", + "issue:44977", + "issue:44987", + "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:42490|issue:43575", + "issue:10000|issue:15354", + "issue:37428|issue:44987", + "issue:43927|issue:45698", + "issue:44561|issue:44779", + "issue:43278|issue:43329", + "issue:27957|issue:34689", + "issue:44857|issue:45588", + "issue:43873|issue:45229", + "issue:41720|issue:45229", + "issue:44568|issue:45491", + "issue:44977|issue:45563", + "issue:43278|issue:43381", + "issue:45372|issue:45588", + "issue:45588|issue:45823", + "issue:43412|issue:45588", + "issue:44448|issue:44466", + "issue:45372|issue:45406", + "issue:43976|issue:45440", + "issue:45198|issue:45803", + "issue:25251|issue:32090", + "issue:43653|issue:44743", + "issue:45245|issue:45563", + "issue:32090|issue:33290", + "issue:44186|issue:45198", + "issue:44206|issue:44479", + "issue:44361|issue:45198", + "issue:42490|issue:43873", + "issue:33357|issue:39692" + ], + "prepared_review_unit_hash": "e3a1bce9c6108e41be27b20946ee20e3b2c51ec913af3d5cde7cde568656f9eb", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36962, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a loose similarity cluster, not a true duplicate cluster: the issues span unrelated models, loaders, tokenizers, attention paths, and memory/runtime failures. No soft edge is strong enough to merge as the same underlying bug or change.", + "confidence": 0.9, + "canonical_issue_reason": "issue:43873 is the broadest and most representative standalone issue in the set, with the highest apparent discussion/inbound activity and a general loading/offloading theme that is closer to a central cluster anchor than the many highly specific bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the best issue representative because it is broad, active, and ties to a common infrastructure problem class (quantization/offloading/memory behavior), even though the rest of the items are not true duplicates of it.", + "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 torch.jit parsing issue and the other is a missing decorator/registration bug in Qwen3.5; different code paths and fixes." + }, + { + "left": "issue:42490", + "right": "issue:43575", + "accept": false, + "reason": "Local-path import warnings and tensor-parallel OOM on Qwen2-57B are unrelated failure modes." + }, + { + "left": "issue:10000", + "right": "issue:15354", + "accept": false, + "reason": "Generic DistilBertModel failure and torch.jit GeneratorExp export limitation on ViT are distinct issues." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "A missing flash-attention utility import and a model-loading failure for a specific repo are not the same underlying bug." + }, + { + "left": "issue:43927", + "right": "issue:45698", + "accept": false, + "reason": "Both concern save/load behavior, but one loses custom token IDs and the other restores the wrong custom module; separate defects." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "API removal breaking trust_remote_code models is unrelated to DeepSeek tokenizer output regressions." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift in eval and an undefined-variable multimodal token bug are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:34689", + "accept": false, + "reason": "Both involve model loading and attention features, but they target different models and different breakages." + }, + { + "left": "issue:44857", + "right": "issue:45588", + "accept": false, + "reason": "AMP/CUDA loss crash and a flash-attention s_aux=None AttributeError are different runtime paths." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Both mention memory pressure, but one is quantization/offloading behavior and the other is a specific multi-GPU inference OOM on Gemma4." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "Auto device-map cudaErrorAssert and CUDA OOM are different symptoms with different likely causes." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token handling and Gemma3 NaN embeddings are unrelated bugs." + }, + { + "left": "issue:44977", + "right": "issue:45563", + "accept": false, + "reason": "Flash-attention generation regression and a stale warning in paged generate() are not the same code-path problem." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Eval/training dtype mismatch and gradient-checkpointing-in-eval are different behaviors." + }, + { + "left": "issue:45372", + "right": "issue:45588", + "accept": false, + "reason": "Gemma4 processor import failure and flash_attention integration crash are separate failures." + }, + { + "left": "issue:45588", + "right": "issue:45823", + "accept": false, + "reason": "An AttributeError in flash_attention integration and a Gemma4 device mismatch during forward are not the same bug." + }, + { + "left": "issue:43412", + "right": "issue:45588", + "accept": false, + "reason": "Executorch export problems and flash-attention integration errors are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Different output behavior and inconsistent serialization are related to v5 behavior but are not the same concrete issue." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Both are Gemma4-related, but one is a missing dependency import and the other is a missing _tokenizer attribute in serve." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version incompatibility and DeepSeekV3MoE divergence are unrelated." + }, + { + "left": "issue:45198", + "right": "issue:45803", + "accept": false, + "reason": "Wav2Vec2 save/tokenization failure and a bare-except bug in Fuyu tensor conversion are different issues." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k output-shape change and Trainer _gpu_broadcast_one TypeError are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and Qwen3.5 recurrent-state reset are different model subsystems." + }, + { + "left": "issue:45245", + "right": "issue:45563", + "accept": false, + "reason": "Category cardinality limit and a stale warning in paged generate() have no common underlying defect." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "Trainer broadcast TypeError and deepspeed Adafactor OOM are different runtime failures." + }, + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different tokenizers, different crash modes, and different affected tasks; not a duplicate." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported center argument in feature extraction and Qwen VL video regression are separate bugs." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLukeTokenizer AttributeError and Wav2Vec2 save/tokenization failure are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43873", + "accept": false, + "reason": "Import warning for local paths and quantization/offloading behavior are not the same issue." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error on a community CLIP model and SigLIP2 docs errors are different problems, one runtime and one documentation/example-related." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 29, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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: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:43761", + "issue:43824", + "issue:43931", + "issue:43994", + "issue:44186", + "issue:44265", + "issue:44279", + "issue:44387", + "issue:44442", + "issue:44462", + "issue:44625", + "issue:44871", + "issue:44977", + "issue:45072", + "issue:45198", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45584", + "issue:45593", + "issue:45712", + "issue:45834", + "issue:45859", + "issue:46121", + "issue:46369" + ], + "soft_pairs": [ + "issue:29127|issue:34567", + "issue:43065|issue:45412", + "issue:43611|issue:46121", + "issue:42915|issue:42994", + "issue:44871|issue:44977", + "issue:43065|issue:43994", + "issue:43066|issue:44462", + "issue:45459|issue:45712", + "issue:43278|issue:43575", + "issue:29942|issue:30064", + "issue:45412|issue:46369", + "issue:45372|issue:45459", + "issue:43994|issue:45412", + "issue:44265|issue:45198", + "issue:43381|issue:43701", + "issue:43653|issue:44442", + "issue:43756|issue:45440", + "issue:43421|issue:45491", + "issue:27957|issue:29942", + "issue:29127|issue:33357", + "issue:43824|issue:44279", + "issue:45440|issue:45593", + "issue:43994|issue:46369", + "issue:43526|issue:45834", + "issue:43526|issue:44186", + "issue:29127|issue:35141", + "issue:44625|issue:45584", + "issue:43575|issue:43701", + "issue:44442|issue:45072", + "issue:42490|issue:43381", + "issue:43575|issue:44387", + "issue:43761|issue:45859", + "issue:42915|issue:43931" + ], + "prepared_review_unit_hash": "2ab49ca1bf75fcc75ac3ad1776cd013886862684377f84b3308b0797bc39ac3e", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37522, + "estimated_input_tokens": 9381, + "estimated_eval_tokens": 19018 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This looks like a heterogeneous Transformers issue bundle rather than a true duplicate cluster. The soft-similarity pairs share keywords or subsystems, but they do not appear to describe the same underlying bug or change.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue really fits this set because the reports span unrelated bugs; the closest thing to a representative item is a concrete, user-facing tokenizer-loading bug.", + "canonical_pr_reason": null, + "best_issue_reason": "Best overall issue representative is the most self-contained and actionable bug report; the rest are mostly model- or subsystem-specific and not duplicate-alike.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "Different areas: LayoutLMv3 error messaging vs TrainerState token counting." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "One is a dummy Conv2d in Sam3PixelDecoder; the other is RT-DETR memory not being released." + }, + { + "left": "issue:43611", + "right": "issue:46121", + "accept": false, + "reason": "Loading via base_model_prefix and rope-parameter JSON handling are unrelated bugs." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "Both involve quantization, but one is a Qwen3Moe FP8 failure and the other is quantized model saving." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma eos token config mismatch vs Qwen3.5 flash-attention generation failure." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Sam3PixelDecoder dummy layer issue is unrelated to SigLIP2 bad AutoModel/pipeline outputs." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Both are tokenizer-loading issues, but the concrete failure modes differ and are not the same bug." + }, + { + "left": "issue:45459", + "right": "issue:45712", + "accept": false, + "reason": "Tokenizer error masking with protobuf missing is unrelated to leftover dummy class cleanup." + }, + { + "left": "issue:43278", + "right": "issue:43575", + "accept": false, + "reason": "Embedding dtype drift in eval mode is unrelated to Qwen2 loading OOM." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void segmentation map processing are different issues." + }, + { + "left": "issue:45412", + "right": "issue:46369", + "accept": false, + "reason": "Both are memory-related, but RT-DETR GC leaks and Sam3-litetext overflow are different code paths." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma 4 processor import error from mistral_common is unrelated to tokenizer protobuf error masking." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 bad inference and RT-DETR memory release are unrelated bugs." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failures from torch_compilable_check are not the same as Wav2Vec2 save/tokenization failures." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient-checkpointing eval-mode behavior and resume_from_checkpoint key mismatch are separate training issues." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and FastSpeech2ConformerTokenizer loading are different tokenizer bugs." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE-layer dropout and DeepseekV3MoE divergence from remote implementation are not the same defect." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token post-processing and Gemma3 NaN embeddings are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:29942", + "accept": false, + "reason": "Both mention Flash Attention 2, but one is XLMRoberta support and the other is failing tests." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error clarity and MacOS CLIP bus error are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Missing Qwen2_5_VL import and a generic dependency issue are not the same bug." + }, + { + "left": "issue:45440", + "right": "issue:45593", + "accept": false, + "reason": "DeepseekV3MoE implementation drift is unrelated to D-FINE auxiliary-loss behavior." + }, + { + "left": "issue:43994", + "right": "issue:46369", + "accept": false, + "reason": "SigLIP2 wrong outputs and Sam3-litetext memory overflow are different symptoms and code paths." + }, + { + "left": "issue:43526", + "right": "issue:45834", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and Kosmos2.5 OCR index error are unrelated image-processing bugs." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to LayoutLMv2Tokenizer NER padding/truncation crashes." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "Error-message clarity in LayoutLMv3 and embedding reinit after resize are unrelated." + }, + { + "left": "issue:44625", + "right": "issue:45584", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Whisper empty-transcription generation failure are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Qwen2 loading OOM and checkpoint resume key mismatch are different problems." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading failure and dtype mismatch in inference are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:43381", + "accept": false, + "reason": "Trailing-slash local model import warning and gradient-checkpointing eval-mode behavior are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "Both are OOM-related, but one is model/tp loading and the other is int4 quantization memory regression." + }, + { + "left": "issue:43761", + "right": "issue:45859", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and CPU-offload incompatibility in Qwen3_5MoeTextRotaryEmbedding are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Qwen3Moe FP8 failure and Qwen3-VL weight-shape mismatch are different model-loading 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": 21, + "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:43232", + "issue:43262", + "issue:43316", + "issue:43334", + "issue:43381", + "issue:43388", + "issue:43452", + "issue:43525", + "issue:43526", + "issue:43531", + "issue:43611", + "issue:43646", + "issue:43720", + "issue:43723", + "issue:43761", + "issue:43873", + "issue:43931", + "issue:44062", + "issue:44077", + "issue:44079", + "issue:44206", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44977", + "issue:45003", + "issue:45216", + "issue:45245", + "issue:45440", + "issue:45479", + "issue:45491", + "issue:45588", + "issue:45907", + "issue:46121", + "issue:46234", + "issue:46399" + ], + "soft_pairs": [ + "issue:43526|issue:44857", + "issue:43334|issue:43531", + "issue:46121|issue:46399", + "issue:44977|issue:45245", + "issue:43262|issue:44206", + "issue:44464|issue:45491", + "issue:27957|issue:29127", + "issue:44743|issue:44811", + "issue:43452|issue:45588", + "issue:27957|issue:33290", + "issue:43646|issue:44077", + "issue:33290|issue:34689", + "issue:44062|issue:45907", + "issue:33290|issue:35141", + "issue:33290|issue:34567", + "issue:43381|issue:43873", + "issue:44442|issue:44898", + "issue:45003|issue:45440", + "issue:43723|issue:45479", + "issue:42490|issue:44743", + "issue:43316|issue:43525", + "issue:43761|issue:44079", + "issue:42175|issue:46234", + "issue:34689|issue:43611", + "issue:43720|issue:44442", + "issue:43232|issue:43388", + "issue:43931|issue:45216", + "issue:44077|issue:44479" + ], + "prepared_review_unit_hash": "ba1c197cf768e52f08100cb975353f8dd727d59be017668eee78412226f7e39b", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "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": "The set is highly heterogeneous: most items are distinct bugs/regressions in different subsystems or model families, and none of the soft pairs look like true duplicates of the same concrete underlying issue.", + "confidence": 0.96, + "canonical_issue_reason": "No single canonical issue: the cluster spans unrelated loading, generation, preprocessing, optimizer, and config bugs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the broadest representative regression in the set, affecting custom model initialization across models.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "Different components and failures: BEiT image processor label reduction vs LwDETR loss crashing under float16/CUDA." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Different Qwen variants and bugs: missing pad_token_id on a text config vs sliding-window recurrence reset." + }, + { + "left": "issue:46121", + "right": "issue:46399", + "accept": false, + "reason": "Unrelated code paths: rope-validation TypeError vs rotary-pos-embedding registration decorator failure." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Different problems entirely: flash-attention generation regression vs category-count limit error." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sampling-rate default bug vs Laser/feature-extractor unsupported-center crash." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Different generation/attention failures: compiled-forward chunked generation inconsistency vs Gemma3 NaN embeddings with sliding-window all-padding cases." + }, + { + "left": "issue:27957", + "right": "issue:29127", + "accept": false, + "reason": "Different subsystems: XLM-RoBERTa Flash Attention 2 support vs LayoutLMv3 error-message clarity." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Different model paths and bugs: Qwen3_5 recurrent state reset vs Whisper batch_decode skipping special tokens." + }, + { + "left": "issue:43452", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer/model loading with gguf_file vs flash-attention sink-less-model AttributeError are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:33290", + "accept": false, + "reason": "One is Flash Attention model support, the other is Adafactor OOM in DeepSpeed; no shared underlying bug." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Both touch initialization, but one is a broad v5 custom-model-init regression and the other is a specific patchtsmixer post_init API rule change." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "Optimizer OOM in DeepSpeed vs a model-loading regression for Llama 3.2 Vision; unrelated failures." + }, + { + "left": "issue:44062", + "right": "issue:45907", + "accept": false, + "reason": "AddedToken keyword duplication vs list_repo_files revision forwarding are unrelated utility bugs." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "DeepSpeed optimizer OOM and embedding reinitialization after resize are distinct issues." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "Training memory usage vs TrainerState token-count tracking are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing eval-mode restriction vs quantization offloading behavior are different code paths." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "Tokenizer loading failure for FastSpeech2ConformerTokenizer vs Perceiver non-default-resolution failure are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils vs DeepseekV3MoE divergence are different bugs." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "AutoTokenizer loading regression in v5 vs zero-loss classification with num_labels=1 are not the same issue." + }, + { + "left": "issue:42490", + "right": "issue:44743", + "accept": false, + "reason": "Local-path trailing-slash import warning vs Qwen3_5 recurrent-state reset; different areas and symptoms." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both concern config APIs, but the described discrepancies are not clearly the same missing attribute or failure mode." + }, + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression vs ModelOutput key assignment semantics are different underlying problems." + }, + { + "left": "issue:42175", + "right": "issue:46234", + "accept": false, + "reason": "Packaging/backend dependency issue vs missing import_utils symbol; no common bug." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Both are version regressions, but they affect different loading paths and failure modes." + }, + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "BitNet packed-weight loading bug vs tokenizer loading failure; unrelated." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs update after sync_gpus vs metrics label dropping are unrelated." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Weight-shape mismatch on model load vs incorrect save_pretrained checkpoint for Qwen3.5 are different bugs." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "PatchTSMixer post_init policy vs Qwen video-input regression; different models and code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43716", + "issue:43742", + "issue:43749", + "issue:43761", + "issue:43825", + "issue:43827", + "issue:43873", + "issue:43931", + "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:45290", + "issue:45632", + "issue:45698", + "issue:45759", + "issue:45803", + "issue:46178", + "issue:46202", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:43761|issue:44188", + "issue:44987|issue:46178", + "issue:43493|issue:43827", + "issue:43873|issue:44871", + "issue:44442|issue:45803", + "issue:42907|issue:45698", + "issue:44464|issue:46421", + "issue:43749|issue:43931", + "issue:43643|issue:44913", + "issue:43316|issue:45759", + "issue:44479|issue:45290", + "issue:44855|issue:45137", + "issue:44190|issue:44496", + "issue:43650|issue:46202", + "issue:43295|issue:43825", + "issue:25251|issue:28282", + "issue:43550|issue:43701", + "issue:43606|issue:43701", + "issue:45632|issue:46424", + "issue:42617|issue:43575", + "issue:43575|issue:43742", + "issue:27957|issue:32090", + "issue:16998|issue:22355", + "issue:43316|issue:43716", + "issue:43065|issue:44617", + "issue:43295|issue:44871" + ], + "prepared_review_unit_hash": "a0f67fb85e284ad8f7435ec60c139355f2478f6dc334e9428c274c2c565a457a", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 26, + "serialized_chars": 34686, + "estimated_input_tokens": 8672, + "estimated_eval_tokens": 17600 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is heterogeneous and the soft-similarity pairs are mostly false positives: they share broad themes like loading, caching, offload, or v5 regressions, but not the same concrete bug or fix path.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "Both involve torch.compile/regression behavior, but one is CLIPVisionModel hidden_states handling and the other is attention-kernel branching; different code paths and symptoms." + }, + { + "left": "issue:44987", + "right": "issue:46178", + "accept": false, + "reason": "Both are loading-time failures, but one is a specific model/package regression and the other is a backend mapping error for tensorflow_text; not the same bug." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation discrepancy vs stale docs referencing pipeline() after v5 removals; one is model correctness, the other is documentation drift." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior vs Gemma eos_token_id config inconsistency; unrelated issues." + }, + { + "left": "issue:44442", + "right": "issue:45803", + "accept": false, + "reason": "Tokenizer loading failure for FastSpeech2ConformerTokenizer vs a bare-except bug in Fuyu tensor conversion; different components and failure modes." + }, + { + "left": "issue:42907", + "right": "issue:45698", + "accept": false, + "reason": "Both touch save/load flows for custom models, but one is dequantized model saving and the other is wrong custom-module selection after save_pretrained; not the same underlying defect." + }, + { + "left": "issue:44464", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked/compiled execution, but one is inconsistent generation outputs and the other is cache lazy-init causing recompilation; different root causes." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "FSDP CPU RAM efficient loading broken vs Qwen3-VL weight-shape mismatch; unrelated loading problems." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Both are config persistence issues, but one loses fields with trust_remote_code and the other resets rotary_pct on reload; different model/config paths." + }, + { + "left": "issue:43316", + "right": "issue:45759", + "accept": false, + "reason": "Both concern multimodal config APIs, but one is Gemma3TextConfig inconsistency and the other is AutoModelForCausalLM.from_config not unwrapping text_config for Qwen; not mergeable as one fix." + }, + { + "left": "issue:44479", + "right": "issue:45290", + "accept": false, + "reason": "Video-input regression in Qwen family vs chat-template crash on tool-call assistant messages; unrelated features and code paths." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 import/parsing failure in DebertaV2Model vs DeepSpeed ZeRO3 deque underflow; completely different failures." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Local dataset loading in an image-classification script vs unrecognized-model config loading; different subsystems." + }, + { + "left": "issue:43650", + "right": "issue:46202", + "accept": false, + "reason": "A vague request and a Gemma pipeline roundtrip tokenization bug are not the same underlying issue." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "Both are v5 regressions around pipeline/processor behavior, but one is processor.tokenizer/images handling and the other is an error message about translation support; different fixes." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Nested-list output change in top_k pipelines vs missing PyTorch import error; unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "torch.compile/SDPA failure in Bamba vs resume_from_checkpoint key mismatch; no shared code path." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch vs checkpoint key mismatch; different bugs." + }, + { + "left": "issue:45632", + "right": "issue:46424", + "accept": false, + "reason": "Both mention cache behavior, but one is trust_remote_code cache-path collisions and the other is max_cache_length handling for quantized cache; not the same defect." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel.py execution problem vs tensor-parallel OOM loading Qwen2-57B; different failure modes." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "OOM during tensor-parallel load vs key error loading MobileLLM-125M; not the same bug." + }, + { + "left": "issue:27957", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 support in XLMRoberta vs NoneType broadcast error in Trainer; unrelated." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "model_max_length question for DeBERTa-V3 vs missing transformers.onnx module; unrelated issues." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API discrepancy vs Mistral-3 dtype mismatch between preprocessor and model; different problems." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder vs CUDA OOM in Sam3Video; same family, but not the same underlying bug." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "v5 processor/tokenizer regression vs Gemma eos_token_id mismatch; unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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: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: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:45381", + "issue:45406", + "issue:45542", + "issue:45563", + "issue:45584", + "issue:45698", + "issue:45803", + "issue:46050", + "issue:46170", + "issue:46178", + "issue:46234", + "issue:46365", + "issue:46369", + "issue:46421" + ], + "soft_pairs": [ + "issue:43540|issue:44514", + "issue:43316|issue:44877", + "issue:44610|issue:46369", + "issue:44162|issue:44792", + "issue:44871|issue:45584", + "issue:43404|issue:43441", + "issue:44077|issue:44534", + "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:45127|issue:45200", + "issue:43388|issue:43688", + "issue:43572|issue:46178", + "issue:43673|issue:44743", + "issue:44617|issue:45584", + "issue:43504|issue:45698", + "issue:44734|issue:45406", + "issue:45542|issue:46234", + "issue:43673|issue:46421", + "issue:43525|issue:43531", + "issue:44792|issue:44977", + "issue:45381|issue:46369", + "issue:43761|issue:44242", + "issue:45563|issue:45584" + ], + "prepared_review_unit_hash": "e53da95d9e968250749f71fb9ab1cd5c575da96f005bd602ead2865a792aaf9e", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36169, + "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 false-positive mix of unrelated Transformers bugs: model-specific regressions, config/serialization issues, generation problems, backend-import failures, and a few video/speech cases. A couple of pairs share broad subsystem vocabulary (cache/chunked prefill, save/load, config fields), but they still describe different concrete failures.", + "confidence": 0.96, + "canonical_issue_reason": "issue:45698 is the broadest, most representative open report: it covers broken custom-module resolution after save_pretrained/from_pretrained, has the strongest discussion activity, and overlaps loosely with the other load/save-related item in the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45698 is the best overall issue candidate because it is open, broadly applicable to custom model loading, and has the most active discussion/inbound references among the issues here.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Both involve multimodal/video processing, but they are different model families and different failures: Qwen3OmniMoe video-input handling vs Qwen2.5-VL batched chat-template padding." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Both are config/API complaints, but the concrete problems differ: Gemma3TextConfig API inconsistency vs granite_speech being blocked by strict config loading." + }, + { + "left": "issue:44610", + "right": "issue:46369", + "accept": false, + "reason": "Different models and different bug classes: OmDet processor image-size mismatch vs sam3-litetext memory overflow." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "No shared underlying bug; ESM2 is a model breakage report, while janus test failure is a separate image-generation regression." + }, + { + "left": "issue:44871", + "right": "issue:45584", + "accept": false, + "reason": "Both mention generation/token IDs, but Gemma-3 eos_token_id config inconsistency and Whisper empty-transcription failure are distinct code paths." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "Same family name is not enough: one is a weight-tying bug in Mistral3ForConditionalGeneration, the other is a FlashAttention failure in Ministral-3." + }, + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Unrelated issues: patchtsmixer post_init policy vs a v5 non-persistent-buffer initialization bug." + }, + { + "left": "issue:45803", + "right": "issue:46365", + "accept": false, + "reason": "Different failure modes and surfaces: Fuyu tensor conversion exception handling vs Trainer import failure tied to datasets." + }, + { + "left": "issue:45803", + "right": "issue:46170", + "accept": false, + "reason": "Bare-except tensor conversion bug is unrelated to the Windows/ROCm torch.distributed import problem." + }, + { + "left": "issue:44871", + "right": "issue:46050", + "accept": false, + "reason": "Both concern model state/config after changes, but Gemma eos_token_id mismatch and sinq quantized save/load corruption are different bugs." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Different multimodal pipelines and symptoms: video-classification processor lookup vs SAM3 vision encoder missing attribute." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Both are loading failures, but one is a gguf/from_pretrained path issue and the other is an unrecognized model_type/config.json problem." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Generic v5 initialization breakage vs a specific GPTNeoX rotary_pct persistence bug; related area, not the same issue." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "Both are regressions, but they affect different models and different behavior: CLIPVision hidden_states vs Qwen VL video-input handling." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "Unrelated topics: LoRA merge with extended vocab causing collapse vs Gemma 4 mm_token_type_ids defaulting for text-only fine-tuning." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Different bugs in different training losses: metric gathering drops tuple labels vs auxiliary-loss normalization in MoE models." + }, + { + "left": "issue:43572", + "right": "issue:46178", + "accept": false, + "reason": "Both involve config/backend compatibility, but StableLm pad_token_idx regression and missing tensorflow_text backend registration are separate issues." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both touch cache/state machinery, but one is missing cache during chunked prefill and the other is recurrent states resetting in modular_qwen3_5." + }, + { + "left": "issue:44617", + "right": "issue:45584", + "accept": false, + "reason": "Different products and symptoms: Sam3Video OOM vs Whisper empty-transcription generation failure." + }, + { + "left": "issue:43504", + "right": "issue:45698", + "accept": false, + "reason": "Both are load/save issues, but BEiT legacy preset loading and wrong custom-module resolution after save_pretrained are not the same concrete bug." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Different transformers serve crashes: KV-cache continuation indexing vs Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:45542", + "right": "issue:46234", + "accept": false, + "reason": "Unrelated backend/import issues: undefined tf backend from tensorboard-only install vs missing falcon_mamba_ssm availability helper." + }, + { + "left": "issue:43673", + "right": "issue:46421", + "accept": false, + "reason": "Both mention chunked prefill and cache, but one reports missing cache in v5.0.0 while the other reports lazy-init recompilation; different observable failures." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Both are config-field regressions, but they affect different models and different missing settings: Llama4 pad_token_id vs Qwen3-MoE sliding_window." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Different models and paths: janus image-generation test failure vs Qwen3.5 flash-attention generation bug." + }, + { + "left": "issue:45381", + "right": "issue:46369", + "accept": false, + "reason": "Different domains and failure modes: Qwen2.5-VL video vision_position_ids bug vs sam3-litetext memory overflow." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "Same general MoE/regression theme is too broad; CLIPVision hidden_states and router load-balancing loss are unrelated concrete issues." + }, + { + "left": "issue:45563", + "right": "issue:45584", + "accept": false, + "reason": "Both are generation-related, but a stale warning for num_return_sequences is not the same bug as Whisper failing on empty transcription after special-token alignment." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 33, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:45588", + "issue:45600", + "issue:45632", + "issue:45803", + "issue:45907", + "issue:46234", + "issue:46369" + ], + "soft_pairs": [ + "issue:44792|issue:45600", + "issue:44534|issue:44913", + "issue:43931|issue:44060", + "issue:44190|issue:44964", + "issue:35141|issue:42490", + "issue:44855|issue:45588", + "issue:44479|issue:44734", + "issue:43827|issue:43994", + "issue:41720|issue:45245", + "issue:44617|issue:46369", + "issue:43065|issue:44610", + "issue:44521|issue:45381", + "issue:44464|issue:45563", + "issue:44933|issue:45362", + "issue:43644|issue:44913", + "issue:43976|issue:44855", + "issue:44077|issue:44877", + "issue:43550|issue:44190", + "issue:45127|issue:45357", + "issue:44462|issue:45907", + "issue:25251|issue:27957", + "issue:43540|issue:44734", + "issue:44112|issue:45600", + "issue:45632|issue:46234", + "issue:43335|issue:44190", + "issue:44737|issue:45803", + "issue:43116|issue:44190" + ], + "prepared_review_unit_hash": "e5b90b5d9db4a8c720234c94cd2380a32680764873ea3873f0b4ef17248b9ddb", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 27, + "serialized_chars": 35181, + "estimated_input_tokens": 8796, + "estimated_eval_tokens": 17848 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: the candidate pairs span unrelated bugs in model loading, config handling, generation, docs, CI, and example scripts. None of the soft edges are safe to collapse into one duplicate group.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44792", + "right": "issue:45600", + "accept": false, + "reason": "Different failures: Janus image-generation test vs Sam3LiteText auto-mapping/CI break." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "One is a v5 non-persistent-buffer initialization bug; the other is GPTNeoXConfig reload losing rotary_pct." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Different models and code paths: Qwen3-VL shape mismatch vs Qwen3-Next tied-weights warning." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Different loaders and symptoms: example-script local dataset issue vs multimodal model load failure." + }, + { + "left": "issue:35141", + "right": "issue:42490", + "accept": false, + "reason": "Embedding reinit on resize is unrelated to the local-path trailing-slash import warning." + }, + { + "left": "issue:44855", + "right": "issue:45588", + "accept": false, + "reason": "Python 3.13 import parse error in DebertaV2Model is unrelated to flash-attention crashing on s_aux=None." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Different subsystems: video-input regression in Qwen VL models vs serve KV-cache continuation indexing." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Docs still referencing pipeline() is not the same as SigLIP2 producing bad outputs with AutoModel/pipeline." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3 auto device-mapping CUDA assert is unrelated to the 2^24 category-count runtime limit." + }, + { + "left": "issue:44617", + "right": "issue:46369", + "accept": false, + "reason": "Same family name, but different model variants and likely different memory causes; too broad to merge." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder is unrelated to the OmDet-Turbo processor/model resolution mismatch." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal, but one is assistant_mask construction and the other is vision_position_ids for qwen2.5-vl video input." + }, + { + "left": "issue:44464", + "right": "issue:45563", + "accept": false, + "reason": "Chunked-generation inconsistency under compile is unrelated to the stale num_return_sequences warning." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "Missing image_utils import is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent-buffer junk filling is a different bug from rotary_pct resetting on reload." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "Python-version support breakage is unrelated to the DebertaV2Model IndentationError." + }, + { + "left": "issue:44077", + "right": "issue:44877", + "accept": false, + "reason": "PatchTSMixer post_init policy change is unrelated to strict granite_speech config loading." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "Torch.compile+SDPA failure in Bamba is unrelated to the image-classification example local-dataset issue." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge collapse with extended vocab is unrelated to incorrect visual encoder keys in save_pretrained." + }, + { + "left": "issue:44462", + "right": "issue:45907", + "accept": false, + "reason": "Tokenizer loading ignoring tokenizer.json is unrelated to list_repo_files failing to forward revision." + }, + { + "left": "issue:25251", + "right": "issue:27957", + "accept": false, + "reason": "Pipeline top_k nesting change and XLMRoberta Flash Attention 2 support are unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError is unrelated to serve-side KV-cache continuation indexing." + }, + { + "left": "issue:44112", + "right": "issue:45600", + "accept": false, + "reason": "Both touch CI, but they concern different stale references in different models/modules." + }, + { + "left": "issue:45632", + "right": "issue:46234", + "accept": false, + "reason": "Cache-path collision for trust_remote_code is unrelated to the missing import_utils helper." + }, + { + "left": "issue:43335", + "right": "issue:44190", + "accept": false, + "reason": "SwitchTransformers sparse-layer bug is unrelated to the example-script dataset loading failure." + }, + { + "left": "issue:44737", + "right": "issue:45803", + "accept": false, + "reason": "XLNet CPU tensor allocation bug is unrelated to the bare-except KeyboardInterrupt swallowing issue." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Different example scripts and different bugs: empty multi-label results vs local dataset loading." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43295|issue:45276", + "issue:45276|issue:45584", + "issue:44938|issue:45542", + "issue:46082|issue:46121", + "issue:36246|issue:41628", + "issue:45632|issue:46421", + "issue:42175|issue:44279", + "issue:45245|issue:45584", + "issue:43404|issue:45216", + "issue:43650|issue:44220", + "issue:44246|issue:45003", + "issue:36246|issue:38617", + "issue:41950|issue:44016", + "issue:44186|issue:44734", + "issue:43479|issue:45375", + "issue:44683|issue:45446", + "issue:43643|issue:44496", + "issue:43673|issue:46424", + "issue:44964|issue:45600", + "issue:43854|issue:44964", + "issue:44016|issue:44220", + "issue:44112|issue:44964", + "issue:27957|issue:35141", + "issue:46123|issue:46369", + "issue:43531|issue:45600", + "issue:43976|issue:44987", + "issue:43526|issue:43901", + "issue:42907|issue:43441" + ], + "prepared_review_unit_hash": "855bbc891532d9a82cdbce666a4c4710bc341858367f15ec3463f8ff823d6f8a", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36072, + "estimated_input_tokens": 9018, + "estimated_eval_tokens": 18292 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is only loosely related by Transformers API/regression themes; the paired issues mostly describe different models, components, or failure modes rather than the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 44246 is the broadest and most representative of the set\u2019s general import/regression theme, but it is still not a true duplicate anchor for the other issues.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44246 has the widest scope and highest discussion activity among the candidates, making it the best single representative issue for the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Both involve embedding/resize behavior, but they are different bugs in different code paths and models." + }, + { + "left": "issue:45276", + "right": "issue:45584", + "accept": false, + "reason": "Unrelated failures: token embedding resize vs Whisper empty-transcription handling after special-token alignment." + }, + { + "left": "issue:44938", + "right": "issue:45542", + "accept": false, + "reason": "Python version load failure and missing TensorFlow backend are unrelated dependency problems." + }, + { + "left": "issue:46082", + "right": "issue:46121", + "accept": false, + "reason": "Different config validation paths: Llama head_dim rules vs rope-parameter JSON list handling." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and different modules." + }, + { + "left": "issue:45632", + "right": "issue:46421", + "accept": false, + "reason": "One is a trust_remote_code cache-path collision; the other is cache lazy init causing recompilation." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "One is about TensorFlow not being included in a torch install extra; the other is a generic dependency issue." + }, + { + "left": "issue:45245", + "right": "issue:45584", + "accept": false, + "reason": "Category-count overflow and empty-transcription generation are completely different failures." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Both mention tied weights, but one is model-specific lm_head tying and the other is a save_pretrained regression." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "No shared bug surface; this looks like a stray/nontechnical issue versus an audio feature extraction bug." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Both touch import behavior, but one is slowness and the other is unsafe sys.modules access." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Different missing imports with no common underlying code-path." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline bug and notebook syntax error are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "Tokenizer padding/NER crash and KV-cache continuation indexing are distinct code paths." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Both involve config defaults/fields, but one is default initialization and the other is strict-field dropping." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Compiled flex_attention breakage on torch>=2.9 is not the same as a version-check bug for AuxRequest import." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "Missing fields with trust_remote_code and unrecognized model_type are different config-loading problems." + }, + { + "left": "issue:43673", + "right": "issue:46424", + "accept": false, + "reason": "Chunked-prefill cache absence and max_cache_length being ignored outside quantized cache are different cache bugs." + }, + { + "left": "issue:44964", + "right": "issue:45600", + "accept": false, + "reason": "Model loading failure and auto_mappings stale references/CI breakage do not describe the same issue." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Different models and symptoms; one is a unit-test loading problem, the other is a general latest-transformers load failure." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error is unrelated to the fbank feature extraction issue." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI stale device override and model loading are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention support for XLMRoberta and embedding reinitialization on resize are different bugs." + }, + { + "left": "issue:46123", + "right": "issue:46369", + "accept": false, + "reason": "Partial-batch result dropping and model memory overflow are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45600", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and stale auto_mappings references are different issues." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python-version compatibility and a specific model-load failure are not duplicates." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "ImageProcessor label reduction and pipeline docs behavior mismatch are unrelated." + }, + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Dequantized save failure and FlashAttention failure in Ministral are distinct model-path 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": 21, + "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:39692", + "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:28282|issue:35141", + "issue:44496|issue:45042", + "issue:43441|issue:45588", + "issue:44016|issue:44062", + "issue:45372|issue:46234", + "issue:43299|issue:45216", + "issue:44792|issue:46424", + "issue:43901|issue:44496", + "issue:43381|issue:45593", + "issue:43824|issue:44933", + "issue:43673|issue:44336", + "issue:43976|issue:44279", + "issue:44496|issue:45759", + "issue:45335|issue:45584", + "issue:44075|issue:44368", + "issue:46082|issue:46399", + "issue:43650|issue:46209", + "issue:43577|issue:45216", + "issue:38175|issue:44617", + "issue:45103|issue:45372", + "issue:44242|issue:45200", + "issue:44279|issue:45405", + "issue:41950|issue:45020", + "issue:43295|issue:44779", + "issue:43299|issue:45357", + "issue:44779|issue:45812", + "issue:44351|issue:45278", + "issue:39692|issue:43901" + ], + "prepared_review_unit_hash": "5739e3405151239b93f628e0c356bceb5f6b65cb1add861dc5c8bf0645d96a26", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36467, + "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 a loose collection of mostly unrelated Transformers issues (imports, model-loading regressions, tokenizer bugs, docs, and a few training/inference edge cases). There are no PRs here, and the soft pairs do not appear to be true duplicates of the same underlying bug/change.", + "confidence": 0.9, + "canonical_issue_reason": "issue:45020 is the broadest umbrella item (recent-version regressions affecting remote_code models), but the cluster is heterogeneous and does not form a clean duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best single representative of the cluster\u2019s general regression theme, but it should not be treated as a true canonical duplicate for the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Different failures: processor/tokenizer regression with images vs resize_token_embeddings not updating decoder.embed_tokens." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "Missing PyTorch import is unrelated to embedding reinitialization after resize." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model/config loading error is unrelated to PIL image processors requiring torchvision." + }, + { + "left": "issue:43441", + "right": "issue:45588", + "accept": false, + "reason": "Both mention flash attention, but one is a Ministral-specific crash and the other is an s_aux=None AttributeError in sink-less models." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and AddedToken special-argument TypeError are distinct bugs." + }, + { + "left": "issue:45372", + "right": "issue:46234", + "accept": false, + "reason": "Different missing imports in different subsystems: Gemma processor loading vs falcon mamba availability helper." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Qwen3VL model loading regression is unrelated to Qwen3.5 save_pretrained checkpoint corruption." + }, + { + "left": "issue:44792", + "right": "issue:46424", + "accept": false, + "reason": "A janus test failure is not the same as cache max length being ignored outside quantized cache." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Documentation mismatch for return_all_scores is unrelated to an unrecognized model config error." + }, + { + "left": "issue:43381", + "right": "issue:45593", + "accept": false, + "reason": "Gradient checkpointing eval-mode behavior is unrelated to D-FINE auxiliary-loss handling." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and different failing modules." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Cache-missing failure during chunked prefill is unrelated to ANSI codes in loading_report." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python-version compatibility and a generic dependency issue are not the same underlying bug." + }, + { + "left": "issue:44496", + "right": "issue:45759", + "accept": false, + "reason": "Different model-loading paths: unrecognized config vs missing text_config unwrap in composite Qwen configs." + }, + { + "left": "issue:45335", + "right": "issue:45584", + "accept": false, + "reason": "Token embedding resize bug is unrelated to Whisper empty-transcription generation failure." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Optimizer argument handling and tie_word_embeddings warnings are distinct issues." + }, + { + "left": "issue:46082", + "right": "issue:46399", + "accept": false, + "reason": "Llama head_dim validation and Qwen3.5 kernel registration are unrelated." + }, + { + "left": "issue:43650", + "right": "issue:46209", + "accept": false, + "reason": "Placeholder data request is unrelated to FP8 inference failure on B200." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "BLIP2 dtype propagation and Qwen3.5 save_pretrained checkpoint issues do not match." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Zero-probabilities in SigLIP2 is unrelated to Sam3Video CUDA OOM." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "auto_docstring annotation parsing crash is unrelated to Gemma processor import failure." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Mixture-of-experts load-balancing loss and Gemma 4 mm_token_type_ids are different model-training bugs." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "A transformers dependency problem is unrelated to a PEFT minimum-version bump." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "Video-classification processor lookup bug is unrelated to remote_code model breakage." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Processor/tokenizer regression with images is unrelated to Deepseek tokenizer output changes." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3VL loading and Qwen3.5 visual-encoder key saving are different code paths." + }, + { + "left": "issue:44779", + "right": "issue:45812", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different model families and likely different root causes." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing HybridCache import is not enough to treat this as the same bug as a broad import-error report." + }, + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "A SigLIP2 example with multiple errors is unrelated to a docs mismatch about return_all_scores." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 36, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:33290", + "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:41950|issue:45823", + "issue:43479|issue:43525", + "issue:44683|issue:45542", + "issue:43976|issue:45405", + "issue:44462|issue:45081", + "issue:44297|issue:44462", + "issue:43299|issue:43854", + "issue:43450|issue:44279", + "issue: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:44361|issue:45593", + "issue:44261|issue:44263", + "issue:44671|issue:44779", + "issue:44016|issue:45823", + "issue:43723|issue:44162", + "issue:42175|issue:43976", + "issue:43901|issue:43994", + "issue:43441|issue:44683" + ], + "prepared_review_unit_hash": "f2d99915d9072d61d264d7676da37e5ddf552ba50d39762c70333bbeacccaf45", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 35414, + "estimated_input_tokens": 8854, + "estimated_eval_tokens": 17964 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it mixes unrelated Transformer v5 regressions, tokenizer/model-loading bugs, processor/config issues, deepspeed training behavior, packaging, and docs. The soft-similarity links are mostly superficial and do not look like true duplicates.", + "confidence": 0.95, + "canonical_issue_reason": "No single canonical issue: the cluster mixes unrelated bugs across tokenizers, multimodal processing, config, deepspeed, packaging, and docs.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the most active/referenced node, but it is not a true duplicate hub and should not be treated as a canonical merge target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Both are multimodal-related, but one is about all-zero assistant masks and the other about temporal position_id scaling; different code paths and fixes." + }, + { + "left": "issue:41950", + "right": "issue:45823", + "accept": false, + "reason": "Video-classification processor selection and Gemma4 device-mismatch during forward are unrelated bugs." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Different config problems: multimodal defaults vs missing pad_token_id on Llama4Config." + }, + { + "left": "issue:44683", + "right": "issue:45542", + "accept": false, + "reason": "Torch flex_attention compilation on 2.9 and TensorFlow packaging/backend availability are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python version support regression and a bumped PEFT dependency are separate release issues." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer ignoring repository tokenizer.json and Mistral regex patch crashing are different loading paths." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "Tokenizer save metadata mismatch is not the same bug as AutoTokenizer ignoring tokenizer.json." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Different model-loading failures for different models; no shared concrete code-path." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor batched-shape bug and a vague dependency issue are not the same defect." + }, + { + "left": "issue:43650", + "right": "issue:46132", + "accept": false, + "reason": "A placeholder 'ADD THE DATA' issue and AttentionInterface.register behavior change are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45907", + "accept": false, + "reason": "Mistral regex patch crash and list_repo_files missing revision forwarding are separate repository-loading bugs." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Inconsistent lm_head serialization and resize_token_embeddings not propagating are different weight-handling problems." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error and init_empty_weights argument mismatch are unrelated." + }, + { + "left": "issue:43441", + "right": "issue:45834", + "accept": false, + "reason": "FlashAttention failure in Ministral and a long-OCR index error in Kosmos2.5 are different model-specific bugs." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Quantized saving failure and dequantized loading with device_map=auto are distinct quantization/offload issues." + }, + { + "left": "issue:43653", + "right": "issue:45593", + "accept": false, + "reason": "Tokenizer special-token registration and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "processor.save_pretrained missing files and gemma-4 zero3 from_pretrained are different save/load issues." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor save-file omission and a Chinese report about filtering slowdown are not the same bug." + }, + { + "left": "issue:45600", + "right": "issue:45712", + "accept": false, + "reason": "Both are cleanup/CI-related, but they target different stale-reference problems in different files." + }, + { + "left": "issue:33290", + "right": "issue:45656", + "accept": false, + "reason": "OOM with Adafactor under deepspeed and optimizer.step being called twice are not clearly the same underlying bug." + }, + { + "left": "issue:44361", + "right": "issue:45593", + "accept": false, + "reason": "MLukeTokenizer AttributeError and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "A q_a_layernorm precision/config issue and torch.split return-value handling are different bugs." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT MLM prediction regression and Deepseek tokenizer output regression are different components and fixes." + }, + { + "left": "issue:44016", + "right": "issue:45823", + "accept": false, + "reason": "Notebook syntax error and Gemma4 PLE device mismatch are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "AutoTokenizer v5 loading issues and ESM2 breakage are different model/loading problems." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "TensorFlow backend packaging and Python version support are separate dependency/release issues." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Documentation mismatch for return_all_scores and incorrect SigLIP2 predictions are not the same bug." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "FlashAttention failure and compiled flex_attention failure are different attention-kernel problems, not one duplicate." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 37, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:44857", + "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:43901|issue:44246", + "issue:43525|issue:43881", + "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:44246|issue:44496", + "issue:44295|issue:45397", + "issue:44336|issue:44373", + "issue:44242|issue:45593", + "issue:43526|issue:45003", + "issue:42491|issue:43299", + "issue:43531|issue:45362", + "issue:41950|issue:44220", + "issue:44246|issue:44871", + "issue:42886|issue:46132", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:44336|issue:44485", + "issue:45003|issue:45561", + "issue:44734|issue:46424", + "issue:44779|issue:45292", + "issue:44336|issue:46424", + "issue:44242|issue:44857" + ], + "prepared_review_unit_hash": "319e88a167032d5f793c72a1a35bc5f1334825790f0d3fa814c04ed3db60d7d8", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36900, + "estimated_input_tokens": 9225, + "estimated_eval_tokens": 18706 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is heterogeneous: most items are unrelated bug reports spanning imports, model loading, cache handling, pipelines, docs, and testing. None of the soft pairs look like true duplicates with the same underlying bug/change, so I rejected them all.", + "confidence": 0.46, + "canonical_issue_reason": "issue:44521 is the most concrete and actionable standalone bug report here: it is open, has multiple inbound references, and describes a specific multimodal apply_chat_template failure with clear behavior to fix. There is no strong duplicate cluster, so it is the best practical canonical issue for this mixed set.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the strongest representative issue overall because it is current, well-scoped, and tied to a core user-facing API with observable broken behavior.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Different subsystems and failure modes: generation cache during chunked prefill vs CLIP hidden_states output." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Both mention apply_chat_template, but one is an audio sampling-rate default bug and the other is a multimodal assistant-mask bug." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer save/load metadata mismatch is unrelated to gemma-4 Zero3 from_pretrained loading." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Missing Qwen2.5-VL export and Python 3.9/3.10 compatibility are different problems." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Docs mismatch for return_all_scores is unrelated to intermittent import slowness." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Different models and code paths: Llama4Config pad_token_id vs glm-4v-9b loading." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Llama4 config field absence and AutoConfig trust_remote_code missing fields are not the same bug." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:43784", + "right": "issue:45542", + "accept": false, + "reason": "Both are import-time issues, but one is sentence-transformers missing nn and the other is TensorFlow backend handling." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "AutoImageProcessor import failure is unrelated to the Python version support issue." + }, + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Attention kernel divergence under torch.compile is unrelated to ANSI color output in loading_report." + }, + { + "left": "issue:45103", + "right": "issue:46399", + "accept": false, + "reason": "Auto-docstring annotation handling and Qwen3.5 kernel registration both involve decorators, but not the same underlying bug." + }, + { + "left": "issue:45561", + "right": "issue:45563", + "accept": false, + "reason": "Testing-utils xdist race and stale num_return_sequences warning are different concerns." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import latency is unrelated to an unrecognized model/config.json model_type error." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Reading position_ids after buffer registration is unrelated to gemma-4 Zero3 loading." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "ANSI terminal handling and a wrong position_ids docstring are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:45593", + "accept": false, + "reason": "Load-balancing loss omission in one model family is unrelated to D-FINE auxiliary loss behavior." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug and sys.modules access safety are different code paths." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3 MoE, but one is LoRA compatibility across versions and the other is a Qwen3VL-MoE loading regression; not the same concrete fix." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Sliding_window behavior in Qwen3-MoE is unrelated to Qwen3.5 chat crashes." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline image-processor lookup and fbank feature extraction are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Import slowness and Gemma-3 eos_token_id inconsistency do not share a concrete bug." + }, + { + "left": "issue:42886", + "right": "issue:46132", + "accept": false, + "reason": "Offline cache loading and AttentionInterface.register behavior are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but for different symbols and different packages." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are missing exports from transformers, but they are different symbols with no evidence of a shared root cause." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI loading-report output and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:45003", + "right": "issue:45561", + "accept": false, + "reason": "Unsafe sys.modules access is unrelated to the xdist captured_info.txt race." + }, + { + "left": "issue:44734", + "right": "issue:46424", + "accept": false, + "reason": "KV-cache continuation tensor indexing and max_cache_length handling are different cache bugs." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Deepseek tokenizer regression and resize_token_embeddings not updating output_embeddings are unrelated." + }, + { + "left": "issue:44336", + "right": "issue:46424", + "accept": false, + "reason": "Terminal color output and cache-length configuration have no shared underlying bug." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "One is a missing auxiliary-loss path; the other is an AMP/CUDA crash in LwDetrImageLoss." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43653", + "issue:43673", + "issue:43901", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44246", + "issue:44295", + "issue:44466", + "issue:44479", + "issue:44483", + "issue:44496", + "issue:44561", + "issue:44749", + "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:46123", + "issue:46170", + "issue:46202", + "issue:46291", + "issue:46369", + "issue:46399", + "issue:46431" + ], + "soft_pairs": [ + "issue:44295|issue:44749", + "issue:43475|issue:46369", + "issue:45859|issue:46399", + "issue:36246|issue:43596", + "issue:46291|issue:46399", + "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:45292|issue:45701", + "issue:45412|issue:46123", + "issue:45375|issue:45759", + "issue:44561|issue:46399", + "issue:44877|issue:45375", + "issue:45685|issue:46431", + "issue:44987|issue:45459", + "issue:44466|issue:45335", + "issue:44995|issue:45200", + "issue:45397|issue:45656", + "issue:45137|issue:45600", + "issue:44479|issue:46170", + "issue:43475|issue:43540", + "issue:43653|issue:44162", + "issue:44483|issue:46202", + "issue:44060|issue:44077" + ], + "prepared_review_unit_hash": "c9623ebb04864d6bb0c7f1059ccd9613c1847212c58c63c9c5fbfab4f0e6e3a3", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36525, + "estimated_input_tokens": 9132, + "estimated_eval_tokens": 18520 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a grab-bag of distinct Transformers bugs/regressions across unrelated subsystems. None of the soft pairs look like the same underlying change or defect, so all soft edges should be rejected.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 44479 is the clearest representative bug report: a concrete v5.3 video-input regression affecting several related Qwen model variants, with a well-defined scope and failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44479 is the strongest standalone issue to anchor on because it is broad, specific, and directly describes a concrete regression in a shared code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Unrelated problems: position_ids buffer handling vs. a slowdown after upgrading Transformers." + }, + { + "left": "issue:43475", + "right": "issue:46369", + "accept": false, + "reason": "Both mention SAM 3, but one is a missing output attribute and the other is a memory overflow in a different model variant." + }, + { + "left": "issue:45859", + "right": "issue:46399", + "accept": false, + "reason": "Same general Qwen family, but different failures: CPU offload incompatibility vs. kernel registration crash." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Completely different areas: missing Qwen2.5 VL import vs. DeepSpeed/BERT zero3 index error." + }, + { + "left": "issue:46291", + "right": "issue:46399", + "accept": false, + "reason": "Both touch kernels, but one is an external kernels dependency change and the other is a Qwen3.5 rotary-pos-emb registration bug." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mismatch for TextClassificationPipeline vs. unsafe sys.modules access in modeling_utils are unrelated." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "Specific Nemotron Flash Attention unsupported error vs. a generic Flash Attention 2.0 issue; not the same bug." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Chunked prefill cache bug and removal of is_torch_fx_available breakage are different code paths." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage is unrelated to Gemma 4 mm_token_type_ids defaulting to zeros." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading vs. sys.modules access safety are distinct failures." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Both involve DeepSpeed/ZeRO3, but the reported symptoms and likely root causes differ." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio processor sampling-rate default bug vs. slow import issue; no shared defect." + }, + { + "left": "issue:45292", + "right": "issue:45584", + "accept": false, + "reason": "Resize-token-embeddings propagation bug is unrelated to Whisper empty-transcription generation failure." + }, + { + "left": "issue:45292", + "right": "issue:45701", + "accept": false, + "reason": "Embedding resize behavior and version-dependent tokenization changes are different issues." + }, + { + "left": "issue:45412", + "right": "issue:46123", + "accept": false, + "reason": "RT-DETR memory cleanup and batch-finalization logic are unrelated." + }, + { + "left": "issue:45375", + "right": "issue:45759", + "accept": false, + "reason": "Both are Qwen multimodal config issues, but one is a missing vision config field and the other is from_config unwrapping behavior." + }, + { + "left": "issue:44561", + "right": "issue:46399", + "accept": false, + "reason": "Breaking trust_remote_code from a removed helper is not the same as a Qwen3.5 kernel registration crash." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Strict granite_speech config loading and Qwen config field dropping are unrelated config problems." + }, + { + "left": "issue:45685", + "right": "issue:46431", + "accept": false, + "reason": "MOE histogram MPS dtype error is unrelated to missing chat template support in a processor." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Loading failure for a specific model/package and protobuf-decoding error handling are different concerns." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both involve weights/embeddings, but one is tied-weight serialization and the other is resize_token_embeddings propagation." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "GlmMoeDsa stale cache crash is unrelated to Gemma 4 multimodal token defaults." + }, + { + "left": "issue:45397", + "right": "issue:45656", + "accept": false, + "reason": "Both mention DeepSpeed, but one is a gemma-4 zero3 loading bug and the other is a double optimizer-step issue." + }, + { + "left": "issue:45137", + "right": "issue:45600", + "accept": false, + "reason": "ZeRO3 deque pop error and auto_mappings CI breakage are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:46170", + "accept": false, + "reason": "Qwen video-input regression vs. Windows ROCm distributed import/type error are unrelated." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Different Qwen/SAM model families and different video-processing failures." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer special-token registration bug vs. ESM2 breakage are unrelated." + }, + { + "left": "issue:44483", + "right": "issue:46202", + "accept": false, + "reason": "Chat completions request validation and Gemma roundtrip tokenization failure are different bugs." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Qwen3-Next tied-weights warning and patchtsmixer post_init allowance are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:42371", + "issue:42886", + "issue:42994", + "issue:43065", + "issue:43232", + "issue:43408", + "issue:43475", + "issue:43482", + "issue:43643", + "issue:43646", + "issue:43673", + "issue:43701", + "issue:43931", + "issue:44038", + "issue:44162", + "issue:44188", + "issue:44297", + "issue:44361", + "issue:44368", + "issue:44623", + "issue:44749", + "issue:44869", + "issue:44871", + "issue:44877", + "issue:44933", + "issue:44977", + "issue:44995", + "issue:45092", + "issue:45229", + "issue:45230", + "issue:45356", + "issue:45381", + "issue:45397", + "issue:45563", + "issue:45584", + "issue:45588", + "issue:45656", + "issue:45987", + "issue:46082", + "issue:46123", + "issue:46132", + "issue:46170", + "issue:46202", + "issue:46421" + ], + "soft_pairs": [ + "issue:42886|issue:45229", + "issue:44995|issue:45588", + "issue:44368|issue:44977", + "issue:43065|issue:43408", + "issue:44368|issue:45563", + "issue:44188|issue:46202", + "issue:44877|issue:46170", + "issue:45987|issue:46082", + "issue:43931|issue:44368", + "issue:45584|issue:46170", + "issue:44188|issue:46421", + "issue:43065|issue:46123", + "issue:44297|issue:44623", + "issue:44361|issue:44749", + "issue:44623|issue:45656", + "issue:44869|issue:45356", + "issue:44749|issue:45656", + "issue:43646|issue:44162", + "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:43475|issue:45381", + "issue:28282|issue:30990", + "issue:41628|issue:44933", + "issue:42371|issue:45092" + ], + "prepared_review_unit_hash": "57216a432441ce494149f742da12c41f497938cafcdb1f2bfa4764ee8b58bcdf", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35575, + "estimated_input_tokens": 8894, + "estimated_eval_tokens": 18044 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster looks like a broad false-positive similarity bucket rather than a duplicate set: it spans unrelated issues in loading, tokenization, generation, config, cache, model-specific regressions, and environment/import failures. No soft edge should be accepted.", + "confidence": 0.9, + "canonical_issue_reason": "Issue:43646 is the broadest v5-related regression in the set and the closest thing to a hub, but the surrounding issues are mostly unrelated, so it should not be treated as a true duplicate anchor for the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "If one issue must represent the cluster, 43646 is the most general and central; however, it only covers a narrow slice of the items and is not a good duplicate target for the rest.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline tokenizer-cache loading failure and multi-GPU CUDA OOM are different bugs with different code paths." + }, + { + "left": "issue:44995", + "right": "issue:45588", + "accept": false, + "reason": "Stale indexer-cache crash and flash-attention s_aux=None AttributeError are unrelated failures." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A warning about tie_word_embeddings during LoRA fine-tuning is not the same as a flash-attention generation bug." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and sam3_video/sam3_tracker config mismatch are distinct SAM 3 issues." + }, + { + "left": "issue:44368", + "right": "issue:45563", + "accept": false, + "reason": "These are different warnings in different subsystems; one is config advice, the other is a paged generation warning bug." + }, + { + "left": "issue:44188", + "right": "issue:46202", + "accept": false, + "reason": "torch.compile attention-kernel divergence is unrelated to a Gemma pipeline roundtrip tokenization failure." + }, + { + "left": "issue:44877", + "right": "issue:46170", + "accept": false, + "reason": "Granite speech config strictness and Windows/ROCm torch.distributed import problems are separate environment/config issues." + }, + { + "left": "issue:45987", + "right": "issue:46082", + "accept": false, + "reason": "StaticCache returning a tensor and LlamaConfig rejecting head_dim are different API/validation problems." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Qwen3-VL weight-shape mismatch is not the same bug as a LoRA tie_word_embeddings warning." + }, + { + "left": "issue:45584", + "right": "issue:46170", + "accept": false, + "reason": "Whisper empty-transcription generation failure is unrelated to the Windows/ROCm torch.distributed import issue." + }, + { + "left": "issue:44188", + "right": "issue:46421", + "accept": false, + "reason": "Diverging attention kernels under torch.compile and cache lazy-init recompilation are different cache/compilation regressions." + }, + { + "left": "issue:43065", + "right": "issue:46123", + "accept": false, + "reason": "A dummy Conv2d in Sam3PixelDecoder and MaskGenerationPipeline dropping the final partial batch are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch and processor.save_pretrained missing files are similar area-wise but not the same concrete failure." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "MLukeTokenizer AttributeError and a performance regression after upgrading Transformers are different issues." + }, + { + "left": "issue:44623", + "right": "issue:45656", + "accept": false, + "reason": "Missing files from processor.save_pretrained and a double optimizer.step under deepspeed are unrelated." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decode crash on trailing replacement character and Kimi tokenizer codec handling are different model/tokenizer bugs." + }, + { + "left": "issue:44749", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer slowdown after upgrade and deepspeed calling optimizer.step twice do not share a root cause." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Generic custom model initialization breakage and ESM2 being broken are separate model-specific regressions." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "resume_from_checkpoint key mismatch is not the same as a tokenizer performance regression." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "sync_gpus generation-kwargs update and chunked-prefill cache-missing are both generation bugs, but not the same concrete defect." + }, + { + "left": "issue:44871", + "right": "issue:46132", + "accept": false, + "reason": "Gemma eos_token_id inconsistency and AttentionInterface.register behavior change are unrelated API issues." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "A generic bug report and a Gemma-4 ZeRO-3 from_pretrained failure are not enough evidence to be duplicates." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure and Qwen2.5-GGUF loading failure are opposite operations and different bugs." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe v5 bug and processor.save_pretrained missing files are unrelated model/subsystem issues." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "Sam3VisionEncoderOutput missing attribute and trust_remote_code AutoConfig missing fields are different remote-code failures." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Sam3 vision encoder attribute loss and Qwen2.5-VL video vision_position_ids are unrelated model-specific bugs." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch not found ImportError and SentenceTransformers loading hang have different root causes and symptoms." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "AutoImageProcessor import failure and image_utils nonexistent import are both import issues, but not the same missing symbol/module." + }, + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "TF32 API warning and old InternVL2 remote-code/meta-init incompatibility are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:30990", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42673", + "issue:42831", + "issue:42898", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43377", + "issue:43504", + "issue:43525", + "issue:43526", + "issue:43704", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:43976", + "issue:44038", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44351", + "issue:44496", + "issue:44568", + "issue:44623", + "issue:44792", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45081", + "issue:45103", + "issue:45137", + "issue:45310", + "issue:45375", + "issue:45406", + "issue:45478", + "issue:45561", + "issue:45712", + "issue:45753", + "issue:46077", + "issue:46123", + "issue:46202", + "issue:46326", + "issue:46399" + ], + "soft_pairs": [ + "issue:45375|issue:45478", + "issue:45406|issue:46202", + "issue:42994|issue:43525", + "issue:43723|issue:46326", + "issue:43784|issue:46399", + "issue:45310|issue:45375", + "issue:45081|issue:45103", + "issue:43873|issue:44945", + "issue:42673|issue:43704", + "issue:44964|issue:45137", + "issue:41628|issue:44351", + "issue:43232|issue:44188", + "issue:44222|issue:45310", + "issue:44945|issue:45753", + "issue:42994|issue:43824", + "issue:46077|issue:46123", + "issue:44038|issue:45310", + "issue:36010|issue:41628", + "issue:43232|issue:44792", + "issue:42831|issue:44222", + "issue:42175|issue:45712", + "issue:44496|issue:45561", + "issue: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", + "issue:30990|issue:35141", + "issue:44623|issue:45310" + ], + "prepared_review_unit_hash": "ba06df8f9d4d8f145c409346184b0a8c6eb34f2a6358a2c5ab22608a1f03bdfa", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36912, + "estimated_input_tokens": 9228, + "estimated_eval_tokens": 18712 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a loose mix of unrelated Transformers issues. There are a couple of exact duplicate reports, but most soft pairs are only superficially similar and should stay separate.", + "confidence": 0.69, + "canonical_issue_reason": "Issue 45310 is the best canonical anchor because it is a concrete, reproducible regression report and has an exact duplicate counterpart (45478) with the same title and symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45310 is the strongest representative issue in this set: specific scope, clear failure mode, and an exact duplicate exist nearby, making it the most suitable cluster anchor.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Same general model family, but different bugs: a missing config field vs a from_pretrained load regression." + }, + { + "left": "issue:45406", + "right": "issue:46202", + "accept": false, + "reason": "Both involve Gemma4, but one is a processor attribute error and the other is a pipeline/tokenization roundtrip failure." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized model saving failure and Llama4Config pad_token_id absence are unrelated failure modes." + }, + { + "left": "issue:43723", + "right": "issue:46326", + "accept": false, + "reason": "Tokenizer loading in v5 and mistral_common continue_final_message are different code paths and symptoms." + }, + { + "left": "issue:43784", + "right": "issue:46399", + "accept": false, + "reason": "Import-time NameError in sentence-transformers is unrelated to the Qwen3.5 kernel registration crash." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Both mention Qwen3.5, but one is a from_pretrained regression and the other is a config-field omission." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Both are AttributeErrors, but they affect different components and root causes." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization/offloading behavior and pipeline-parallel output correctness are separate bugs." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "These are the same reported bug with the same title: Qwen3ForCausalLM VRAM leak across multiple dataloader threads." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "A model-load failure for Phi-4 multimodal is unrelated to a DeepSpeed ZeRO3 deque underflow." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but for different missing symbols and different regressions." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Generation kwargs handling after sync_gpus is unrelated to torch.compile attention-kernel branching." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "Both involve MoE/FP8-adjacent areas, but one is save_pretrained on FP8 and the other is a from_pretrained load regression." + }, + { + "left": "issue:44945", + "right": "issue:45753", + "accept": false, + "reason": "Incorrect pipeline-parallel output and an infinite-loop image bug are different issues." + }, + { + "left": "issue:42994", + "right": "issue:43824", + "accept": false, + "reason": "Quantized model saving and missing Qwen2_5_VLForConditionalGeneration import are unrelated." + }, + { + "left": "issue:46077", + "right": "issue:46123", + "accept": false, + "reason": "Bidirectional attention masking and final partial-batch result dropping are different pipeline/model bugs." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Both mention Qwen and versioning, but the model variants and failure modes do not match." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Different import regressions: GenerationMixin versus AutoImageProcessor." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation sync kwargs bug is unrelated to the janus test failure." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both touch FP8, but one is an accuracy issue and the other is a save_pretrained failure." + }, + { + "left": "issue:42175", + "right": "issue:45712", + "accept": false, + "reason": "TensorFlow backend packaging and leftover dummy-class cleanup are unrelated." + }, + { + "left": "issue:44496", + "right": "issue:45561", + "accept": false, + "reason": "Model config recognition failure and a pytest-xdist file race are not the same bug." + }, + { + "left": "issue:43526", + "right": "issue:45561", + "accept": false, + "reason": "BeitImageProcessorFast label handling and xdist captured_info.txt races are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:45561", + "accept": false, + "reason": "Legacy preset loading failure and xdist file contention are different problems." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "TensorFlow backend install issue is unrelated to a missing image_utils import." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI padding-mask behavior and MoE load-balancing loss gating are separate model bugs." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer cleanup behavior in v5 and add_special_tokens BOS/EOS behavior are distinct changes." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility and missing image_utils import are unrelated." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift in eval and a ZeRO3 deque pop error are unrelated." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "Sentence Transformers loading stall and embedding resize reinitialization are different issues." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Processor save_pretrained missing files is unrelated to the Qwen3.5 MoE from_pretrained regression." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The anchor choice is supported by the packet, the duplicate claims are conservative, and the soft-edge rejections mostly distinguish different failure modes rather than over-merging them." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 41, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:22355", + "issue:27957", + "issue:28282", + "issue:30333", + "issue:33290", + "issue:34689", + "issue:35141", + "issue:36331", + "issue:42222", + "issue:42831", + "issue:42898", + "issue:43122", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43784", + "issue:43824", + "issue:43874", + "issue:44016", + "issue:44060", + "issue:44075", + "issue:44079", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44448", + "issue:44568", + "issue:44623", + "issue:44734", + "issue:44779", + "issue:44869", + "issue:44933", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45200", + "issue:45245", + "issue:45310", + "issue:45356", + "issue:45405", + "issue:45656", + "issue:45712", + "issue:45812", + "issue:46132", + "issue:46209", + "issue:46231", + "issue:46431" + ], + "soft_pairs": [ + "issue:30333|issue:33290", + "issue:42222|issue:45200", + "issue:43479|issue:44869", + "issue:42898|issue:45812", + "issue:43479|issue:44079", + "issue:43784|issue:43824", + "issue:43723|issue:44568", + "issue:42898|issue:44779", + "issue:44016|issue:46231", + "issue:44734|issue:45103", + "issue:42898|issue:45356", + "issue:43874|issue:45092", + "issue:44295|issue:45656", + "issue:44075|issue:46231", + "issue:42898|issue:44448", + "issue:44060|issue:45125", + "issue:44222|issue:44297", + "issue:45405|issue:45712", + "issue:44933|issue:45712", + "issue:44297|issue:45656", + "issue:27957|issue:30333", + "issue:44222|issue:44623", + "issue:30333|issue:34689", + "issue:44448|issue:46132", + "issue:42831|issue:44623", + "issue:46209|issue:46431", + "issue:43479|issue:43644", + "issue:30333|issue:36331", + "issue:30333|issue:35141", + "issue:22355|issue:28282", + "issue:45245|issue:46132", + "issue:44297|issue:45310", + "issue:43122|issue:46209" + ], + "prepared_review_unit_hash": "b2ee7b8c1c0552c676d649818d6d1fec0b31d9e34a60c001ac0aa83258615c6e", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "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 set is highly heterogeneous; the soft edges mostly connect superficially similar transformer/tokenizer/FP8/training bugs, but they do not look like the same underlying defect. No soft pair should be merged.", + "confidence": 0.95, + "canonical_issue_reason": "Issue 45812 is the clearest representative bug report in this mixed set: it describes a concrete, high-impact v4\u2192v5 tokenizer regression affecting all Granite models, with a specific observable failure mode.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45812 is the best issue for triage because it is precise, broadly impactful, and regression-oriented; the other issues are either narrower, orthogonal, or only loosely related by subsystem.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:30333", + "right": "issue:33290", + "accept": false, + "reason": "MLflow reporting on failed training jobs is unrelated to Deepspeed OOMs in Adafactor." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "VitPose breakage and Gemma 4 multimodal token defaults are different model-specific bugs." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Phi4MultimodalConfig default-init bug is unrelated to Whisper timestamp decode crashing." + }, + { + "left": "issue:42898", + "right": "issue:45812", + "accept": false, + "reason": "Both are v5 tokenizer regressions, but one is about clean_up_tokenization_spaces behavior and the other about Granite token IDs; different code paths and symptoms." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config default initialization and ModelOutput key assignment are distinct internal issues." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Sentence-transformers import NameError and Qwen2_5_VL missing export are unrelated import failures." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Both involve tokenizers, but one is about AutoTokenizer loading in v5 and the other is a model-specific add_special_tokens regression." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "General tokenization behavior change and DeepSeek tokenizer regression are different bugs." + }, + { + "left": "issue:44016", + "right": "issue:46231", + "accept": false, + "reason": "Notebook syntax error and a documentation parameter-name mismatch are not the same issue." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "Serve KV-cache continuation crash and auto_docstring annotation handling are unrelated code paths." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but they concern different models and different failure modes." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Missing image-patch method in GLM46V processor and InternVL2 meta-init incompatibility are separate multimodal bugs." + }, + { + "left": "issue:44295", + "right": "issue:45656", + "accept": false, + "reason": "Buffer registration/readback bug and Deepspeed double optimizer step are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:46231", + "accept": false, + "reason": "SGD arguments being ignored is a runtime optimizer bug, not a documentation typo." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Different tokenizer behavior regression vs Pegasus model output regression; not the same defect." + }, + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Incorrect tied-weights warning and missing tensor-parallel plan are different Qwen3.5-MoE issues." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "FP8 MoE save_pretrained and tokenizer save_pretrained metadata mismatch are distinct serialization problems." + }, + { + "left": "issue:45405", + "right": "issue:45712", + "accept": false, + "reason": "PEFT version pinning and leftover dummy classes are unrelated maintenance issues." + }, + { + "left": "issue:44933", + "right": "issue:45712", + "accept": false, + "reason": "Missing image_utils import and dummy class cleanup are different repository hygiene problems." + }, + { + "left": "issue:44297", + "right": "issue:45656", + "accept": false, + "reason": "Tokenizer config save metadata and Deepspeed double-step behavior are unrelated." + }, + { + "left": "issue:27957", + "right": "issue:30333", + "accept": false, + "reason": "Flash Attention 2 on XLM-Roberta and MLflow failure reporting are unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "FP8 MoE save_pretrained and processor.save_pretrained missing files are both save-related, but not the same object or code path." + }, + { + "left": "issue:30333", + "right": "issue:34689", + "accept": false, + "reason": "Failed-job MLflow handling is unrelated to Llama 3.2 Vision model loading regression." + }, + { + "left": "issue:44448", + "right": "issue:46132", + "accept": false, + "reason": "Pegasus output regression and AttentionInterface.register behavior change are separate issues." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue and processor save_pretrained file omission do not share the same underlying bug." + }, + { + "left": "issue:46209", + "right": "issue:46431", + "accept": false, + "reason": "FP8 inference on B200 and missing chat template in a processor are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Incorrect multimodal config defaults and junk-filled non-persistent buffers are distinct initialization bugs." + }, + { + "left": "issue:30333", + "right": "issue:36331", + "accept": false, + "reason": "MLflow failure reporting and CustomTrainer compute_loss signature mismatch are unrelated trainer issues." + }, + { + "left": "issue:30333", + "right": "issue:35141", + "accept": false, + "reason": "Failed-job reporting and embedding reinitialization after resize are different bugs." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Missing transformers.onnx module and missing PyTorch dependency are separate import errors." + }, + { + "left": "issue:45245", + "right": "issue:46132", + "accept": false, + "reason": "Category-count runtime limit and AttentionInterface.register behavior change are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch and Qwen3.5-MoE from_pretrained error are different regressions." + }, + { + "left": "issue:43122", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer behavior change across versions and FP8 inference on B200 are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43232", + "issue:43408", + "issue:43475", + "issue:43644", + "issue:43723", + "issue:43824", + "issue:44038", + "issue:44190", + "issue:44206", + "issue:44222", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44295", + "issue:44360", + "issue:44373", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44623", + "issue:44734", + "issue:44829", + "issue:44871", + "issue:44977", + "issue:44991", + "issue:45071", + "issue:45092", + "issue:45103", + "issue:45310", + "issue:45357", + "issue:45381", + "issue:45406", + "issue:45834", + "issue:46002", + "issue:46050", + "issue:46082", + "issue:46305", + "issue:46328", + "issue:46402", + "issue:46421", + "issue:46424" + ], + "soft_pairs": [ + "issue:43475|issue:46305", + "issue:41628|issue:42757", + "issue:43723|issue:44977", + "issue:45071|issue:46082", + "issue:45381|issue:46305", + "issue:44206|issue:45103", + "issue:43232|issue:46424", + "issue:44190|issue:46082", + "issue:44222|issue:44484", + "issue:44734|issue:46421", + "issue:44295|issue:44623", + "issue:44263|issue:45092", + "issue:43232|issue:45834", + "issue:44373|issue:44485", + "issue:43723|issue:44871", + "issue:43723|issue:44991", + "issue:44483|issue:45406", + "issue:44279|issue:46328", + "issue:44263|issue:44360", + "issue:42907|issue:46050", + "issue:30990|issue:38175", + "issue:45092|issue:45357", + "issue:43644|issue:46305", + "issue:44829|issue:46002", + "issue:44261|issue:45092", + "issue:42831|issue:45310", + "issue:41628|issue:43824", + "issue:42831|issue:44038", + "issue:43408|issue:46402" + ], + "prepared_review_unit_hash": "7a9ffaf4477106e910be7c97e7162d1c4885ac431d5b1e56ce08e84bbb8ad724", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35554, + "estimated_input_tokens": 8889, + "estimated_eval_tokens": 18034 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous and does not represent one coherent duplicate set. The only plausible duplicate pair is the v5 tokenizer-loading regression (43723/44991); all other soft edges look like distinct bugs in different subsystems or models.", + "confidence": 0.58, + "canonical_issue_reason": "Issue 43723 is the best umbrella candidate because it states the broad v5 tokenizer-loading failure, which is the closest match to the only plausible duplicate pair in the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "43723 is the most representative issue here: it is broad enough to subsume a specific tokenizer-loading instance like 44991, while the rest of the cluster is unrelated.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43475", + "right": "issue:46305", + "accept": false, + "reason": "Different models and failure modes: SAM3 video encoder attribute error vs MoonshineStreaming sliding-window off-by-one." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they involve different missing symbols in different packages; not the same bug." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Tokenizer-loading regression vs flash-attention generation quality issue; unrelated code paths." + }, + { + "left": "issue:45071", + "right": "issue:46082", + "accept": false, + "reason": "Both concern config validation, but they are different checks on different configs and are not the same underlying defect." + }, + { + "left": "issue:45381", + "right": "issue:46305", + "accept": false, + "reason": "Both are off-by-one style reports, but they affect different models/components and do not look like one shared bug." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "Feature extractor argument crash and auto-docstring AttributeError are unrelated." + }, + { + "left": "issue:43232", + "right": "issue:46424", + "accept": false, + "reason": "Both are generation/cache related, but one is a sync_gpus kwargs bug and the other is cache-length handling for quantized cache." + }, + { + "left": "issue:44190", + "right": "issue:46082", + "accept": false, + "reason": "Local dataset loading in an example script vs LlamaConfig head_dim validation; unrelated." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "Saving FP8 MoE weights and max_shard_size behavior are different save_pretrained concerns." + }, + { + "left": "issue:44734", + "right": "issue:46421", + "accept": false, + "reason": "Serve KV-cache indexing crash vs cache lazy-init recompilation; different concrete problems." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Position_ids buffer access bug and missing processor save files are unrelated." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "DSA indexer torch.split behavior vs InternVL2 meta-init incompatibility; not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:45834", + "accept": false, + "reason": "Generation kwargs after sync_gpus and Kosmos2.5 OCR index error are unrelated." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "A wrong docstring is not the same as a RoPE implementation bug." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer-loading regression vs eos_token_id configuration inconsistency; different symptoms and fixes." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": true, + "reason": "Both describe v5 tokenizer-loading failures; 44991 looks like a model-specific instance of the broader regression in 43723." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat/completions request rejection and Gemma4Processor missing _tokenizer are unrelated server/processor failures." + }, + { + "left": "issue:44279", + "right": "issue:46328", + "accept": false, + "reason": "Dependency trouble vs a pad() performance issue; no shared underlying bug." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Same broad DSA area, but one is about split outputs and the other about a missing ReLU; distinct fixes." + }, + { + "left": "issue:42907", + "right": "issue:46050", + "accept": false, + "reason": "Both involve save/load of quantized models, but different quantization methods and failure points." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Completely different models and symptoms: loading hang vs zero probabilities." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "Remote-code meta-init incompatibility vs incorrect visual encoder keys on save_pretrained; different bugs." + }, + { + "left": "issue:43644", + "right": "issue:46305", + "accept": false, + "reason": "Non-persistent buffer junk in v5.0 vs a streaming window off-by-one; unrelated." + }, + { + "left": "issue:44829", + "right": "issue:46002", + "accept": false, + "reason": "Degenerate training with flash_attention_3 and NaNs on checkout/init are different classification-head issues." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "Missing rms_norm_eps in MLA q_a_layernorm and InternVL2 meta-init incompatibility are unrelated." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue vs Qwen3.5 MoE from_pretrained error; different code paths." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they concern different classes/symbols and do not look like one shared defect." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "Accuracy regression in FP8 vs Qwen3-VL-Moe bug; not the same problem." + }, + { + "left": "issue:43408", + "right": "issue:46402", + "accept": false, + "reason": "SAM3 model-type warning and SAM3 text-encoder weight loading are different issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet and is conservative. The only accepted soft edge, 43723/44991, is a credible same-bug/tokenizer-loading match, and the other rejections are appropriately cautious rather than overstated." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:43761", + "issue:43784", + "issue:44060", + "issue:44117", + "issue:44188", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44368", + "issue:44373", + "issue:44488", + "issue:44521", + "issue:44568", + "issue:44704", + "issue:44913", + "issue:44945", + "issue:45278", + "issue:45593", + "issue:45753", + "issue:46002", + "issue:46082", + "issue:46132", + "issue:46305", + "issue:46326", + "issue:46369" + ], + "soft_pairs": [ + "issue:43479|issue:44913", + "issue:43761|issue:46082", + "issue:44263|issue:44373", + "issue:39401|issue:44568", + "issue:22355|issue:27957", + "issue:42757|issue:45278", + "issue:27957|issue:30990", + "issue:42886|issue:43066", + "issue:42886|issue:42994", + "issue:22355|issue:28218", + "issue:44368|issue:46305", + "issue:42831|issue:45753", + "issue:43408|issue:46369", + "issue:43232|issue:45593", + "issue:43576|issue:43784", + "issue:44295|issue:44297", + "issue:43550|issue:46132", + "issue:43646|issue:46002", + "issue:44117|issue:44704", + "issue:42371|issue:44263", + "issue:43756|issue:44060", + "issue:30990|issue:33357", + "issue:44188|issue:44373", + "issue:42907|issue:44945", + "issue:43122|issue:44521", + "issue:39401|issue:44488", + "issue:41628|issue:45278", + "issue:43122|issue:46326", + "issue:44261|issue:44373" + ], + "prepared_review_unit_hash": "cb70cbd768e6da7372972bab4c5885772eac0cef6e22d7dad304a88aa5de77d8", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35977, + "estimated_input_tokens": 8995, + "estimated_eval_tokens": 18246 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The items are mostly unrelated standalone issues across tokenization, config, import, quantization, generation, and model-specific regressions. None of the proposed soft pairs look like the same underlying bug or change, so there is no strong duplicate cluster here.", + "confidence": 0.96, + "canonical_issue_reason": "issue:42886 is the clearest concrete regression to anchor triage because it has a specific failure mode, version, and repro; however, the overall set is heterogeneous rather than one duplicate family.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42886 is the best single representative issue because it is precise and reproducible (offline tokenizer cache loading regression), but it should not be treated as a canonical duplicate for the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Different bugs: multimodal config defaulting vs GPTNeoX config reload losing rotary_pct." + }, + { + "left": "issue:43761", + "right": "issue:46082", + "accept": false, + "reason": "Different code paths: CLIPVision hidden_states regression vs LlamaConfig head_dim validation." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "Runtime indexing/splitting bug vs a docstring wording issue; not the same bug." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer offset_mapping regression vs BOS/EOS special-token insertion failure." + }, + { + "left": "issue:22355", + "right": "issue:27957", + "accept": false, + "reason": "Missing transformers.onnx module import and XLMRoberta FlashAttention2 support are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import from huggingface_hub is not the same as broad v5 import breakages." + }, + { + "left": "issue:27957", + "right": "issue:30990", + "accept": false, + "reason": "FlashAttention2 model support issue vs sentence-transformers loading hang; different symptoms and causes." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache loading failure is unrelated to tokenizer decoder type selection in v5." + }, + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Tokenizer cache/offline bug vs quantized model saving failure." + }, + { + "left": "issue:22355", + "right": "issue:28218", + "accept": false, + "reason": "ONNX import failure and extra-space tokenization behavior are unrelated." + }, + { + "left": "issue:44368", + "right": "issue:46305", + "accept": false, + "reason": "A tied-weights warning is not the same as a streaming mask off-by-one bug." + }, + { + "left": "issue:42831", + "right": "issue:45753", + "accept": false, + "reason": "FineGrainedFP8 accuracy issue vs multimodal infinite-loop behavior; different failures." + }, + { + "left": "issue:43408", + "right": "issue:46369", + "accept": false, + "reason": "Model-type warning on instantiation vs memory overflow in a different model path." + }, + { + "left": "issue:43232", + "right": "issue:45593", + "accept": false, + "reason": "Generation kwargs after sync_gpus and D-FINE auxiliary-loss behavior are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken CLI env command vs import-time NameError in sentence-transformers integration." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "position_ids buffer read error vs tokenizer_class metadata mismatch on save_pretrained." + }, + { + "left": "issue:43550", + "right": "issue:46132", + "accept": false, + "reason": "torch.compile SDPA failure in Bamba and AttentionInterface.register behavior change do not share a concrete root cause." + }, + { + "left": "issue:43646", + "right": "issue:46002", + "accept": false, + "reason": "Custom model initialization breakage in v5 vs NaNs in classification heads are different regressions." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "TOKENIZER_MAPPING_NAMES returning None and AutoProcessor kwargs forwarding are separate loader bugs." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 settings warning/behavior is unrelated to torch.split output handling." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch vs Qwen3-Next tied-weights warning; not the same issue." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading hang and MacOS bus error are different failure modes with no shared concrete bug." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Torch.compile attention-kernel divergence is unrelated to a docstring typo." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Quantized/dequantized save failure vs pipeline parallelism output correctness are different problems." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Tokenization drift across versions vs multimodal assistant_masks all-zero bug are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Qwen3 offset-mapping bug and failure to load cjvt/sleng-bert are not the same underlying issue." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Single missing AutoImageProcessor import vs broad post-upgrade import errors are too different to merge." + }, + { + "left": "issue:43761", + "right": "issue:44373", + "accept": false, + "reason": "A CLIPVision forward regression is not a docstring issue." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "MLA q_a_layernorm precision/config bug is unrelated to a docstring mistake." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 44, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "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:38617", + "issue:39401", + "issue:42175", + "issue:42673", + "issue:42757", + "issue:42907", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43493", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44222", + "issue:44295", + "issue:44297", + "issue:44351", + "issue:44360", + "issue:44485", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44779", + "issue:44829", + "issue:44938", + "issue:44945", + "issue:44964", + "issue:45072", + "issue:45200", + "issue:45216", + "issue:45278", + "issue:45290", + "issue:45478", + "issue:45600", + "issue:45753" + ], + "soft_pairs": [ + "issue:42907|issue:43937", + "issue:43784|issue:45600", + "issue:43976|issue:44945", + "issue:44704|issue:45290", + "issue:42907|issue:44964", + "issue:39401|issue:44521", + "issue:44038|issue:44297", + "issue:43116|issue:45200", + "issue:44360|issue:44485", + "issue:44625|issue:44829", + "issue:44038|issue:44295", + "issue:43122|issue:44514", + "issue:43408|issue:43475", + "issue:43122|issue:44779", + "issue:44038|issue:44222", + "issue:42907|issue:43452", + "issue:45216|issue:45478", + "issue:43824|issue:44351", + "issue:22355|issue:29159", + "issue:42757|issue:44351", + "issue:36296|issue:43116", + "issue:43278|issue:44945", + "issue:42175|issue:44938", + "issue:43493|issue:45753", + "issue:43066|issue:44779", + "issue:22355|issue:29127", + "issue:42673|issue:45072", + "issue:43992|issue:44704", + "issue:38617|issue:45278", + "issue:43937|issue:44485" + ], + "prepared_review_unit_hash": "ad0c029c9c7a4ca920560f3f8e3fdf856e028e09a99df3a46c074b02b5160d96", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 35759, + "estimated_input_tokens": 8940, + "estimated_eval_tokens": 18136 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All soft pairs are false positives; these items span unrelated bugs across imports, tokenizers, model loading, chat templates, and training, not one duplicate cluster.", + "confidence": 0.97, + "canonical_issue_reason": "No canonical issue: the cluster is too heterogeneous to map to one underlying bug.", + "canonical_pr_reason": "No PRs present in the cluster.", + "best_issue_reason": "No issue is a strong global representative because the accepted-overlap signal is absent and every soft edge is a mismatch.", + "best_pr_reason": "No PRs present.", + "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Dequantized save failure for Ministrals/Devstrals vs GLM-5 generation config validation; different models and code paths." + }, + { + "left": "issue:43784", + "right": "issue:45600", + "accept": false, + "reason": "sentence-transformers import NameError vs Sam3LiteText auto-mappings CI breakage; unrelated import/mapping bugs." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility issue vs incorrect output under pipeline parallelism; different failure modes." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor kwarg forwarding to cached_file vs chat-template crash on tool-call assistant messages; different subsystems." + }, + { + "left": "issue:42907", + "right": "issue:44964", + "accept": false, + "reason": "Save/dequantization bug vs inability to load Phi-4 multimodal model; different model families and paths." + }, + { + "left": "issue:39401", + "right": "issue:44521", + "accept": false, + "reason": "Qwen3 offset_mapping regression vs multimodal assistant_mask generation; tokenization offsets are not the same bug as chat-template masking." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-Moe from_pretrained error vs tokenizer_config tokenizer_class mismatch on save; different load/save issues." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example script multi-label classification returning empty results vs Gemma 4 mm_token_type_ids defaulting; unrelated example vs model-input bug." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer/ReLU discussion vs GLM-5 RoPE implementation; entirely different technical areas." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "Qwen3.5 num_labels propagation vs flash_attention_3 degenerate training; config plumbing and attention-kernel bugs are distinct." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Qwen3-VL-Moe load failure vs position_ids buffer access error; not the same underlying bug." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer regression across versions vs batched apply_chat_template crash with padding=False; different code paths." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "SAM3 model-type warning vs missing fpn_position_embeddings attribute in vision encoder output; related family, but not one bug." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Generic tokenization differences in v5 vs Deepseek-specific incorrect tokenization; too broad to be the same defect." + }, + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Qwen3-VL-Moe from_pretrained error vs FP8 save_pretrained issue; load-time and save-time bugs differ." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Dequantized saving bug vs gguf_file loading breakage for AutoTokenizer/AutoModelForCausalLM; different serialization paths." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3.5 save_pretrained checkpoint regression vs Qwen3.5 Moe from_pretrained error; same model family, but different save/load failures." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Missing Qwen2_5_VLForConditionalGeneration export vs missing HybridCache export; both import errors, but for different symbols and modules." + }, + { + "left": "issue:22355", + "right": "issue:29159", + "accept": false, + "reason": "Missing transformers.onnx module vs slow/fast tokenizer inconsistency; import failure and tokenizer behavior bug are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "huggingface_hub missing is_offline_mode vs transformers missing HybridCache; different packages and missing symbols." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor-parallel training bug vs example script multi-label classification issue; distinct workloads and symptoms." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "BF16-to-FP32 embedding dtype change in evaluate vs incorrect pipeline-parallel outputs; different bugs." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Missing tensorflow in extras install vs Python 3.14 load failure; packaging/dependency issue versus interpreter compatibility." + }, + { + "left": "issue:43493", + "right": "issue:45753", + "accept": false, + "reason": "SigLIP2 HF vs JAX discrepancy vs Qwen3_5 infinite loop on a specific image; unrelated model families." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 vs Deepseek tokenizer incorrect results; both tokenizer regressions, but not the same concrete defect." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "onnx import failure vs LayoutLMv3 error-message clarity; unrelated import and UX bugs." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM with multiple dataloader threads vs dtype mismatch CI failure; different mechanisms and symptoms." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder load vs AutoProcessor kwarg forwarding; unrelated model loading vs processor plumbing." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "Specific missing layer_type_validation import vs broad 'many import errors' report; 45278 is an umbrella symptom, not the same bug." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GLM-5 generation config invalid vs GLM-5 RoPE implementation; same model family but different concrete 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": 21, + "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:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:43064", + "issue:43116", + "issue:43122", + "issue:43441", + "issue:43450", + "issue:43493", + "issue:43611", + "issue:43673", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43906", + "issue:43937", + "issue:43976", + "issue:43986", + "issue:44038", + "issue:44075", + "issue:44117", + "issue:44188", + "issue:44291", + "issue:44355", + "issue:44380", + "issue:44410", + "issue:44466", + "issue:44779", + "issue:44936", + "issue:44977", + "issue:45071", + "issue:45081", + "issue:45083", + "issue:45310", + "issue:45357", + "issue:45405", + "issue:45446", + "issue:45478", + "issue:45563", + "issue:45588", + "issue:46002", + "issue:46077", + "issue:46231", + "issue:46298", + "issue:46328" + ], + "soft_pairs": [ + "issue:44188|issue:46328", + "issue:43611|issue:46002", + "issue:44355|issue:46231", + "issue:43116|issue:43493", + "issue:42371|issue:43874", + "issue:42222|issue:43441", + "issue:44038|issue:45478", + "issue:43976|issue:45446", + "issue:45357|issue:45478", + "issue:44075|issue:44936", + "issue:43116|issue:43825", + "issue:45083|issue:45446", + "issue:43064|issue:43673", + "issue:33453|issue:39401", + "issue:39401|issue:44779", + "issue:43937|issue:45563", + "issue:44117|issue:45588", + "issue:44380|issue:46077", + "issue:42673|issue:43906", + "issue:43827|issue:45446", + "issue:43450|issue:43986", + "issue:43673|issue:46328", + "issue:44466|issue:46002", + "issue:44291|issue:45071", + "issue:43937|issue:44977", + "issue:45405|issue:45446", + "issue:44410|issue:45310", + "issue:43122|issue:46298", + "issue:44117|issue:45081" + ], + "prepared_review_unit_hash": "56e71963df02f4a9f7ffd4bc90e52bbc33c45d97a06a153b736d75b5eee26cde", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 29, + "serialized_chars": 35703, + "estimated_input_tokens": 8926, + "estimated_eval_tokens": 18108 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous Transformers v5 regression reports across unrelated subsystems. The only especially close pair is the two Qwen3.5 MoE `from_pretrained` reports, which look duplicate-like from the titles alone, but the metadata still does not prove a shared root cause. Issue:43122 is a reasonable representative tokenizer-regression issue, but the cluster as a whole is too mixed to treat as one bug.", + "confidence": 0.55, + "canonical_issue_reason": "issue:43122 is a reasonable representative tokenizer-regression report: its title clearly points to a version-dependent tokenization change, and it sits within the tokenizer-related subset of the cluster. I\u2019m not relying on unseen repro/body details.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43122 is the best representative issue to keep because it is one of the clearest tokenizer-regression reports in the set and serves as a sensible anchor for that small subgroup, even though the overall cluster remains mixed.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44188", + "right": "issue:46328", + "accept": false, + "reason": "Attention-kernel divergence and padding-performance regression are different bugs in different paths." + }, + { + "left": "issue:43611", + "right": "issue:46002", + "accept": false, + "reason": "Model-loading `base_model_prefix` failure and NaNs in classification heads are distinct symptoms and likely distinct causes." + }, + { + "left": "issue:44355", + "right": "issue:46231", + "accept": false, + "reason": "Compiled Python file errors and a docs parameter-name mismatch are unrelated." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Empty multi-label classification outputs and a SigLIP2 implementation discrepancy are different issues." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 API guidance is unrelated to a missing image-processor method." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "VitPose model breakage and Ministral-3 FlashAttention failure are different code paths." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Very similar Qwen3.5 MoE `from_pretrained` failures with identical-looking titles, but metadata alone does not prove the same underlying bug." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python-version compatibility and a PyTorch version-check bug are different problem classes." + }, + { + "left": "issue:45357", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is a `save_pretrained` serialization issue and the other is a `from_pretrained` load failure." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Ignored SGD args and `trainer.evaluate()` failing after `trainer.train()` are separate training bugs." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Classification example output bug and pipeline translation messaging/docs issue are unrelated." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "Qwen3 omni feature-length helper and flex_attention version checking are different subsystems." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption and GenerationMixin cache handling are distinct issues." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Both tokenizer-related, but the loading regression and the Qwen3 offset-mapping bug are not the same concrete failure." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Offset-mapping correctness and Deepseek tokenization output changes are different tokenizer bugs." + }, + { + "left": "issue:43937", + "right": "issue:45563", + "accept": false, + "reason": "GenerationConfig validation and a stale warning about `num_return_sequences` are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45588", + "accept": false, + "reason": "Tokenizer mapping lookup handling and a flash-attention `s_aux=None` crash are not the same bug." + }, + { + "left": "issue:44380", + "right": "issue:46077", + "accept": false, + "reason": "Different attention-mask/scaling problems in different models and backends." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "VRAM leakage across dataloader threads and an isolated reproduction issue are only loosely related at best." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing removed pipelines and an import/version check bug are different issues." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video-processor shape mismatch and missing `torchvision` crash are distinct failure modes." + }, + { + "left": "issue:43673", + "right": "issue:46328", + "accept": false, + "reason": "Generation cache missing and feature-extractor padding performance are unrelated subsystems." + }, + { + "left": "issue:44466", + "right": "issue:46002", + "accept": false, + "reason": "Tied-weight serialization inconsistency and NaNs after checkout/init do not appear to share a root cause." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` argument handling and `PretrainedConfig` type checking are different bugs." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "Invalid generation config and Qwen3.5 flash-attention generation failure are unrelated." + }, + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "A packaging/version pinning issue is not the same as a PyTorch version check bug." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Different Qwen-family load problems with different symptoms and likely different code paths." + }, + { + "left": "issue:43122", + "right": "issue:46298", + "accept": false, + "reason": "Tokenizer regression and noisy logging are unrelated." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "Tokenizer mapping-name handling and the Mistral regex patch crash are different tokenizer bugs." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The clustering is broadly plausible, but a few claims are too strong for the evidence available in the packet. In particular, calling issue:44038 and issue:45478 the \"same underlying bug\" is not well-grounded from the metadata alone; they are clearly similar, but the packet does not prove duplication. Also, the canonical/representative reasoning for issue:43122 leans on reproduction/body details that are not present in the packet. Please soften those verdicts and keep the summary more conservative, e.g. \"most likely duplicate-like pair\" and \"reasonable representative tokenizer-regression issue.\"" + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:44263", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44877", + "issue:44933", + "issue:45071", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45362", + "issue:45464", + "issue:45478", + "issue:45542", + "issue:45561", + "issue:45656", + "issue:46178", + "issue:46231", + "issue:46328" + ], + "soft_pairs": [ + "issue:45341|issue:45561", + "issue:38617|issue:41628", + "issue:45341|issue:46328", + "issue:44829|issue:45216", + "issue:42222|issue:46178", + "issue:42617|issue:45341", + "issue:45362|issue:46328", + "issue:43408|issue:43742", + "issue:42673|issue:43582", + "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:42222|issue:45542", + "issue:44410|issue:45478", + "issue:42371|issue:42831", + "issue:44075|issue:45656", + "issue:43012|issue:43577", + "issue:44483|issue:45464", + "issue:42175|issue:42222", + "issue:43097|issue:44493", + "issue:43673|issue:45341", + "issue:41628|issue:45071", + "issue:44263|issue:44485" + ], + "prepared_review_unit_hash": "52cf101f77fbda400c12258bd6a023b52c547964b1e89a230d4660bfb266307d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35945, + "estimated_input_tokens": 8987, + "estimated_eval_tokens": 18230 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: it mixes unrelated import errors, tokenizer regressions, backend/config issues, model-specific failures, and performance/behavior bugs. None of the soft pairs looks like the same underlying issue with a plausible single fix, so there is no strong canonical duplicate here.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue: the items are not one bug family and the closest-looking pairs still point to different code paths or symptoms.", + "canonical_pr_reason": "No pull requests were provided in the cluster.", + "best_issue_reason": "No best representative issue: this is not a tight duplicate cluster, so any one issue would misrepresent the rest.", + "best_pr_reason": "No pull requests were provided in the cluster.", + "soft_edge_verdicts": [ + { + "left": "issue:45341", + "right": "issue:45561", + "accept": false, + "reason": "Both mention testing_utils, but one is a vague \"little bug\" and the other is a specific xdist race on captured_info.txt; not the same concrete bug." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they involve different symbols and code paths (`layer_type_validation` vs `AutoImageProcessor`)." + }, + { + "left": "issue:45341", + "right": "issue:46328", + "accept": false, + "reason": "One is a testing_utils bug and the other is a SequenceFeatureExtractor.pad performance issue; unrelated functionality." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "Flash-attention training degeneration and Qwen3.5 save_pretrained checkpoint corruption are different regressions." + }, + { + "left": "issue:42222", + "right": "issue:46178", + "accept": false, + "reason": "VitPose breakage and a tensorflow_text backend mapping error are unrelated failures." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "3d_parallel.py execution failure is unrelated to a generic testing_utils bug." + }, + { + "left": "issue:45362", + "right": "issue:46328", + "accept": false, + "reason": "Qwen3.5 chat crashes and feature-extractor padding inefficiency are different code paths and symptoms." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "Model-type warning for sam3_tracker vs KeyError loading MobileLLM are unrelated model-loading issues." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leak in multi-threaded Qwen3.5 inference is not the same as an Apple Silicon TypeError in caching_allocator_warmup." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Offset mapping bug and runtime post-processor update bug are both tokenizer-related but distinct underlying problems." + }, + { + "left": "issue:44877", + "right": "issue:46231", + "accept": false, + "reason": "Strict config loading for granite_speech and a docs parameter-name mismatch are unrelated." + }, + { + "left": "issue:42490", + "right": "issue:42673", + "accept": false, + "reason": "Local-path trailing-slash import warning is unrelated to the Qwen3.5 VRAM leak." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 API warning and MLA rms_norm_eps precision issue do not share the same code path." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "A save_pretrained shard-size question and a Qwen3.5 MoE from_pretrained error are not the same bug." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "ImportError for configuration_utils and `_torch_extract_fbank_features()` are unrelated issues." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GenerationConfig validation for GLM-5 and an unrecognized model/config.json `model_type` error are different problems." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A missing import from image_utils and broad import errors after upgrade are not specific enough to treat as the same bug." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": false, + "reason": "Tokenizer decoder type issue and tokenization output differences are related to tokenizers generally but not the same concrete defect." + }, + { + "left": "issue:42222", + "right": "issue:45542", + "accept": false, + "reason": "VitPose model breakage and a tensorboard-without-TF backend error are unrelated package/backend issues." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Missing layer projections in qwen3next and Qwen3.5 MoE from_pretrained failure are different model-specific issues." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 behavior API warning and FineGrainedFP8 accuracy regression are not the same underlying bug." + }, + { + "left": "issue:44075", + "right": "issue:45656", + "accept": false, + "reason": "SGD args not used and optimizer.step called twice with deepspeed are distinct optimizer bugs." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "A compile-time float32 warning for bf16 models is different from model.dtype staying float32 after loading Blip2." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both touch chat/completions, but one is a general request-acceptance failure and the other is streaming inference on Qwen3.5-0.8B; not the same concrete issue." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow missing from a torch install and broken vitpose models are unrelated." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder and unexpected position-id keys are different regression classes." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing in chunked_prefill and a testing_utils bug are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "ImportError for AutoImageProcessor and PreTrainedConfig type-checking breakage are different API regressions." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "torch.split return-value handling and GLM-5 RoPE implementation 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": 21, + "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:43874|issue:44261", + "issue:45071|issue:45542", + "issue:27957|issue:43352", + "issue:44484|issue:45478", + "issue:44263|issue:46172", + "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:45003|issue:45341", + "issue:43986|issue:45341", + "issue:44933|issue:46231", + "issue:43673|issue:46123", + "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:43502|issue:43992", + "issue:44062|issue:44117", + "issue:44410|issue:45375", + "issue:44038|issue:44410", + "issue:42890|issue:46298", + "issue:38617|issue:44351", + "issue:43097|issue:43873" + ], + "prepared_review_unit_hash": "9e425f4e53b74e13853e24ee46aadf00204bf2d389b49431f7778b16b98d6f09", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37659, + "estimated_input_tokens": 9415, + "estimated_eval_tokens": 19086 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is very heterogeneous: most items are distinct model-specific regressions, import errors, or unrelated docs/tests issues. I did not find any soft pair that looks like the same underlying bug/change, so all soft edges are rejected.", + "confidence": 0.86, + "canonical_issue_reason": "issue:43097 is the best anchor because it\u2019s a broad v5 release regression with downstream impact, whereas most other items are narrower model- or subsystem-specific failures.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43097 is the most representative and highest-level issue in the set: a release-wide API removal/regression that plausibly sits near the center of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "Different bugs in different multimodal models: missing image-patch helper vs missing rms_norm_eps in a layernorm config." + }, + { + "left": "issue:45071", + "right": "issue:45542", + "accept": false, + "reason": "Unrelated failures: PretrainedConfig type-checking regression vs TensorFlow backend import handling." + }, + { + "left": "issue:27957", + "right": "issue:43352", + "accept": false, + "reason": "Both mention Flash Attention 2, but they concern different models and different support paths." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "save_pretrained shard-size question is unrelated to Qwen3.5 MoE loading errors." + }, + { + "left": "issue:44263", + "right": "issue:46172", + "accept": false, + "reason": "Different code paths: GlmMoeDsaIndexer torch.split behavior vs deepseek-v4 dtype/transpose logic." + }, + { + "left": "issue:44360", + "right": "issue:46172", + "accept": false, + "reason": "DSA indexer ReLU omission is unrelated to deepseek-v4 comb transpose behavior." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Both are Qwen-family, but one is a weight-shape load mismatch and the other is missing attention projections." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both involve flash attention, but the models, versions, and failure modes are different enough to be separate bugs." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Different Qwen-VL/MoE loading problems with no clear shared underlying defect." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Tie-embeddings removal is a release API regression; gguf_file loading is a separate tokenizer/model-loading issue." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer misconfiguration is unrelated to config type-checking failures." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "sys.modules access bug and a testing utility bug are different code paths and symptoms." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "Video-model loading without torchvision is unrelated to the testing_utils bug." + }, + { + "left": "issue:44933", + "right": "issue:46231", + "accept": false, + "reason": "Missing image_utils import and a docs parameter-name mismatch are not the same issue." + }, + { + "left": "issue:43673", + "right": "issue:46123", + "accept": false, + "reason": "Cache handling in GenerationMixin and final-batch pipeline bookkeeping are separate bugs." + }, + { + "left": "issue:45753", + "right": "issue:46369", + "accept": false, + "reason": "Infinite-loop generation on one image is unrelated to memory overflow in a different model." + }, + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Tokenizer mapping returning None and a specific model load failure are not clearly the same defect." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "No meaningful overlap: torchvision-dependent processor crash vs shard-size behavior in save_pretrained." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Different model internals: missing image patch count method vs torch.split return-value handling." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding-window bug is unrelated to a testing utility issue." + }, + { + "left": "issue:45440", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeekV3 MoE divergence and deepseek-v4 comb transpose are different implementation problems." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "No torchvision crash and missing ReasoningEffort import are unrelated import/dependency failures." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "Precision/epsilon issue in MLA q_a_layernorm is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:43097", + "right": "issue:45588", + "accept": false, + "reason": "Tie-embeddings API removal is unrelated to flash_attention.py crashing on s_aux=None." + }, + { + "left": "issue:43097", + "right": "issue:46050", + "accept": false, + "reason": "API removal/regression is distinct from quantized model save/load corruption." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2+PEFT optimizer-state corruption and Liger application with model_init are different training paths." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are import errors, but they involve different symbols from different packages and likely different causes." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "local_files_only network leakage is unrelated to missing embed_tokens.weight on UMT5Encoder load." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "AddedToken keyword collision and tokenizer mapping returning None are separate tokenizer bugs." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Missing qwen3next projections and missing deepstack_visual_indexes are different config/model-definition issues." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Qwen3-VL-Moe from_pretrained failure is not the same as qwen3next layer-0 projection omission." + }, + { + "left": "issue:42890", + "right": "issue:46298", + "accept": false, + "reason": "Flaky integration test from missing set_seed and noisy logging are unrelated." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports from different modules; no shared underlying bug is evident." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Tie-embeddings API removal is unrelated to offloading/quantization behavior." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "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:44368", + "issue:43012|issue:43550", + "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": "59431ae160898bac43f0c230dbe02f076158a241ffcc45b9a9ee42e4fc27c224", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37444, + "estimated_input_tokens": 9361, + "estimated_eval_tokens": 18978 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft links are all false positives; this is a heterogeneous set of unrelated import, training, model, backend, and warning issues rather than one duplicate cluster.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue: the items do not share one underlying bug or code path closely enough to consolidate.", + "canonical_pr_reason": null, + "best_issue_reason": "No strong global representative exists because the issues span unrelated failures across different subsystems and model families.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44936", + "right": "issue:46365", + "accept": false, + "reason": "Different failures: evaluation breaks after training vs import-time Trainer conflict with datasets." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "Different code paths: cached_file kwargs forwarding vs auto-docstring annotation introspection." + }, + { + "left": "issue:44380", + "right": "issue:46421", + "accept": false, + "reason": "Attention scaling under SDPA/FlashAttention is unrelated to cache lazy initialization recompilation." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Missing torchvision for video AutoProcessor is unrelated to TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:43824", + "right": "issue:43986", + "accept": false, + "reason": "Both are import/loading problems, but one is a missing exported class and the other is a missing optional dependency." + }, + { + "left": "issue:44938", + "right": "issue:46178", + "accept": false, + "reason": "Python 3.14 load failure is unrelated to a bad BACKENDS_MAPPING backend name." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "Wrong tokenizer decoder type and an isolated reproduction of another issue are not the same bug." + }, + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "xLSTM training blockers and Bamba torch.compile+SDPA failure affect different models and paths." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing embed_tokens.weight on load is unrelated to max_shard_size behavior in save_pretrained." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "A bfloat16 precision warning is not the same as the tie_word_embeddings warning." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline task messaging in v5 is unrelated to saving Pipeline objects." + }, + { + "left": "issue:45712", + "right": "issue:46298", + "accept": false, + "reason": "Dummy PT object leakage is unrelated to a noisy logging message." + }, + { + "left": "issue:42907", + "right": "issue:45753", + "accept": false, + "reason": "Saving dequantized models and a Qwen3_5 image loop are different bugs." + }, + { + "left": "issue:43097", + "right": "issue:46298", + "accept": false, + "reason": "Config API removal is unrelated to noisy logging." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "Nemotron FlashAttention 2 support gap is not the same as GPT2 attention scaling ignored under backends." + }, + { + "left": "issue:43976", + "right": "issue:43352", + "accept": false, + "reason": "Python version incompatibility is unrelated to a model-specific FlashAttention support issue." + }, + { + "left": "issue:44162", + "right": "issue:46172", + "accept": false, + "reason": "ESM2 breakage and deepseek-v4 transpose behavior are different failures." + }, + { + "left": "issue:45584", + "right": "issue:46365", + "accept": false, + "reason": "Whisper empty-transcription generation failure is unrelated to importing Trainer with datasets installed." + }, + { + "left": "issue:44077", + "right": "issue:46365", + "accept": false, + "reason": "patchtsmixer post_init validation has nothing to do with the Trainer import conflict." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "Attention backend scaling bug and zero-loss classification bug are different code paths." + }, + { + "left": "issue:44380", + "right": "issue:46132", + "accept": false, + "reason": "Generic AttentionInterface register behavior change is not the same as GPT2 attention scaling being ignored." + }, + { + "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 on MoE models and Qwen3.5 MoE from_pretrained errors are not the same concrete bug." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Chat completions API rejection is unrelated to Qwen2_5_VLProcessor batched padding crash." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Missing Gemma3Config vocab_size and Apple Silicon allocator TypeError are unrelated." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training blockers are unrelated to BeitImageProcessorFast reduce_labels behavior." + }, + { + "left": "issue:42175", + "right": "issue:46365", + "accept": false, + "reason": "Backend packaging for pip install '.[torch]' is unrelated to a Trainer import conflict with datasets." + }, + { + "left": "issue:42490", + "right": "issue:43012", + "accept": false, + "reason": "Local-path trailing-slash warning and bfloat16 precision warning are different issues." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "A precision warning is not the same as a torch>=2.9 flex_attention failure." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Different missing imports: is_offline_mode from huggingface_hub vs Qwen2_5_VLForConditionalGeneration from transformers." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Importing layer_type_validation and making API requests despite local_files_only are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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: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: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:42994|issue:44655", + "issue:36683|issue:43475", + "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:43352|issue:43937", + "issue:36683|issue:44077", + "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:44625", + "issue:43992|issue:45310", + "issue:43064|issue:44492", + "issue:44075|issue:46365", + "issue:43010|issue:45278", + "issue:44355|issue:44855", + "issue:43867|issue:46202" + ], + "prepared_review_unit_hash": "761607b7b3c05db5d856b202fe79e4c2cdf02fadbda4a393101a7b8da02c386c", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36740, + "estimated_input_tokens": 9185, + "estimated_eval_tokens": 18626 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose collection of Transformers issues around v5 regressions, import/load failures, model-specific config bugs, and docs/behavior mismatches. I don\u2019t see any soft pair that clearly represents the same underlying bug or change.", + "confidence": 0.39, + "canonical_issue_reason": "issue:45278 is the broadest umbrella-style report in the set ('many import errors after update from 4.57.0 to 5.5.0'), so it is the best representative issue if one must be chosen.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the most general, high-signal issue here and best fits the overall v5 import/regression theme.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45600", + "right": "issue:46298", + "accept": false, + "reason": "One is a CI break from stale Sam3LiteText auto-mappings; the other is a noisy logging message. Different bug classes." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Both mention Qwen3-VL/Moe, but one is a loading bug and the other is fp8 inference support; not the same concrete defect." + }, + { + "left": "issue:45698", + "right": "issue:46172", + "accept": false, + "reason": "Different problems: custom-module reload after save_pretrained versus a deepseek-v4 tensor transpose issue." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIP attentions missing versus GPT2 attention scaling ignored under SDPA/FlashAttention are separate code paths and symptoms." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both are CLIP regressions, but one concerns attentions and the other hidden_states; distinct missing outputs." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "An API request acceptance regression is unrelated to chat template crashes on tool-call messages." + }, + { + "left": "issue:44625", + "right": "issue:46365", + "accept": false, + "reason": "Qwen3.5 num_labels propagation is unrelated to a Trainer import conflict with datasets." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3VL-Moe loading breakage is not the same as a generic state_dict-sorting load error." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Quantized model saving failure and pipeline save_pretrained failure are both save-related, but they are different failures." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Gemma3Config vocab_size missing and SAM 3 Video missing fpn_position_embeddings are unrelated model-specific attribute errors." + }, + { + "left": "issue:43066", + "right": "issue:46209", + "accept": false, + "reason": "Tokenizer decoder type issue and FP8 inference on B200 are unrelated regressions." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Smollm3 RoPE-layer mismatch and GLM-5 RoPE implementation are different model/configuration issues." + }, + { + "left": "issue:45276", + "right": "issue:46077", + "accept": false, + "reason": "Gemma4 token embedding resize bug and bidirectional attention mask bug affect different mechanisms." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Docs mismatch for return_all_scores is not the same as pipeline serialization failure." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Different import errors from different dependencies: transformers.generation vs huggingface_hub." + }, + { + "left": "issue:43611", + "right": "issue:44655", + "accept": false, + "reason": "base_model_prefix load failure and pipeline save_pretrained failure are distinct serialization/load paths." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Nemotron FlashAttention 2.0 support and GLM-5 GenerationConfig validation are unrelated model-specific bugs." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Both are attribute/config issues, but Gemma3Config vocab_size and patchtsmixer post_init allowance are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "A generic load error with sorted state_dict is not the same as Qwen3.5 checkpoint corruption on save_pretrained." + }, + { + "left": "issue:44625", + "right": "issue:46077", + "accept": false, + "reason": "Qwen3.5 num_labels propagation and Gemma4 attention-mask behavior are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Python version compatibility for 5.1.0 and Python 3.14 load failures are different environment/version regressions." + }, + { + "left": "issue:46077", + "right": "issue:46082", + "accept": false, + "reason": "Gemma4 attention mask bug and LlamaConfig head_dim validation are separate config behaviors." + }, + { + "left": "issue:43352", + "right": "issue:45593", + "accept": false, + "reason": "FlashAttention support limitation for NemotronH and D-FINE auxiliary-loss handling are unrelated." + }, + { + "left": "issue:36683", + "right": "issue:43525", + "accept": false, + "reason": "Gemma3Config lacks vocab_size, while Llama4Config lacks pad_token_id; similar shape, but different models and fixes." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "Missing CLIP attentions and Qwen3.5 checkpoint save regression are unrelated issues." + }, + { + "left": "issue:36683", + "right": "issue:44625", + "accept": false, + "reason": "Both are config propagation issues, but for different models and different fields." + }, + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Missing embed_tokens.weight in UMT5Encoder is not the same as Qwen3.5 MoE from_pretrained loading error." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Distributed optimizer-state corruption and a cache-strategy typo are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:46365", + "accept": false, + "reason": "SGD args not used in optimizer code and Trainer import conflict with datasets are different bugs." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Cache/layer update decoration and broad import errors after upgrade are different issue areas." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Compiled Python file loading errors and Python 3.13 torch.jit script parsing errors are different failure modes." + }, + { + "left": "issue:43867", + "right": "issue:46202", + "accept": false, + "reason": "Generic sorted-state_dict load error and Gemma pipeline roundtrip failure are not the same underlying bug." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "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:30990", + "issue:31795", + "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:46285" + ], + "soft_pairs": [ + "issue:43867|issue:45375", + "issue:44938|issue:45071", + "issue:41669|issue:44246", + "issue:36683|issue:43064", + "issue:42890|issue:45561", + "issue:43010|issue:45850", + "issue:43992|issue:45478", + "issue:43867|issue:45357", + "issue:43618|issue:44077", + "issue:43010|issue:43976", + "issue:36296|issue:43575", + "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:45600", + "issue:42890|issue:44112", + "issue:43723|issue:44556", + "issue:44387|issue:44556", + "issue:44222|issue:45125", + "issue:43010|issue:43065", + "issue:43576|issue:46132", + "issue:42548|issue:44351", + "issue:43502|issue:44485", + "issue:45125|issue:45478", + "issue:44355|issue:44749", + "issue:30333|issue:33453", + "issue:45125|issue:45375", + "issue:43010|issue:43232", + "issue:43502|issue:44222", + "issue:30990|issue:31795" + ], + "prepared_review_unit_hash": "020e1b4ae0f443f97ad3bb476c062223778daeed01725a7dff62a4d41ceccc0d", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37116, + "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 mostly heterogeneous: many items are about unrelated loading, model-specific, CI, performance, or API-regression topics. I found no soft pair that looks like the same underlying bug/change with a plausible single-PR merge path, so all candidate edges are rejected.", + "confidence": 0.77, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43502 is the clearest standalone bug report in the set (offline loading still making API calls), but it is not a duplicate center for the rest of the items.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "Different failures: a general state_dict loading error vs a Qwen3.5 vision config field being dropped." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Both are v5-related breakages, but one is Python 3.14 import failure and the other is PretrainedConfig type-checking; not the same bug." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Both mention import slowdown, but 41669 is a specific import-* cleanup proposal while 44246 is an intermittent import performance complaint." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Unrelated: Gemma3Config missing vocab_size versus FSDP2/PEFT optimizer-state corruption." + }, + { + "left": "issue:42890", + "right": "issue:45561", + "accept": false, + "reason": "Test flakiness from missing set_seed is unrelated to xdist races on captured_info.txt." + }, + { + "left": "issue:43010", + "right": "issue:45850", + "accept": false, + "reason": "A no_grad decorator change for cache updates is not the same as a broad 'too many models break' complaint." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "Both are from_pretrained problems, but they involve different models and different missing/error paths." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "State-dict loading error and incorrect save_pretrained visual-encoder keys are different directions of failure." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to patchtsmixer post_init validation rules." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "No-grad decoration for cache updates is unrelated to a Python-version compatibility issue." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor-parallel training bug and tensor-parallel OOM on a specific model are different concrete problems." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch-missing import error is unrelated to tokenizer loading regression." + }, + { + "left": "issue:43618", + "right": "issue:46285", + "accept": false, + "reason": "Different CLIP regressions: attentions field assignment vs model structure inconsistency." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "CLIPOutput attentions and float16 CUDA crash in LwDetrImageLoss are unrelated bugs." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "A DSA indexer ReLU discussion is unrelated to DeepseekV3 implementation divergence." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "Offline API calls and AddedToken special-argument duplication are unrelated loading/tokenization bugs." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "Seed-related test flakiness is unrelated to Phi-4 multimodal model loading failure." + }, + { + "left": "issue:42890", + "right": "issue:45600", + "accept": false, + "reason": "SamHQ test seeding issue is unrelated to auto_mappings referencing removed Sam3LiteText configs." + }, + { + "left": "issue:42890", + "right": "issue:44112", + "accept": false, + "reason": "One is a seed-driven integration-test flake; the other is a stale GraniteSpeech device-override test." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer loading regression and checkpoint reload incompatibility across versions are related only at a broad level, not the same bug." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 quantization memory regression is different from checkpoint reload failure after upgrade." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained on MoE is not the same as missing _tp_plan for tensor parallelism." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache no_grad decoration and Sam3PixelDecoder dummy Conv2d are unrelated code changes." + }, + { + "left": "issue:43576", + "right": "issue:46132", + "accept": false, + "reason": "Broken env command in v5 and AttentionInterface.register behavior change are separate CLI/API issues." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "Cannot import PreTrainedModel and cannot import HybridCache are different export/import regressions." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "Offline loading making API calls is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Missing _tp_plan and Qwen3.5 from_pretrained error are model-specific but not the same concrete failure." + }, + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Compiled Python file errors are unrelated to a post-upgrade data-processing slowdown." + }, + { + "left": "issue:30333", + "right": "issue:33453", + "accept": false, + "reason": "MLFlow reporting on failed training and tokenizer loading regression do not share the same underlying bug." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Tensor-parallel plan missing is unrelated to a missing deepstack_visual_indexes field in vision config." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache update no_grad decoration is not the same as _update_model_kwargs_for_generation after sync_gpus." + }, + { + "left": "issue:43502", + "right": "issue:44222", + "accept": false, + "reason": "Local-files-only API leakage and FP8 save_pretrained behavior are different loading vs saving issues." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Sentence Transformers load hanging and confusing forward docs are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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:31795", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:43299", + "issue:43502", + "issue:43526", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:43901", + "issue:44038", + "issue:44222", + "issue:44230", + "issue:44273", + "issue:44315", + "issue:44355", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44908", + "issue:44938", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45125", + "issue:45230", + "issue:45362", + "issue:45468", + "issue:45542", + "issue:45850", + "issue:46095", + "issue:46143", + "issue:46144", + "issue:46209" + ], + "soft_pairs": [ + "issue: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:33453|issue:34689", + "issue:44315|issue:44360", + "issue:27957|issue:33453", + "issue:42757|issue:43576", + "issue:43576|issue:45071", + "issue:43576|issue:43824", + "issue:31795|issue:38617", + "issue:44961|issue:45468", + "issue:44222|issue:44230", + "issue:43576|issue:44938", + "issue:41628|issue:42548", + "issue:46143|issue:46144", + "issue:44908|issue:44961", + "issue:44485|issue:44961", + "issue:36296|issue:43901", + "issue:36296|issue:43526", + "issue:44315|issue:46209", + "issue:43673|issue:44492", + "issue:44623|issue:45230", + "issue:45362|issue:45850", + "issue:44273|issue:45542", + "issue:44038|issue:45125", + "issue:44230|issue:46095", + "issue:44393|issue:45071", + "issue:44273|issue:45003", + "issue:43576|issue:44556" + ], + "prepared_review_unit_hash": "edf4b57a04afba782c8fb91d005ffe201c727a81cc700af12f0e1909d8a04255", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 36721, + "estimated_input_tokens": 9181, + "estimated_eval_tokens": 18618 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a noisy similarity cluster of unrelated Transformer regressions, import errors, docs issues, and model-specific bugs. I did not find any soft pair that clearly describes the same underlying issue.", + "confidence": 0.91, + "canonical_issue_reason": "Issue 45850 is the closest umbrella report: it broadly captures the theme of v5 breaking multiple models, which is the only common thread across the set. The other issues are narrower and mostly unrelated symptoms.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45850 is the best representative of the cluster\u2019s loose theme because it is broad and generic enough to subsume many of the v5-regression style complaints, though it is still not a true duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both are model-loading regressions in v5, but one is Qwen3VL MoE-specific and the other is a generic checkpoint reload failure; different underlying bugs." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "The first is about unwanted API calls with local_files_only, while the second is an unrelated, underspecified issue with no clear overlap." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "Accuracy problems with FP8 training are unrelated to a missing image-processor method causing an AttributeError." + }, + { + "left": "issue:30333", + "right": "issue:36296", + "accept": false, + "reason": "MLflow reporting after failed jobs is unrelated to a tensor-parallel training bug." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled Python file errors and PretrainedConfig type-checking regressions are different failure modes." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs still referencing pipeline() is unrelated to the other issue, which does not identify the same bug." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "One is about compiled Python files, the other about save_pretrained on Pipeline objects; different code paths." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Both concern loading, but one is tokenizer loading and the other is Llama Vision model loading; not the same underlying defect." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger kernel application in model_init has nothing to do with a DSA indexer missing ReLU." + }, + { + "left": "issue:27957", + "right": "issue:33453", + "accept": false, + "reason": "Flash Attention 2 support for XLMRoberta is unrelated to tokenizer loading regression." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "An import error from huggingface_hub is not the same as a broken transformers env CLI command." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "A broken env subcommand and a PretrainedConfig type-checking regression are distinct issues." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "The env CLI bug is unrelated to a specific model import failure." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Documentation confusion around forward arguments is unrelated to an ImportError for a missing symbol." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "No evidence these describe the same bug; one is empty/unclear, the other is a Gemma audio positional encoding issue." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "Both mention FP8/MoE, but one is about save_pretrained and the other about inference support; not the same concrete bug." + }, + { + "left": "issue:43576", + "right": "issue:44938", + "accept": false, + "reason": "A broken CLI command is unrelated to Python 3.14 load failures." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Different top-level import errors for different symbols; likely separate missing-export problems." + }, + { + "left": "issue:46143", + "right": "issue:46144", + "accept": false, + "reason": "Both are about RoFormer, but one is kwargs passthrough and the other is attention-interface usage; related area, not the same bug." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwarg handling is unrelated to the other issue." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "A GLM-5 RoPE implementation discussion is not the same as the other issue." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training and TextClassificationPipeline docs are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor-parallel training does not match a BeitImageProcessorFast reduce_labels bug." + }, + { + "left": "issue:44315", + "right": "issue:46209", + "accept": false, + "reason": "Liger Kernel model_init behavior and FP8 inference on B200 are different model/runtime problems." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "A cache-missing runtime bug is unrelated to a typo in cache strategy naming." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "Missing files from processor.save_pretrained is not the same as an unspecified bug report." + }, + { + "left": "issue:45362", + "right": "issue:45850", + "accept": false, + "reason": "A Qwen3.5 chat crash may fit the broader v5-breakage theme, but it is a distinct model-specific failure and not the same underlying bug." + }, + { + "left": "issue:44273", + "right": "issue:45542", + "accept": false, + "reason": "Lazy loading issues are unrelated to TensorBoard/TensorFlow backend setup problems." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Both touch Qwen/MoE/tensor parallelism, but they describe different model families and different concrete failures." + }, + { + "left": "issue:44230", + "right": "issue:46095", + "accept": false, + "reason": "FP8 inference support and a weight-conversion script request are different changes." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bounding-box output issues are unrelated to PretrainedConfig type checking." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading and unsafe sys.modules access are distinct bugs." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "The env CLI issue is unrelated to checkpoint reload failures after upgrading to v5.2/v5.3." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 52, + "review_unit_count": 53, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-10000-451", + "nodes": [ + "issue:31515", + "issue:31795", + "issue:33453", + "issue:34689", + "issue:41669", + "issue:42371", + "issue:42890", + "issue: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: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:44393|issue:45753", + "issue:41669|issue:43673", + "issue:43525|issue:45850", + "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:43408|issue:44315", + "issue:42890|issue:44908", + "issue:44297|issue:45230", + "issue:44295|issue:45230", + "issue:44246|issue:44273", + "issue:43352|issue:44393", + "issue:43576|issue:44908", + "issue:43976|issue:44393", + "issue:44908|issue:45071", + "issue:44704|issue:44908", + "issue:41669|issue:44351", + "issue:44492|issue:46328", + "issue:41669|issue:45341", + "issue:43976|issue:45850", + "issue:44393|issue:45478", + "issue:44077|issue:44861", + "issue:43976|issue:46218", + "issue:43827|issue:44908", + "issue:31795|issue:33453", + "issue:44315|issue:44829", + "issue:43576|issue:44273", + "issue:45230|issue:45310", + "issue:44655|issue:44861", + "issue:44360|issue:45468", + "issue:44273|issue:44908", + "issue:43576|issue:46218", + "issue:46018|issue:46172", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:41669|issue:43704", + "issue:31795|issue:34689", + "issue:31515|issue:31795", + "issue:43704|issue:44485", + "issue:44485|issue:45468" + ], + "prepared_review_unit_hash": "1273ea2cfb737f9561a9b2f07e8de97686a4861819c5a0b28baa61b6c8eaf09f", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 43, + "serialized_chars": 39866, + "estimated_input_tokens": 9967, + "estimated_eval_tokens": 20190 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly noisy cluster of unrelated Transformers issues. There are at least two believable duplicate pairs: 44556/45092 around v5 checkpoint-loading/meta-init incompatibility, and the stronger exact-title pair 45310/45478 about the Qwen3.5 MoE from_pretrained failure. The rest are only loosely theme-related.", + "confidence": 0.83, + "canonical_issue_reason": "issue:45310 is the strongest canonical duplicate anchor because 45478 has the exact same title and they describe the same Qwen3.5 MoE from_pretrained failure; 44556/45092 is a separate duplicate-like pair, not the same one.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44556 is the best overall cluster representative because it is a concrete, user-facing v5 loading/reload regression and is more descriptive than the broader noisy issues.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44393", + "right": "issue:45753", + "accept": false, + "reason": "Same model family, but bbox-output hallucination and an infinite-loop/image-specific failure are different bugs." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-time slowdown from wildcard imports is unrelated to missing cache state in chunked prefill." + }, + { + "left": "issue:43525", + "right": "issue:45850", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is specific and concrete; the other is a vague broad-breakage complaint." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": true, + "reason": "Both point to the same v5 checkpoint-loading/meta-init compatibility regression and could plausibly be fixed together." + }, + { + "left": "issue:43572", + "right": "issue:45850", + "accept": false, + "reason": "Model-specific pad_token_idx regression is not the same bug as a broad 'many models break' report." + }, + { + "left": "issue:44492", + "right": "issue:45850", + "accept": false, + "reason": "A typo in cache-strategy text is unrelated to broad model-loading breakage." + }, + { + "left": "issue:45278", + "right": "issue:45850", + "accept": false, + "reason": "Both are upgrade-related complaints, but one is import errors and the other is a vague generalization; not the same concrete defect." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 API settings and a cache-strategy typo are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:46218", + "accept": false, + "reason": "Import slowdown from wildcard imports is not the same as a Qwen3_5 import failure." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "Type-mismatch warning for sam3 models and Liger Kernel model_init behavior are different code paths." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Missing set_seed in tests and inverse_sqrt scheduler kwargs handling are unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer metadata mismatch is not the same as a generic placeholder bug report." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "position_ids buffer error and a placeholder bug report are unrelated." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Both touch loading, but import slowness and lazy-loading failure are not the same concrete bug." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash-Attention support gating and Qwen3-VL bbox hallucination are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "Broken env CLI and scheduler kwargs handling are different issues." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version support mismatch is unrelated to Qwen3-VL bbox output errors." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs being ignored and PretrainedConfig type checking are different bugs." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "Both involve kwargs, but they affect different APIs and failure modes." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-time slowdown and missing HybridCache import are not the same issue." + }, + { + "left": "issue:44492", + "right": "issue:46328", + "accept": false, + "reason": "Cache-strategy typo and a SequenceFeatureExtractor.pad performance regression are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import slowdown from wildcard imports has nothing to do with a testing_utils bug." + }, + { + "left": "issue:43976", + "right": "issue:45850", + "accept": false, + "reason": "Python compatibility issues are not the same as broad model-breakage complaints." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different symptoms in the same family: bbox output issue versus Qwen3.5 MoE from_pretrained failure." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "Optional post_init validation and tied-weights key handling are separate bugs." + }, + { + "left": "issue:43976", + "right": "issue:46218", + "accept": false, + "reason": "Python support mismatch is unrelated to a specific Qwen3_5 import error." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still referencing pipeline() is unrelated to scheduler kwargs handling." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Forward-doc confusion and tokenizer loading regression are unrelated." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application with model_init and flash_attention_3 training degeneration are different problems." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env CLI and lazy loading not working are too broad to call the same bug." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "A generic placeholder bug report is not the same as the Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Pipeline save_pretrained failure and _get_tied_weight_keys AttributeError are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer discussion and Gemma-4 audio positional encoding are unrelated." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading regressions and scheduler kwargs handling are different areas." + }, + { + "left": "issue:43576", + "right": "issue:46218", + "accept": false, + "reason": "Broken env CLI and Qwen3_5 import failure are different bugs." + }, + { + "left": "issue:46018", + "right": "issue:46172", + "accept": false, + "reason": "DeepSeek-V4 shared-expert gating and a tensor transpose issue are not the same concrete defect." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Timestamp calculation in a processor is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp calculation and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import slowdown and VRAM leak in dataloader threads are different bugs." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Forward-doc confusion is unrelated to a Llama model-loading breakage." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Slow checkpoint loading and confusing forward documentation are not the same issue." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leak in Qwen3ForCausalLM and GLM-5 RoPE implementation are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "Different model families and different implementation areas; not a duplicate." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The summary is a bit too strong for the packet. In particular, the cluster is not limited to a single believable duplicate pair: issue:45310 and issue:45478 have the same title and are a much more obvious duplicate-like match than the narrative suggests. Because of that, the claim that 44556/45092 is the only believable soft edge is not fully grounded, and the canonical-anchor reasoning is not conservative enough." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 1, + "component_count": 21, + "review_unit_index": 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: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:44492", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44938", + "issue:44998", + "issue:45290", + "issue:45310", + "issue:45468", + "issue:45850", + "issue:46018", + "issue:46095", + "issue:46144", + "issue:46218" + ], + "soft_pairs": [ + "issue:43298|issue:46095", + "issue:45850|issue:46218", + "issue:42371|issue:43704", + "issue:43502|issue:43519", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:44263|issue:45468", + "issue:43606|issue:45468", + "issue:44261|issue:45468", + "issue:44351|issue:46218", + "issue:43352|issue:46218", + "issue:42913|issue:43066", + "issue:44908|issue:44998", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44998|issue:45850", + "issue:44861|issue:45290", + "issue:44938|issue:44998", + "issue:43525|issue:44861", + "issue: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": "39f46537d12dbf654b70087c4ffe56c7dc02cf62eeee273b3d521a7b6765b4d3", + "component_budget": { + "node_count": 451, + "item_count": 451, + "soft_pair_count": 1666, + "serialized_chars": 862694, + "estimated_input_tokens": 215674, + "estimated_eval_tokens": 431604 + }, + "budget": { + "node_count": 42, + "item_count": 42, + "soft_pair_count": 39, + "serialized_chars": 35120, + "estimated_input_tokens": 8780, + "estimated_eval_tokens": 17816 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The items are mostly unrelated issues that only share superficial wording. None of the soft pairs look like the same underlying bug or change, so no duplicate merges are justified.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue: this cluster is an overbroad lexical grouping spanning unrelated bugs, regressions, and feature requests.", + "canonical_pr_reason": "No PRs are present in the cluster.", + "best_issue_reason": "No global best issue stands out; none of these issues can safely represent the cluster as a whole.", + "best_pr_reason": "No PRs are present in the cluster.", + "soft_edge_verdicts": [ + { + "left": "issue:43298", + "right": "issue:46095", + "accept": false, + "reason": "Different problems: continuous batching support for audio models vs a DeepSeek-V4 weight conversion question." + }, + { + "left": "issue:45850", + "right": "issue:46218", + "accept": false, + "reason": "A vague version-wide breakage report is not the same as a specific import failure for Qwen3.5." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 API guidance and a VRAM leak in Qwen3ForCausalLM are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "One is unwanted network requests with local_files_only; the other is an incorrect timestamp calculation." + }, + { + "left": "issue:43519", + "right": "issue:44623", + "accept": false, + "reason": "Timestamp logic in Qwen3VL is unrelated to missing files in processor.save_pretrained." + }, + { + "left": "issue:43519", + "right": "issue:43704", + "accept": false, + "reason": "These describe different failures: timestamp math vs VRAM leakage across dataloader threads." + }, + { + "left": "issue:44263", + "right": "issue:45468", + "accept": false, + "reason": "A torch.split return-value issue is not the same as Gemma4 audio positional encoding behavior." + }, + { + "left": "issue:43606", + "right": "issue:45468", + "accept": false, + "reason": "CPU offload device mismatch and audio positional encoding are different code paths and bugs." + }, + { + "left": "issue:44261", + "right": "issue:45468", + "accept": false, + "reason": "RMS norm epsilon precision discussion does not match Gemma4AudioRelPositionalEncoding." + }, + { + "left": "issue:44351", + "right": "issue:46218", + "accept": false, + "reason": "HybridCache import failure and Qwen3_5ForCausalLM import failure are distinct module issues." + }, + { + "left": "issue:43352", + "right": "issue:46218", + "accept": false, + "reason": "Flash Attention 2 support for Nemotron and a Qwen3.5 import error are unrelated." + }, + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both involve tokenizer changes, but they are different concrete regressions: behavior drift vs decoder type selection." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Scheduler keyword handling is unrelated to the unrelated 'Unemployment' placeholder issue." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A cache-strategy typo is not the same as the unrelated 'Unemployment' issue." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIPOutput attentions regression and the unrelated 'Unemployment' issue do not match." + }, + { + "left": "issue:44998", + "right": "issue:45850", + "accept": false, + "reason": "A nonsensical placeholder issue is not the same as a broad 5.0 compatibility complaint." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Tied-weight key handling and chat template crashes are different failures." + }, + { + "left": "issue:44938", + "right": "issue:44998", + "accept": false, + "reason": "Python 3.14 import failure is unrelated to the 'Unemployment' issue." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config is distinct from _tied_weight_keys list handling." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression is unrelated to tied-weight key crashes." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Slow from_pretrained checkpoint loading and a missing layer_type_validation import are different bugs." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "FP8 inference support for Qwen3-VL/Qwen3.5 MoE is not the same as processor.save_pretrained file loss." + }, + { + "left": "issue:30333", + "right": "issue:31515", + "accept": false, + "reason": "MLFlow reporting on failed training jobs is unrelated to slow checkpoint loading." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "A slow checkpoint load is not the same as a tensor parallel training bug." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "FP8 support work and a Qwen3.5 MoE from_pretrained error are different concrete changes/failures." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Checkpoint loading performance and tokenizer loading regression are separate problems." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching support for audio models is unrelated to FP8 inference support." + }, + { + "left": "issue:43717", + "right": "issue:46144", + "accept": false, + "reason": "Mamba init_weights usage and RoFormer attention-interface usage are different model implementation issues." + }, + { + "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": "A ModernBERT type-hint bug is unrelated to tokenizer regex warnings." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2Former ignore_value and loss-kwargs forwarding are different code paths." + }, + { + "left": "issue:43240", + "right": "issue:46144", + "accept": false, + "reason": "Loss kwargs forwarding and RoFormer attention-interface usage are unrelated bugs." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former ignore_value and tokenizer regex warnings are not the same issue." + }, + { + "left": "issue:42936", + "right": "issue:46144", + "accept": false, + "reason": "Mask2Former ignore_value handling does not overlap with RoFormer attention implementation." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "Ignore_value usage and a quantization_operation typo are unrelated." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Missing loss kwargs and ModernBERT norm_eps typing are unrelated." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "A type-hint issue and a typo in quantization_operation are different problems." + }, + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Llama4TextL2Norm/RMS norm discussion is not the same as ModernBERT norm_eps typing." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "Mask2Former ignore_value and ModernBERT norm_eps typing are unrelated." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": "477f17e52149011010c89fc0015c0327625f6fafcc2fab8ae7e0a9817cf5a0d4", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2875, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 2875, + "estimated_input_tokens": 719, + "estimated_eval_tokens": 1694 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The three issues are unrelated feature requests in different areas: token truncation behavior, a ReduceLROnPlateau-like scheduler feature, and Wav2Vec2 word boundary offsets. This does not look like a duplicate cluster.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:10519", + "right": "issue:11307", + "accept": false, + "reason": "Different topics and code paths: token truncation direction vs. speech model word offset extraction. No evidence of the same underlying bug or change." + }, + { + "left": "issue:10519", + "right": "issue:11005", + "accept": false, + "reason": "One is about truncation strategy in tokenization; the other asks for learning-rate scheduler functionality. They are unrelated feature requests." + }, + { + "left": "issue:11005", + "right": "issue:11307", + "accept": false, + "reason": "Scheduler behavior request vs. Wav2Vec2 time-offset extraction. No shared concrete bug or mergeable code-path 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": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-13244-3", + "nodes": [ + "issue:13244", + "issue:15980", + "issue:16618" + ], + "soft_pairs": [ + "issue:13244|issue:16618", + "issue:15980|issue:16618" + ], + "prepared_review_unit_hash": "749c4db89dabf7abc5e75e3519ce18024c184de211f12d9bbc07dec6e6ecba2a", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2477, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 2477, + "estimated_input_tokens": 620, + "estimated_eval_tokens": 1496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This does not look like a single duplicate bug cluster. 16618 is the broadest model-loading issue, but 15980 is about private-model login/error messaging and 13244 is about Tapas tokenization behavior, so the three are likely separate problems.", + "confidence": 0.79, + "canonical_issue_reason": "Issue 16618 is the best cluster representative because it is the most general model-loading complaint; the other issues are narrower and point to different failure modes.", + "canonical_pr_reason": null, + "best_issue_reason": "16618 is the best issue to anchor the cluster only because it is the closest to a generic loading problem, but it is not a strong duplicate of the others.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tokenization behavior for Tapas and failing to load bert-base-uncased are different code paths and different bugs." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Private-model login/error-message handling is not the same underlying issue as a generic public model loading failure." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 4, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-26350-2", + "nodes": [ + "issue:26350", + "issue:33260" + ], + "soft_pairs": [ + "issue:26350|issue:33260" + ], + "prepared_review_unit_hash": "5f4ee6f66f7a79ba543d2007bc735686053324d94ac9acc8beaf7f6c9d02aa7b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1589, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two unrelated community contribution issues: one for Flash Attention 2 support on more architectures and one for GGUF support on more architectures. They share a similar framing but not the same underlying feature 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:26350", + "right": "issue:33260", + "accept": false, + "reason": "Both are broad community contribution requests, but they target different features and code paths (Flash Attention 2 vs GGUF support). Similar wording is not enough to treat them as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 5, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-26413-2", + "nodes": [ + "issue:26413", + "issue:31501" + ], + "soft_pairs": [ + "issue:26413|issue:31501" + ], + "prepared_review_unit_hash": "81e6c243e340ee80a42f56d93a6a91f81bfa9f913bb122fb640baa53c4b69f3b", + "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 both concern training failures, but they describe different symptoms and likely different root causes: one is a specific `resume_from_checkpoint` epoch-iterator error, the other is a broader report that FineWeb SLM training does not start. I would not deduplicate them.", + "confidence": 0.92, + "canonical_issue_reason": "Issue #26413 is the more specific and technically concrete report, with an explicit failure message and a clearly scoped code path (`resume_from_checkpoint`).", + "canonical_pr_reason": null, + "best_issue_reason": "#26413 is the best representative because it names the exact exception and failure point, making it more actionable as the canonical issue for triage.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26413", + "right": "issue:31501", + "accept": false, + "reason": "The reports are too different to treat as the same bug: #26413 is a checkpoint-resume iterator crash, while #31501 is a generic startup/training-init failure. Same broad area, but not 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": 6, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": "970d8ddb664c0143ff2b746ee7f92a0d1ee54aed40a6841a0b248e525d6556ae", + "component_budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13531, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 24, + "serialized_chars": 13531, + "estimated_input_tokens": 3383, + "estimated_eval_tokens": 7022 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This is a mixed i18n cluster: 27088, 28936, and 38435 are clearly separate language-labeled translation requests; 43834/44365/44370 form a highly similar placeholder-title subgroup that may be duplicates; and 46222/46226 are an exact duplicate pair. The safest representative is 43834, but the cluster is not one single duplicate set.", + "confidence": 0.9, + "canonical_issue_reason": "43834 is the best representative of the dense placeholder-title subgroup and sits near the center of the cluster, but it should be treated conservatively because the cluster also contains separate language-specific requests and an exact Thai duplicate pair.", + "canonical_pr_reason": null, + "best_issue_reason": "43834 is the strongest global issue candidate because it belongs to the most internally consistent near-duplicate subgroup and is more representative than the exact two-item Thai pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Different language-specific translation requests; only the broad docs/i18n theme overlaps." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Different language-specific translation requests; only the broad docs/i18n theme overlaps." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Turkish-specific request vs a generic placeholder-title issue; not enough evidence of the same report." + }, + { + "left": "issue:27088", + "right": "issue:38435", + "accept": false, + "reason": "Turkish and Romanian translation requests are distinct issues." + }, + { + "left": "issue:27088", + "right": "issue:28936", + "accept": false, + "reason": "Turkish and Spanish translation requests are distinct issues." + }, + { + "left": "issue:28936", + "right": "issue:43834", + "accept": false, + "reason": "Spanish-specific request vs a generic placeholder-title issue; not the same concrete issue." + }, + { + "left": "issue:28936", + "right": "issue:44365", + "accept": false, + "reason": "Different translation requests; same broad i18n theme is not enough to merge." + }, + { + "left": "issue:28936", + "right": "issue:38435", + "accept": false, + "reason": "Spanish and Romanian translation requests are distinct issues." + }, + { + "left": "issue:28936", + "right": "issue:44370", + "accept": false, + "reason": "Different translation requests; same broad i18n theme is not enough to merge." + }, + { + "left": "issue:38435", + "right": "issue:44370", + "accept": false, + "reason": "Romanian-specific request vs a generic placeholder-title issue; not enough evidence of the same report." + }, + { + "left": "issue:38435", + "right": "issue:44365", + "accept": false, + "reason": "Romanian-specific request vs a generic placeholder-title issue; not enough evidence of the same report." + }, + { + "left": "issue:38435", + "right": "issue:43834", + "accept": false, + "reason": "Romanian-specific request vs a generic placeholder-title issue; not enough evidence of the same report." + }, + { + "left": "issue:38435", + "right": "issue:46222", + "accept": false, + "reason": "Different titles and no evidence they describe the same underlying report." + }, + { + "left": "issue:38435", + "right": "issue:46226", + "accept": false, + "reason": "Different titles and no evidence they describe the same underlying report." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": true, + "reason": "Identical placeholder-style title and extremely high similarity; these look like the same underlying issue." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": true, + "reason": "Identical placeholder-style title and extremely high similarity; these look like the same underlying issue." + }, + { + "left": "issue:43834", + "right": "issue:46222", + "accept": false, + "reason": "Only broad docs-translation similarity is visible; the titles do not show the same concrete report." + }, + { + "left": "issue:43834", + "right": "issue:46226", + "accept": false, + "reason": "Only broad docs-translation similarity is visible; the titles do not show the same concrete report." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": true, + "reason": "Identical placeholder-style title and near-identical bodies; this is a strong duplicate match." + }, + { + "left": "issue:44365", + "right": "issue:46222", + "accept": false, + "reason": "Only broad docs-translation similarity is visible; not enough to treat as the same issue." + }, + { + "left": "issue:44365", + "right": "issue:46226", + "accept": false, + "reason": "Only broad docs-translation similarity is visible; not enough to treat as the same issue." + }, + { + "left": "issue:44370", + "right": "issue:46222", + "accept": false, + "reason": "Only broad docs-translation similarity is visible; not enough to treat as the same issue." + }, + { + "left": "issue:44370", + "right": "issue:46226", + "accept": false, + "reason": "Only broad docs-translation similarity is visible; not enough to treat as the same issue." + }, + { + "left": "issue:46222", + "right": "issue:46226", + "accept": true, + "reason": "Same exact title and identical report shape; this is a true duplicate pair." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The duplicate call on 46222/46226 is well grounded, but the cluster summary overreaches by describing the rest as distinct issues in different languages. The packet only directly supports language-specific separation for 27088/28936/38435; it does not support that claim for the identical placeholder-title trio 43834/44365/44370, which have extremely high similarity and could also be near-duplicates. Please make that distinction more conservative and avoid asserting different-language separation where the packet does not show it." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "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": "6d1d0b11348ba24ae1e66c649f1c97c32fef8659eb7ad8ee1796e0eb2ae87338", + "component_budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21662, + "estimated_input_tokens": 5416, + "estimated_eval_tokens": 11088 + }, + "budget": { + "node_count": 16, + "item_count": 16, + "soft_pair_count": 36, + "serialized_chars": 21662, + "estimated_input_tokens": 5416, + "estimated_eval_tokens": 11088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is a loose collection of model-request issues that all say \u201cadd/support model X,\u201d but they describe different models and code changes, not the same underlying bug or feature. No PRs are present.", + "confidence": 0.95, + "canonical_issue_reason": "No true canonical duplicate exists here: the issues request different model additions/architectures and should remain separate.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45322 is the cleanest, most self-contained model-request example in the set, so it is the best representative if one issue must be chosen.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "Different model requests: EuroBert config vs DEIMv2. Same broad theme, not the same change." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config request is unrelated to Qwen3-TTS support." + }, + { + "left": "issue:36817", + "right": "issue:36879", + "accept": false, + "reason": "EuroBert and RF-DETR are unrelated model additions." + }, + { + "left": "issue:36879", + "right": "issue:41211", + "accept": false, + "reason": "RF-DETR and DEIMv2 are different models and different implementation paths." + }, + { + "left": "issue:36879", + "right": "issue:43909", + "accept": false, + "reason": "RF-DETR and LFM2.5 Audio are unrelated additions." + }, + { + "left": "issue:36879", + "right": "issue:44121", + "accept": false, + "reason": "RF-DETR and OpenAI weight-sparse transformer are different model requests." + }, + { + "left": "issue:36879", + "right": "issue:43837", + "accept": false, + "reason": "RF-DETR and Qwen3-ASR are different concrete models." + }, + { + "left": "issue:36879", + "right": "issue:43671", + "accept": false, + "reason": "RF-DETR and Qwen3-TTS are unrelated model additions." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are different model families and code changes." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT and circuitgpt/weight-sparse transformer are unrelated requests." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT and LFM2.5 Audio are different models with different support needs." + }, + { + "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 distinct model requests, not duplicates." + }, + { + "left": "issue:41115", + "right": "issue:43837", + "accept": false, + "reason": "MiniCPM3 architecture support is unrelated to Qwen3-ASR support." + }, + { + "left": "issue:41115", + "right": "issue:44121", + "accept": false, + "reason": "MiniCPM3 and circuit-sparse transformer are different concrete models." + }, + { + "left": "issue:41115", + "right": "issue:42503", + "accept": false, + "reason": "MiniCPM3 and ModernVBERT are separate model additions." + }, + { + "left": "issue:41115", + "right": "issue:45027", + "accept": false, + "reason": "MiniCPM3 and Voxtral-4B-TTS-2603 are different model integrations." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT and Qwen3-ASR are different models and code paths." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are unrelated additions." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT and generic BERT-like RoPE support are not the same concrete change." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like models with RoPE and Voxtral TTS support are unrelated requests." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are related by name but are distinct speech models and separate changes." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are different model integrations." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model requests." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS and the weight-sparse transformer request are different model changes." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS and LFM2.5 Audio are different audio-model support requests." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio are different model integrations." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are unrelated model requests." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR and circuit-sparse transformer are different concrete changes." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio and SAM3-LiteText are different model families." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Weight-sparse transformer support and EUPE are unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral-4B-TTS-2603 are distinct speech-model requests." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral-4B-TTS-2603 are unrelated model support tasks." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are different model requests." + }, + { + "left": "issue:45322", + "right": "issue:46133", + "accept": false, + "reason": "EUPE and TIPSv2 are separate vision-language model requests from different authors and model families." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 8, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-38740-2", + "nodes": [ + "issue:38740", + "issue:46443" + ], + "soft_pairs": [ + "issue:38740|issue:46443" + ], + "prepared_review_unit_hash": "e2a21758b76ea719607ee1ca2730a4c737b77ce94c670c6ba1e136a407248b14", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1505, + "estimated_input_tokens": 377, + "estimated_eval_tokens": 1010 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one is a docs request about adding `pruna`, the other is a Vision RoPE refactor. The soft similarity signal is not enough to treat them as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:38740 ([DOCS] Add `pruna` as optimization framework)", + "right": "issue:46443 (Vision RoPE refactor)", + "accept": false, + "reason": "Different topics and problem types: documentation/framework addition vs model architecture refactor; no shared code path or duplicate 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": 9, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41084-2", + "nodes": [ + "issue:41084", + "issue:43037" + ], + "soft_pairs": [ + "issue:41084|issue:43037" + ], + "prepared_review_unit_hash": "38168fa2b04c108ee3c9c3d0d24748561ce27798a199cceec46424fafedd67ca", + "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": "The two issues are not duplicates: one is about block decoding, the other about DeepSeek v3.2 support. No PRs are present.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Unrelated topics and no shared concrete bug/change path; a decoding setting request is not the same as model-version support." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 10, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "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": "fe2e0c70170ddbcbf5d8d720ce0cb6cae7b63548a06a8d14c740cba5c8594740", + "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": "The two closed issues are related only by a broad theme of custom trainer compatibility, but they describe different problems: one is about `remove_unused_column` customizability in `Trainer`, the other about FLOPS estimation for custom trainers. They should not be deduplicated.", + "confidence": 0.94, + "canonical_issue_reason": "No single canonical issue: the topics overlap only at the level of custom trainer support, but the underlying bugs are different code paths and would need separate fixes.", + "canonical_pr_reason": null, + "best_issue_reason": "If a representative issue is needed, #43598 is the better anchor because it names a specific Trainer behavior and a concrete customization point, making the scope clearer than the FLOPS-accounting request.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Different underlying issues: one targets `remove_unused_column` handling, the other FLOPS counting for custom trainers; same subsystem, 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": 11, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43684-2", + "nodes": [ + "issue:43684", + "issue:44405" + ], + "soft_pairs": [ + "issue:43684|issue:44405" + ], + "prepared_review_unit_hash": "b7d62275c67c8a6ece832e4eb16a4182481eba65941907f289e78286224e8538", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1612, + "estimated_input_tokens": 403, + "estimated_eval_tokens": 1062 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues are both about adding missing model support, but they target different model families and different AutoModel heads, so they are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": "Issue #43684 is the broader request, covering registration support for Qwen3-Omni in AutoModel and AutoModelForConditionalGeneration, so it is the better representative of the cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "#43684 is the best issue to keep as the cluster representative because it describes the more general model-registration gap rather than a single task-specific head.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different underlying changes: Qwen3-Omni registration for AutoModel/AutoModelForConditionalGeneration versus AutoModelForSequenceClassification support for Qwen3.5. They involve different configs and different code paths, so they should not be merged as the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": "d71da42388bf8ecba4b20ca1c37e5863f01a6dd86a6720cd96c2496544f37f74", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4265, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4265, + "estimated_input_tokens": 1067, + "estimated_eval_tokens": 2390 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The three \"huggingface \" issues appear to be duplicates of the same report; \"Fett\" looks unrelated. I would treat the earliest huggingface issue as the canonical issue and the other two as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": "Issue #43845 is the earliest of the near-identical \"huggingface \" reports and is the best representative of that duplicate cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue #43845 is the best global issue choice because it is the first clear member of the repeated \"huggingface \" group, while #43835 has a different title and no visible overlap in content.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Different titles and no content evidence beyond weak similarity; not enough to conclude the same underlying issue." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Different titles and no shared details indicating the same bug." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles and no concrete overlap in the issue descriptions." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": true, + "reason": "Near-identical title and timing strongly suggest duplicate reports of the same underlying issue." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": true, + "reason": "Same title and same apparent topic; likely a duplicate of the same report." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": true, + "reason": "Same title and repeated topic with no distinguishing details, so they likely describe the same issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded overall. The three identical \"huggingface \" issues are reasonably grouped as a likely duplicate cluster based on the matching titles and high soft-similarity scores, while #43835 is plausibly separate given its different title and much weaker similarity. The earliest huggingface report (#43845) is a conservative canonical choice." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "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": "05e822059b0772b7fcced7205cbde2681ffebe8aac92aaebbb1ef50b1fe9dfa9", + "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": "The cluster is made up of low-signal, likely spam or placeholder issues with only superficial textual similarity. They do not appear to describe the same underlying bug or request, so they should not be deduplicated together.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are generic spam-like placeholders; no evidence they report the same underlying issue." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Both are generic spam-like placeholders; superficial similarity is not enough to treat them as duplicates." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "One is spam-like, the other is a vague documentation suggestion; they do not match on a concrete problem." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "The titles are unrelated and do not indicate the same bug or request." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "A vague documentation suggestion is not the same as a spam-like placeholder issue." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "These do not describe the same concrete request or failure mode." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are spam-like placeholders, but there is no evidence they refer to the same underlying report." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "The titles suggest different content; no shared bug/request is apparent." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "No concrete overlap beyond superficial similarity, so they should not be merged." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 14, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45145-2", + "nodes": [ + "issue:45145", + "issue:45151" + ], + "soft_pairs": [ + "issue:45145|issue:45151" + ], + "prepared_review_unit_hash": "c656e317459e162e74ff56ae3ed46844b4d8596d79e2711ee4a6d92d02138a57", + "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": "#45145 is the best canonical issue record because it was opened first and has the only inbound reference. #45151 looks highly similar, but the available evidence is limited to matching titles and soft similarity, so the duplicate call should stay conservative.", + "confidence": 0.88, + "canonical_issue_reason": "#45145 is earlier and has the only inbound reference, making it the best canonical issue to keep.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the strongest issue representative for the cluster because it is earliest and slightly better connected.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": false, + "reason": "The issues have the same title and appear closely related, but only soft-similarity evidence is available here; without body text or other distinguishing detail, this is not enough to confirm the same underlying request." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The canonical selection is mostly grounded (#45145 is earlier and has one inbound reference vs. none). However, the duplicate-issue conclusion is overstated for the packet: the evidence shown is only soft similarity plus identical titles, and there is no body text or other content here to support \u201cnear-exact,\u201d \u201csame apparent content,\u201d or \u201cno distinguishing technical differences.\u201d The soft edge verdict should be more conservative, e.g. \u201clikely duplicate / likely same request,\u201d not a strong accept." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 15, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45295-2", + "nodes": [ + "issue:45295", + "issue:45373" + ], + "soft_pairs": [], + "prepared_review_unit_hash": "f5fd95fada2f530ab0999e343c17c16266cde0c7182bb22509972ca44827cd49", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 0, + "serialized_chars": 1286, + "estimated_input_tokens": 322, + "estimated_eval_tokens": 900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues describe the same missing Gemma 4 sequence-classification support; the later, more specific open issue is the better cluster representative.", + "confidence": 0.94, + "canonical_issue_reason": "issue #45373 is the clearest canonical tracker: it is open, explicitly names the missing class `Gemma4ForSequenceClassification`, and states the module-level gap relative to Gemma 2/3.", + "canonical_pr_reason": null, + "best_issue_reason": "issue #45373 best captures the concrete request and likely remaining work, while #45295 is a shorter earlier duplicate with less detail and is already closed.", + "best_pr_reason": null, + "soft_edge_verdicts": [] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 16, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45672-2", + "nodes": [ + "issue:45672", + "issue:45674" + ], + "soft_pairs": [ + "issue:45672|issue:45674" + ], + "prepared_review_unit_hash": "2847beeb0078578be6da3098a2f55e1603fa1dd4b47e1e5a98c9ababbe29d676", + "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 to 4-bit inference, but they describe different bugs and should not be deduplicated. One is a torch.finfo TypeError triggered by uint8 audio-module weights in Gemma4; the other is a BitsAndBytesConfig regression where llm_int8_skip_modules drops the default lm_head exclusion and causes an AssertionError.", + "confidence": 0.97, + "canonical_issue_reason": "No canonical issue for the pair: issue 45672 is an audio-module uint8/torch.finfo failure, while issue 45674 is a separate BitsAndBytesConfig default-exclusion regression.", + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good global duplicate target for the other; they have different root causes and would need separate fixes.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45672", + "right": "issue:45674", + "accept": false, + "reason": "Different failure modes and code paths: 45672 is about torch.finfo on uint8 weights in audio modules, while 45674 is about llm_int8_skip_modules clearing the default lm_head exclusion in 4-bit inference." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 17, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "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": "63e87cac3a1210700acfbc5c8b45439c303df3563eb4a25cf56c93d2c707a593", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1476, + "estimated_input_tokens": 369, + "estimated_eval_tokens": 994 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are closed issues with only person-name/date titles and no body, links, or shared targets; they do not appear to describe the same bug or request, so this is not a duplicate cluster.", + "confidence": 0.98, + "canonical_issue_reason": "No clear canonical issue: neither issue contains enough technical detail, and they do not share a specific problem statement or target.", + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good global representative for a technical cluster because both look like isolated, content-free entries rather than actionable bug reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45762", + "right": "issue:45763", + "accept": false, + "reason": "The titles are different personal-name/date strings with no shared code context, body, links, or target; they do not look like the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 18, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45764-2", + "nodes": [ + "issue:45764", + "issue:45765" + ], + "soft_pairs": [ + "issue:45764|issue:45765" + ], + "prepared_review_unit_hash": "85f59221a84bcfbcef99891707844afb90e5d3b4fc596e3257676fe0010aff84", + "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 are just empty title-only reports with different personal-name/date strings and no shared technical context, so they should not be treated as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45764", + "right": "issue:45765", + "accept": false, + "reason": "No evidence of the same underlying bug or request: different titles, no body, no linked targets, and no shared technical context." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 19, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "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": "6cae1f4b9275f29c0c333e4321b1a44927132884998a4d977974712a30a4bc12", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1542, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1542, + "estimated_input_tokens": 386, + "estimated_eval_tokens": 1028 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are unrelated: one requests a new Qwen3_5 token-classification model for value modeling, while the other asks for FP8 training support. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": "No canonical issue: the requests target different problems and would not be resolved by the same change.", + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a good cluster representative because they are about separate feature requests with no shared underlying bug or fix.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45810", + "right": "issue:46295", + "accept": false, + "reason": "Different topics and code paths: adding a model class versus enabling FP8 training. They do not describe the same underlying issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 20, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-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": "2da42aea1df98d74cd979b96ff096e9aa7a951cf9f3ab2b3620dfc6f5d37d9fb", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5727, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 8, + "serialized_chars": 5727, + "estimated_input_tokens": 1432, + "estimated_eval_tokens": 3120 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster appears to split into two duplicate-looking/same-topic issue pairs: 45870/45871 and 45872/45873. Issue 45869 looks related thematically but not like the same issue as either pair.", + "confidence": 0.79, + "canonical_issue_reason": "45870 is a reasonable representative for the 45870/45871 same-topic pair because it is the earlier of the two and their titles/body lengths match closely; however, the cluster does not look like one single duplicate set overall.", + "canonical_pr_reason": null, + "best_issue_reason": "45870 is the best issue to keep as a representative for the strongest same-topic pair, since it is the earlier of the matching 45870/45871 issues and there is no stronger evidence tying 45869 into the same set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45869", + "right": "issue:45870", + "accept": false, + "reason": "They share financial/DataForge context, but one is a dataset/free-sample post and the other is a prediction example; they do not look like the same issue." + }, + { + "left": "issue:45869", + "right": "issue:45871", + "accept": false, + "reason": "They share financial/DataForge context, but one is a dataset/free-sample post and the other is a prediction example; they do not look like the same issue." + }, + { + "left": "issue:45870", + "right": "issue:45871", + "accept": true, + "reason": "Identical title/body-length pair with very strong soft-similarity; these appear to be duplicate or same-topic issues." + }, + { + "left": "issue:45870", + "right": "issue:45872", + "accept": false, + "reason": "Both are about DataForge finance-related content, but they describe different topics: an example versus a scraping-service request." + }, + { + "left": "issue:45870", + "right": "issue:45873", + "accept": false, + "reason": "Both are about DataForge finance-related content, but they describe different topics: an example versus a scraping-service request." + }, + { + "left": "issue:45871", + "right": "issue:45872", + "accept": false, + "reason": "Both are about DataForge finance-related content, but they describe different topics: an example versus a scraping-service request." + }, + { + "left": "issue:45871", + "right": "issue:45873", + "accept": false, + "reason": "Both are about DataForge finance-related content, but they describe different topics: an example versus a scraping-service request." + }, + { + "left": "issue:45872", + "right": "issue:45873", + "accept": true, + "reason": "Identical title/body-length pair with very strong soft-similarity; these appear to be duplicate or same-topic issues." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster split is plausible, but the writeup overstates the evidence. The packet only shows identical titles and body lengths plus soft-similarity scores; it does not prove the pairs are exact duplicate title/body matches. Please soften the language to something like \"appear to be duplicate/same-topic issues\" and avoid claiming exact duplicates or same underlying issue unless the bodies or explicit links support it. The 45869 non-match call is reasonably conservative, but the cluster-level summary is too strong." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 21, + "component_count": 21, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46439-2", + "nodes": [ + "issue:46439", + "issue:46441" + ], + "soft_pairs": [ + "issue:46439|issue:46441" + ], + "prepared_review_unit_hash": "389a1a13a49eadbd479354e9053f3462e5bb239d236f61cea9a035fc294ee44b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1636, + "estimated_input_tokens": 409, + "estimated_eval_tokens": 1074 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two issues are closely related around static-cache and linear-attention handling, but the packet does not establish that the Qwen3.5 `KeyError: 'linear_attention'` and the wrong-shape `RuntimeError` are the same root bug. They should be treated as related reports rather than confirmed duplicates.", + "confidence": 0.61, + "canonical_issue_reason": "Issue 46439 is the better canonical issue because it describes the broader mechanism (`early_initialization` corrupting linear-attention layers) instead of a model-specific symptom.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 46439 is the best representative for the cluster since it is the more general diagnostic report, while 46441 appears to be a narrower failure report in the same area.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:46439", + "right": "issue:46441", + "accept": false, + "reason": "The titles suggest a shared static-cache/linear-attention area, but the evidence is too thin to conclude the `KeyError` and wrong-shape `RuntimeError` come from the same underlying bug." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The conclusion is a bit too strong for the evidence in the packet. We only have titles and a soft-similarity signal, so saying the two issues are the same underlying bug is not fully grounded. A more conservative verdict would be that they appear related and may involve the same static-cache/linear-attention area, but the packet does not establish that the KeyError and wrong-shape RuntimeError are manifestations of one root cause." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 1, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-39895-14", + "nodes": [ + "pull_request:39895", + "pull_request:41855", + "pull_request:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45170", + "pull_request:45213", + "pull_request:45655", + "pull_request:46048", + "pull_request:46267" + ], + "soft_pairs": [ + "pull_request:39895|pull_request:43424", + "pull_request:41855|pull_request:43823", + "pull_request:41855|pull_request:43424", + "pull_request:42668|pull_request:43823", + "pull_request:43251|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:43424|pull_request:44827", + "pull_request:43424|pull_request:43823", + "pull_request:43823|pull_request:45655", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45170", + "pull_request:44827|pull_request:46267", + "pull_request:44827|pull_request:46048", + "pull_request:44827|pull_request:45655", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "327b0e7c28b236270ae23f60ff7a574975905568302999540eee4b506fad9b7b", + "component_budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 19, + "serialized_chars": 38083, + "estimated_input_tokens": 9521, + "estimated_eval_tokens": 19298 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 19, + "serialized_chars": 38083, + "estimated_input_tokens": 9521, + "estimated_eval_tokens": 19298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "This cluster is mostly a set of unrelated pull requests across different models, docs, and utilities. The only clear duplicate-like pair is the two `fixed_cross_entropy` changes, which address the same loss helper in the same code path.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the best representative of the only true duplicate pair: it is directly tied to issue 43240, changes the same `fixed_cross_entropy` helper, and reads like the more complete fix for passing through `cross_entropy` kwargs.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the strongest standalone candidate overall because it has the explicit issue target, a focused fix, and the clearest linkage to the underlying bug.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Different domains and code paths: VideoPrism model addition vs ExecuTorch dynamic-shape export test." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43823", + "accept": false, + "reason": "Unrelated tokenizer utility fix vs new MobileLLM model implementation." + }, + { + "left": "pull_request:41855", + "right": "pull_request:43424", + "accept": false, + "reason": "Tokenizer method addition and ExecuTorch export testing are not the same underlying change." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Processor auto-loading robustness and MobileLLM model addition are different features." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Loss helper kwargs fix is unrelated to adding a new model." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "The loss helper fix and Mistral4 test/model changes are unrelated code paths." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "ExecuTorch export testing and Mistral4 test/model work do not fix the same bug." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Different concerns: dynamic export test vs new MobileLLM model." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45655", + "accept": false, + "reason": "New model addition and `cls.config` resolution ordering are unrelated." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "MobileLLM model addition and Falcon output collection refactor are separate changes." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon refactor and model-creation skill docs are unrelated." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon standardized output refactor and Mistral4 test fixes do not share the same concrete bug." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 tests vs `layrnorm` typo normalization is a broad subsystem overlap only, not the same fix." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46267", + "accept": false, + "reason": "Mistral4 test fixes and serve chat parsing updates are unrelated." + }, + { + "left": "pull_request:44827", + "right": "pull_request:46048", + "accept": false, + "reason": "Mistral4 tests and adding a TDT loss kernel are different changes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45655", + "accept": false, + "reason": "Mistral4 test/model work and config resolution ordering are not the same underlying issue." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Model tests and the new-model skill documentation are not mergeable as one duplicate PR." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "A model implementation PR and a repository skill/documentation PR are unrelated." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Both PRs modify `fixed_cross_entropy` in the same file to support passing `cross_entropy` kwargs for the same issue 43240; they are alternative fixes for the same code-path bug." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The only accepted pair, 43251/43254, is clearly mergeable into one PR for the same bugfix: both change `src/transformers/loss/loss_utils.py`, both target issue 43240, and both extend `fixed_cross_entropy` to forward supported `cross_entropy` kwargs. The other soft-edge rejections are also cautious and consistent with the file lists and titles." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-3", + "nodes": [ + "pull_request:40546", + "pull_request:42781", + "pull_request:46331" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781", + "pull_request:42781|pull_request:46331" + ], + "prepared_review_unit_hash": "10ac43826f39335e5faa28af853569cd4912fe54304608461c5cfd944844ba39", + "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": "The cluster contains three distinct pull requests, not duplicate changes: an initial VibeVoice implementation, a later VibeVoice Realtime add-on, and an unrelated Parakeet-RNNT integration. Both soft edges should be rejected.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 40546 is the most central representative: it introduces the core VibeVoice implementation and integration surface, while 42781 is a separate realtime extension and 46331 targets a different model family.", + "best_issue_reason": null, + "best_pr_reason": "PR 40546 is the best standalone representative for the cluster because it is the original full VibeVoice implementation with the broadest scope and clearest subject matter.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Related topic, but not the same underlying change: 40546 implements core VibeVoice, while 42781 adds VibeVoice Realtime with acoustic-tokenizer and generation changes. They overlap in files but are not mergeable as one duplicate PR." + }, + { + "left": "pull_request:42781", + "right": "pull_request:46331", + "accept": false, + "reason": "Different model families and different concrete changes. The shared auto-mapping files are generic integration plumbing, not evidence of the same bug or feature." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40755-2", + "nodes": [ + "pull_request:40755", + "pull_request:41224" + ], + "soft_pairs": [ + "pull_request:40755|pull_request:41224" + ], + "prepared_review_unit_hash": "2d0021d37b7bf37e5e5cb7f94d5da50cb24b8b64f5891e4039a6e94d69f7d407", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are unrelated feature additions in different model families: TimesFM covariate forecasting vs. DINOv3 image classification support. They do not look like the same underlying bug or change.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs target different code paths, different models, and different user-facing features, so there is no duplicate anchor in this cluster.", + "best_issue_reason": null, + "best_pr_reason": "No global best PR: neither PR is a duplicate of the other, and each addresses a separate feature area with no meaningful mergeable overlap.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different subsystems and different changesets: TimesFM forecasting with covariates versus DINOv3 image classification support. They are not the same underlying change and should not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41356-2", + "nodes": [ + "pull_request:41356", + "pull_request:44339" + ], + "soft_pairs": [ + "pull_request:41356|pull_request:44339" + ], + "prepared_review_unit_hash": "aead1308c167b0bae6ff11bf9358340baa5e76f816618383418b505a10a9cd01", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6050, + "estimated_input_tokens": 1513, + "estimated_eval_tokens": 3282 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two DEIMv2 pull requests target the same model family, but they are not the same change: PR 41356 is an early/basic model + image processor add, while PR 44339 is a much broader follow-up that adds loss code, auto-mappings, conversion tooling, modular files, and more integration work.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44339 is the more complete and representative DEIMv2 implementation; PR 41356 looks like an earlier partial version of the same model bring-up.", + "best_issue_reason": null, + "best_pr_reason": "PR 44339 is the strongest single artifact for the cluster because it subsumes the basic model work and includes the broader integration needed for a full Transformers addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": false, + "reason": "Same model area, but not the same concrete change: 44339 expands the implementation substantially and adds new code paths beyond the basic PR 41356. They look like related iterations, not mergeable duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41797-3", + "nodes": [ + "pull_request:41797", + "pull_request:45186", + "pull_request:45630" + ], + "soft_pairs": [ + "pull_request:41797|pull_request:45186", + "pull_request:45186|pull_request:45630" + ], + "prepared_review_unit_hash": "b50793534bd4dea6da266cde310c14ef12e916d865e74c016ef03637d4db9955", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10435, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 10435, + "estimated_input_tokens": 2609, + "estimated_eval_tokens": 5474 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate model-addition PRs that only overlap on the standard Transformers boilerplate. The shared files and \"soft_similarity\" signals are not enough to make them duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR: 41797 adds DeepSeek OCR, 45186 adds Isaac, and 45630 adds Kimi2-6; they target different models and code paths.", + "best_issue_reason": null, + "best_pr_reason": "45630 is the strongest representative of the group: it has the most review activity and a broad, well-scoped model integration across auto mappings and model files.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different model additions (DeepSeek OCR vs Isaac). The overlap is only in shared Transformers scaffolding, not the same underlying bug or change." + }, + { + "left": "pull_request:45186", + "right": "pull_request:45630", + "accept": false, + "reason": "Different model additions (Isaac vs Kimi2-6). Shared docs/auto-registration files are generic boilerplate, not evidence of 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": 6, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42975-2", + "nodes": [ + "pull_request:42975", + "pull_request:42976" + ], + "soft_pairs": [ + "pull_request:42975|pull_request:42976" + ], + "prepared_review_unit_hash": "3d1f6d0fe6fa77a556b6d6f8155a4843dffb6ea33ac1f2f872635f86e0280831", + "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": "Two open PRs both modernize GitHub Actions in workflow files, but they target different dependency families and different underlying maintenance issues. They are related thematically, not duplicates.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the broader representative of the cluster\u2019s theme because it upgrades multiple core workflow actions for Node 24 compatibility across many workflows. It is still not a duplicate of 42976, which focuses on docker/build-push-action upgrades and a different workflow subset.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Reject: although both are workflow-action maintenance PRs, 42975 is about GitHub Actions runtime/version compatibility (checkout/cache/upload-artifact), while 42976 primarily upgrades docker/build-push-action and related workflow action versions. They do not fix the same concrete code-path problem and would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-42978-4", + "nodes": [ + "pull_request:42978", + "pull_request:43448", + "pull_request:43451", + "pull_request:45490" + ], + "soft_pairs": [ + "pull_request:42978|pull_request:43448", + "pull_request:42978|pull_request:43451", + "pull_request:42978|pull_request:45490", + "pull_request:43448|pull_request:43451", + "pull_request:43448|pull_request:45490", + "pull_request:43451|pull_request:45490" + ], + "prepared_review_unit_hash": "ec5d7773f7acf6fc5179e57c0fd0c94771f81bc68e444db0a7b8a150f0b38aed", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14253, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14253, + "estimated_input_tokens": 3564, + "estimated_eval_tokens": 7384 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All four pull requests are new model additions for different architectures (ViT NEPA, Molmo, Molmo2, CTSM). They share some registration/docs boilerplate, but they do not fix the same concrete code-path problem, so none should be deduplicated against another.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR is a duplicate canonical for this cluster; if one must be chosen as a representative, Molmo2 is the most substantial and complete addition, but it is still a distinct model from the others.", + "best_issue_reason": null, + "best_pr_reason": "PR 43451 (Molmo2) is the strongest representative on breadth and completeness, with the highest review activity and the most extensive model/processor coverage, but it is not a duplicate of the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model families: ViT NEPA vs Molmo. Shared auto/doc updates are generic model-integration boilerplate, not the same underlying change or bug fix." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "ViT NEPA and Molmo2 are distinct architectures and code paths. Overlap is limited to common registration/docs files, so they are not mergeable as one PR." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "ViT NEPA is a vision model add-on, while CTSM is a time-series model add-on. They touch unrelated modeling code and do not address the same change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are related by naming but still separate model implementations with different files and functionality; this is not the same concrete fix/change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo and CTSM are unrelated model additions. Shared files are just integration points, which is insufficient for duplicate PR merging." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM solve different model-support tasks in different subsystems, with no shared bug or code-path target beyond generic auto/registry updates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43291-4", + "nodes": [ + "pull_request:43291", + "pull_request:43665", + "pull_request:44815", + "pull_request:46380" + ], + "soft_pairs": [ + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43291|pull_request:46380", + "pull_request:43665|pull_request:44815", + "pull_request:43665|pull_request:46380", + "pull_request:44815|pull_request:46380" + ], + "prepared_review_unit_hash": "efe380863eed3728627fb670d4657896a91793b2599323477e780ddff2c3f47c", + "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": "No strong duplicate cluster here: the four PRs target different bugs or even look like unrelated test noise. I would not merge any soft-edge pair.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43291 is the clearest single bugfix: it fixes Whisper decode behavior in one concrete code path and has a focused scope.", + "best_issue_reason": null, + "best_pr_reason": "PR 43291 is the best representative of a real, narrowly scoped defect fix. The others are either unrelated, much broader, or appear to be test-placeholder noise.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Not the same underlying change: 43291 fixes Whisper decoding, while 43665 is just test-file edits with bogus assertions and no concrete bugfix path." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Different bugs and code paths: Whisper tokenization/decoding versus FP8 dequantization and model loading." + }, + { + "left": "pull_request:43291", + "right": "pull_request:46380", + "accept": false, + "reason": "43291 is a Whisper tokenizer decode fix; 46380 is about generation config/watermarking validation and broader generation/model changes, not the same defect." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "43665 looks like placeholder test breakage, not a real fix; 44815 is a real FP8 dequant fix. No shared underlying bug." + }, + { + "left": "pull_request:43665", + "right": "pull_request:46380", + "accept": false, + "reason": "43665 is nonsensical test-only noise, while 46380 is a substantive generation-validation change. Not duplicates." + }, + { + "left": "pull_request:44815", + "right": "pull_request:46380", + "accept": false, + "reason": "Both are broad PRs but they fix different concrete problems: FP8 dequant/padding versus generation watermarking validation and related model updates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 9, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-2", + "nodes": [ + "pull_request:43488", + "pull_request:43532" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532" + ], + "prepared_review_unit_hash": "13184443d3d1bdfe1f1e9568bb5fe84c5be14a0ca6e6fdfe0bcd33ed503d5789", + "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": "Two unrelated non-merge PRs with only an incidental overlap in a Vit test file. One is a setup/versioning repo-bot check, the other is a workflow/CI diff demo touching multiple automation files.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "pull_request:43488 is the more self-contained and clearly scoped artifact; its change is centered on setup/versioning and a small test tweak, making it the cleaner representative of the cluster.", + "best_issue_reason": null, + "best_pr_reason": "pull_request:43488 is the most suitable representative because its intent is explicit and its diff is narrower and easier to triage than the broader CI/workflow-oriented PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Reject: they do not fix the same underlying problem. The overlap in tests/models/vit/test_modeling_vit.py is incidental, while the rest of the diffs target different areas (setup/versioning vs CI workflows/testing utilities)." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 10, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43612-2", + "nodes": [ + "pull_request:43612", + "pull_request:43613" + ], + "soft_pairs": [ + "pull_request:43612|pull_request:43613" + ], + "prepared_review_unit_hash": "8c40c3b9f45fa820e8bb889d3f0741cb0ee72dd968a224eb0478700b3ee36f1c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate feature PRs for different segmentation pipelines: promptable concept segmentation vs promptable visual segmentation. They overlap in docs, registration, and pipeline scaffolding, but they target different models/code paths and should not be treated as duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: 43612 and 43613 implement different pipeline features with different task docs, pipeline modules, model integrations, and tests.", + "best_issue_reason": null, + "best_pr_reason": "No single best PR for the cluster. If forced to choose a representative, 43613 is broader, but it is still a distinct change from 43612 rather than the same underlying fix/feature.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different underlying features and code paths: concept segmentation vs visual segmentation. Shared docs/registration scaffolding is not enough to merge them as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 11, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43651-2", + "nodes": [ + "pull_request:43651", + "pull_request:43663" + ], + "soft_pairs": [ + "pull_request:43651|pull_request:43663" + ], + "prepared_review_unit_hash": "59eec95683166ad01e427fa147482226fecbf2c7007e569c3acc347789bac692", + "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": "Both PRs touch `trainer.py` but address different trainer extension points: one changes gradient-accumulation loss scaling, the other extracts signature-column filtering into an overridable helper. They are not the same underlying change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43651 is the more concrete bug-fix style change, with a direct training-step behavior adjustment and clearer user impact.", + "best_issue_reason": null, + "best_pr_reason": "PR #43651 is the strongest standalone representative of the cluster because it targets a specific runtime correctness issue rather than a refactor/extension hook.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Same file and similar theme, but they fix different code paths: loss scaling vs. signature column filtering. They could not plausibly be merged as one PR without creating an unrelated mixed 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": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-2", + "nodes": [ + "pull_request:43656", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842" + ], + "prepared_review_unit_hash": "3d8f43278d2e928b21ece9f44d8659132c048757023342fffbc9de17d5a32731", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3664, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3664, + "estimated_input_tokens": 916, + "estimated_eval_tokens": 2088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related CLI type-annotation fixes in the same file, but they address different runtime failures: one broad annotation/forward-reference cleanup, the other a specific `TypeAdapter` import/annotation issue when `pydantic` is absent. They should not be treated as duplicates.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43842 is the more precise, concrete fix: it directly addresses the `TypeAdapter` NameError path by moving the import under `TYPE_CHECKING` and quoting the annotation. Its scope is narrower and more clearly aligned with the stated bug than #43656.", + "best_issue_reason": null, + "best_pr_reason": "PR #43842 is the best representative for this cluster because it cleanly describes a specific runtime NameError scenario and has a tight, self-contained code change. #43656 appears to be a broader annotation-related cleanup in the same module rather than the same exact bug.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": false, + "reason": "Same file and broadly similar theme, but the fixes are not the same underlying bug. #43656 is about annotation/forward-reference cleanup, while #43842 specifically fixes `TypeAdapter` import handling when `pydantic` is not installed. They would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43779-2", + "nodes": [ + "pull_request:43779", + "pull_request:43816" + ], + "soft_pairs": [ + "pull_request:43779|pull_request:43816" + ], + "prepared_review_unit_hash": "e7c674adcc031aa8883109a5dfc678b4407fb20b10764302092e384d031ba83a", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4362, + "estimated_input_tokens": 1091, + "estimated_eval_tokens": 2438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests appear to implement the same SwanLab integration change: adding run id/resume support in `integration_utils.py` for issue #43698, with one variant also expanding the docs. They are strong duplicate PR candidates.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43816 is the cleaner representative: it carries the same SwanLab id/resume support plus the accompanying documentation updates, making it the more complete version of the shared change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43816 is the best global representative because it appears to include the full feature change and docs for SwanLab run id/resume support, while #43779 is the narrower implementation-focused variant.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43779", + "right": "pull_request:43816", + "accept": true, + "reason": "Same underlying SwanLab callback enhancement for id/resume support in the same file and same issue target; the doc differences do not change the core code-path change, so these are duplicate PRs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target issue #43698, touch the same file, and the previews indicate the same SwanLab id/resume support with #43816 adding docs. The duplicate/mergeable verdict is conservative enough given the shared target and overlapping code change." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43996", + "pull_request:43997", + "pull_request:44007", + "pull_request:44010", + "pull_request:44013", + "pull_request:44017", + "pull_request:44018", + "pull_request:44019", + "pull_request:44044", + "pull_request:44056", + "pull_request:44059", + "pull_request:44066", + "pull_request:44068", + "pull_request:44071", + "pull_request:44072", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44085", + "pull_request:44086", + "pull_request:44098", + "pull_request:44101", + "pull_request:44116", + "pull_request:44129", + "pull_request:44154", + "pull_request:44161", + "pull_request:44722" + ], + "soft_pairs": [ + "pull_request:43996|pull_request:44085", + "pull_request:43996|pull_request:44044", + "pull_request:44066|pull_request:44085", + "pull_request:44007|pull_request:44072", + "pull_request:44072|pull_request:44722", + "pull_request:44066|pull_request:44072", + "pull_request:44013|pull_request:44044", + "pull_request:44066|pull_request:44086", + "pull_request:44066|pull_request:44071", + "pull_request:44018|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44066|pull_request:44068", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44018|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44085", + "pull_request:44018|pull_request:44086", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request: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": "543940efd02bbd59903612f7e79baa212d62f122e4cb7c8de4d48e0128954719", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84888, + "estimated_input_tokens": 21222, + "estimated_eval_tokens": 42700 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75083, + "estimated_input_tokens": 18771, + "estimated_eval_tokens": 37798 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a set of model-specific refactors to standardized output tracing/capture_outputs, not one single bug fix. The only clear duplicate-style matches are within the same concrete code paths (notably GPT-J/CodeGen and GPT-Neo), while most cross-model pairs should stay separate.", + "confidence": 0.83, + "canonical_issue_reason": "Issue #43979 is the shared upstream tracker for this output-tracing refactor work; the PRs in this cluster largely point back to it.", + "canonical_pr_reason": "PR #44722 is the best canonical representative: it is the latest, most complete GPT-J/CodeGen variant and appears to supersede earlier near-identical attempts on the same code path.", + "best_issue_reason": "No standalone issue artifact is present in the cluster; #43979 is the only common issue target and is the appropriate umbrella issue.", + "best_pr_reason": "PR #44722 is the strongest single PR in the cluster because it addresses the same concrete GPT-J/CodeGen tracing implementation as earlier duplicates, with the most evolved patch among them.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Different models and code paths; same umbrella refactor theme is not enough to merge them." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Both are output-tracing refactors, but they touch different model implementations and are not the same concrete fix." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both edit GPT-J output-tracing logic in the same file and look like the same underlying refactor." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "Different model families and distinct implementations; only the general pattern matches." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "EfficientNet and GPT-J/CodeGen are unrelated code paths, so these are not duplicates." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen versus EfficientNet are separate model refactors, not one shared change." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa V2 are separate model-specific output-tracing changes." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen and MGP-STR are unrelated implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are different code paths, despite similar decorator migration." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both are GPT-Neo tracing refactors on the same file and appear to implement the same change." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are different model implementations." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-J/CodeGen and GPT-Neo are distinct refactors for different models." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet and GPT-J/CodeGen are unrelated code paths." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are different model changes, not a single fix." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "ResNet and GPT-J are different implementations; same issue target does not make them duplicates." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo and ResNet are unrelated model refactors." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT are separate code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are unrelated model implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are different model-specific changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are not the same underlying code-path fix." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44722", + "accept": false, + "reason": "MPT and GPT-J/CodeGen are different model refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44085", + "accept": false, + "reason": "GPT-Neo and GPT-J are different concrete implementations." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are unrelated changes." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-Neo and MGP-STR are separate model paths." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44085", + "accept": false, + "reason": "MPT and GPT-J are distinct model implementations." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are unrelated beyond sharing a decorator migration theme." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR do not share the same code path." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 and GPT-J/CodeGen are different concrete refactors." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are separate model implementations." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J/CodeGen are unrelated model changes." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 are distinct implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are separate model-specific changes." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-2 and MPT are not the same concrete fix." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different model paths." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "VisualBert and LongT5 are unrelated implementations." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44161", + "accept": false, + "reason": "MPNet and LongT5 are separate model refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44017", + "accept": false, + "reason": "CVT/FNet and SegFormer are different model implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44161", + "accept": false, + "reason": "CVT/FNet and LongT5 are unrelated code paths." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are different model-specific changes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are unrelated implementations." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44074", + "accept": false, + "reason": "RegNet and TextNet are different model refactors." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44018", + "accept": false, + "reason": "RegNet and GPT-Neo are unrelated code paths." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44161", + "accept": false, + "reason": "RegNet and LongT5 are separate model changes." + }, + { + "left": "pull_request:44074", + "right": "pull_request:44161", + "accept": false, + "reason": "TextNet and LongT5 are different implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44129", + "accept": false, + "reason": "MPT and SpeechT5 are unrelated code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are separate model-specific refactors." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44018", + "accept": false, + "reason": "SqueezeBert and GPT-Neo are different implementations." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44044", + "accept": false, + "reason": "SqueezeBert and DeBERTa V2 are unrelated model changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa V2 and SpeechT5 are different code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa V2 and VITS are unrelated model refactors." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "This is the same GPT-J/CodeGen output-tracing change as the later PR, just an earlier version." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both appear to be the same GPT-J output-tracing refactor; one is an earlier duplicate of the other." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are different model implementations." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet/ResNet are unrelated model refactors." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are different concrete changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa V2 and SpeechT5 are separate model-specific changes." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa V2 and VITS are unrelated implementations." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The clustering is grounded in the packet: it correctly treats this as a set of model-specific output-tracing refactors rather than one unified bugfix, and the canonical issue/PR choices are supported by the shared issue target and the later GPT-J/CodeGen patch. The soft-edge verdicts are conservative overall, and the accepted pairs are same-code-path duplicates that look mergeable into a single PR for the same fix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 34, + "review_unit_index": 2, + "review_unit_count": 2, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43996-28", + "nodes": [ + "pull_request:43997", + "pull_request:44010", + "pull_request:44024", + "pull_request:44056", + "pull_request:44073", + "pull_request:44074", + "pull_request:44076", + "pull_request:44129" + ], + "soft_pairs": [ + "pull_request:44024|pull_request:44076", + "pull_request:44073|pull_request:44074", + "pull_request:44056|pull_request:44073", + "pull_request:44056|pull_request:44074", + "pull_request:43997|pull_request:44010", + "pull_request:44010|pull_request:44129" + ], + "prepared_review_unit_hash": "079de8624b438fcca4070b100d5bf82c0693a9f5199430662af63b9ec6208e1e", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84888, + "estimated_input_tokens": 21222, + "estimated_eval_tokens": 42700 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17126, + "estimated_input_tokens": 4282, + "estimated_eval_tokens": 8820 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "All items are model-specific refactors around standardized output tracing/capture_outputs, but they affect different architectures and code paths. They are thematically similar, not duplicates of the same underlying change.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44073 is the clearest representative of the cluster: it directly targets standardized output tracing/capture_outputs and shows the core decorator + recorder pattern in a typical model implementation.", + "best_issue_reason": null, + "best_pr_reason": "PR 44073 is the best global representative because its title and diff most clearly express the shared refactor pattern without being overly narrow or model-quirky.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Both are output-tracing refactors, but for different models (FocalNet vs ImageGPT) and different implementation details; not the same concrete change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "Same general standardized-output migration, but VisualBert and TextNet are separate model implementations with distinct refactors." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "Both touch output capture plumbing, but MPNet and VisualBert are unrelated model code paths and cannot plausibly be merged into one PR." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "Shared theme only; MPNet and TextNet are different models and the changes are not the same bug or fix." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are separate models; these are parallel refactors, not duplicate fixes." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "Both are standardized output refactors, but SqueezeBert vs SpeechT5 targets different code paths and concrete behavior." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 15, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-14", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030", + "pull_request:46438" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030", + "pull_request:44000|pull_request:46438", + "pull_request:44027|pull_request:46438", + "pull_request:44029|pull_request:46438", + "pull_request:44028|pull_request:46438", + "pull_request:43998|pull_request:46438" + ], + "prepared_review_unit_hash": "4c9f3b49f4870ebdac8dae942e6c0efc9f1dd1052b11e375f68bb6f26aaafc29", + "component_budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 44, + "serialized_chars": 43575, + "estimated_input_tokens": 10894, + "estimated_eval_tokens": 22044 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 44, + "serialized_chars": 43575, + "estimated_input_tokens": 10894, + "estimated_eval_tokens": 22044 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are thematically similar output-tracing refactors, but they target different model families and code paths, so they are not duplicates of the same underlying change.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "No issue is present; 44003 is the best representative PR because it applies the tracing refactor across two closely related Mamba-family files and shows the clearest use of the new helper infrastructure.", + "best_issue_reason": null, + "best_pr_reason": "44003 is the strongest single PR in the set to anchor the cluster, but it should remain a representative example rather than absorbing the others as duplicates.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models (timm_backbone vs upernet); same refactor pattern only." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "timm_backbone vs vision_text_dual_encoder; unrelated code paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "timm_backbone vs univnet; different model families and behavior." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "timm_backbone vs mamba/falcon_mamba; same tracing theme, not the same change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "timm_backbone vs rwkv; distinct architectures and implementations." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "MobileNetV1 vs UnivNet; different model code paths." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "MobileNetV1 vs UperNet; only the refactor style matches." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "MobileNetV1 vs DepthAnything; separate models, not one underlying bug." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "MobileNetV1 vs VisionEncoderDecoder; distinct implementations." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "MobileNetV1 vs SpeechEncoderDecoder; related naming only." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "vision_text_dual_encoder vs univnet; different model families." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "vision_text_dual_encoder vs upernet; same tracing motif, different code." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "vision_text_dual_encoder vs mamba/falcon_mamba; not the same concrete change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "vision_text_dual_encoder vs depth_anything; separate model targets." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "UnivNet vs UperNet; unrelated subsystems." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "UnivNet vs mamba/falcon_mamba; not one shared fix." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "UnivNet vs DepthAnything; different models and outputs." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "UperNet vs RWKV; distinct model families." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "UperNet vs mamba/falcon_mamba; only similar refactoring." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Mamba/FalconMamba vs CodeGen; different code paths and semantics." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "timm_backbone vs CodeGen; unrelated models." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "UnivNet vs CodeGen; separate architectures." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "UperNet vs CodeGen; not the same underlying change." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "vision_text_dual_encoder vs CodeGen; different model families." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "DepthAnything vs VisionEncoderDecoder; distinct refactors on different models." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "DepthAnything vs SpeechEncoderDecoder; separate code paths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "DepthAnything vs SuperPoint; unrelated implementations." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "DepthAnything vs RWKV; different model families." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "DepthAnything vs DPR; not the same concrete bug/change." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "VisionEncoderDecoder vs RWKV; unrelated subsystems." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "VisionEncoderDecoder vs SpeechEncoderDecoder; similar wrapper refactor only." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "VisionEncoderDecoder vs SuperPoint; different model stacks." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "VisionEncoderDecoder vs DPR; separate code paths." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "SpeechEncoderDecoder vs SuperPoint; distinct model implementations." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "SpeechEncoderDecoder vs RWKV; unrelated models." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "SpeechEncoderDecoder vs DPR; different functionality." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "SuperPoint vs RWKV; no shared underlying bug." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "SuperPoint vs DPR; separate code paths." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "RWKV vs DPR; unrelated model families." + }, + { + "left": "pull_request:44000", + "right": "pull_request:46438", + "accept": false, + "reason": "Model refactor PR vs modular-conversion tooling; completely different domain." + }, + { + "left": "pull_request:44027", + "right": "pull_request:46438", + "accept": false, + "reason": "SpeechEncoderDecoder model change vs conversion script optimization; unrelated." + }, + { + "left": "pull_request:44029", + "right": "pull_request:46438", + "accept": false, + "reason": "RWKV model refactor vs build/tooling speedup; not the same change." + }, + { + "left": "pull_request:44028", + "right": "pull_request:46438", + "accept": false, + "reason": "SuperPoint model refactor vs modular conversion pool reuse; different areas." + }, + { + "left": "pull_request:43998", + "right": "pull_request:46438", + "accept": false, + "reason": "timm_backbone model refactor vs tooling optimization; 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": 16, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-2", + "nodes": [ + "pull_request:44662", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401" + ], + "prepared_review_unit_hash": "ed413d7592f76f70127184f39671161c7ffc6b3409d570471205d790eac9865a", + "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 are two separate model-addition PRs (PenguinVL vs Voxtral TTS). They share framework/auto-registration files, but they target different models and code paths, so they should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 44662 is the stronger cluster representative: it is a fuller, more mature implementation with substantial review activity and is not a draft. PR 45401 is a separate Voxtral TTS addition, not the same change.", + "best_issue_reason": null, + "best_pr_reason": "PR 44662 is the best overall PR in this set as a representative artifact, but only because it is the more developed standalone model implementation; it is not a duplicate of PR 45401.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different underlying features: PenguinVL implementation vs Voxtral TTS support. Shared registration/docs files are generic framework 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": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-4", + "nodes": [ + "pull_request:44979", + "pull_request:45363", + "pull_request:45666", + "pull_request:46339" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363", + "pull_request:45363|pull_request:46339", + "pull_request:45666|pull_request:46339" + ], + "prepared_review_unit_hash": "d002d9b9d668bd7c75fe82ca7225f7c277eee45841f0121ed5e496efc52604e6", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 9632, + "estimated_input_tokens": 2408, + "estimated_eval_tokens": 5072 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 3, + "serialized_chars": 9632, + "estimated_input_tokens": 2408, + "estimated_eval_tokens": 5072 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs form a related feature line around KernelConfig-based n-to-1 kernel fusion, but only the closest pair looks like the same concrete change path. The module_fusion API PR is foundational, not a duplicate of the KernelConfig fusion work.", + "confidence": 0.61, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46339 is the broadest and most mature representative of the KernelConfig fusion line: it touches core loading, hub kernel integration, model loading, and tests, and its title suggests a later simplification/extension of the same mechanism.", + "best_issue_reason": null, + "best_pr_reason": "PR 46339 is the best single representative for the cluster because it appears to subsume earlier KernelConfig fusion work and exercises the same kernel-loading path most comprehensively.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "44979 introduces the Module Fusion API scaffolding, while 45363 implements n-to-1 kernel fusion via KernelConfig. They are related, but not the same underlying change." + }, + { + "left": "pull_request:45363", + "right": "pull_request:46339", + "accept": false, + "reason": "Both are about KernelConfig fusion, but 46339 is a later extension/simplification that changes the core loading path and does not clearly duplicate the exact feature change in 45363." + }, + { + "left": "pull_request:45666", + "right": "pull_request:46339", + "accept": true, + "reason": "These look like the same concrete n-to-1 KernelConfig fusion change family, with heavy overlap in kernel-loading files; 46339 reads like a follow-on refinement of the same implementation path." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays within the evidence in the packet, and the soft-edge calls are appropriately cautious overall. The only potentially close pair, 45666 vs 46339, is supported by the shared KernelConfig / n-to-1 fusion theme and overlapping kernel-loading files, while 44979 vs 45363 is clearly more foundational than duplicate. The representative PR choice is also grounded in the broader file coverage and the later 'extended & simplified' title of 46339." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 18, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45086-2", + "nodes": [ + "pull_request:45086", + "pull_request:45317" + ], + "soft_pairs": [ + "pull_request:45086|pull_request:45317" + ], + "prepared_review_unit_hash": "2256876bb6709f19d139d1af2390bd7e8c7606af4b76007cf107e64baf2b9dbf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same Mistral regex patching AttributeError in the same code path and target the same issue; the later PR adds test coverage but does not change the underlying fix.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45317 is the better canonical PR because it contains the same fix as #45086 plus tests, making it the more complete implementation.", + "best_issue_reason": null, + "best_pr_reason": "PR #45317 is the strongest standalone candidate: same bug fix, same target issue, and it includes validation coverage in tests.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same underlying AttributeError fix in _patch_mistral_regex, same file and issue target; the second PR is a superset with tests, so they are mergeable as one change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45115-2", + "nodes": [ + "pull_request:45115", + "pull_request:45569" + ], + "soft_pairs": [ + "pull_request:45115|pull_request:45569" + ], + "prepared_review_unit_hash": "38307df2659058477dfb0cd2b5fc856796fb1fe8e60d01e128b3ac350f2117e5", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6066, + "estimated_input_tokens": 1517, + "estimated_eval_tokens": 3290 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two related but distinct NemotronH refactors: one focuses on inheriting GraniteMoeHybrid-style weight/config plumbing, the other splits NemotronH into dense/sparse model families with docs and auto-mapping updates. They overlap in shared NemotronH files but do not look like the same mergeable code change.", + "confidence": 0.86, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45569 is the broader and more complete NemotronH architecture change, covering the new dense/sparse model split plus docs and auto-mapping updates; PR 45115 is a narrower inheritance/refactor step.", + "best_issue_reason": null, + "best_pr_reason": "PR 45569 is the best representative of the cluster because it addresses the higher-level NemotronH redesign end-to-end, while PR 45115 is a partial implementation focused on conversion and inheritance plumbing.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45115", + "right": "pull_request:45569", + "accept": false, + "reason": "They share NemotronH infrastructure files, but one is a GraniteMoeHybrid inheritance/conversion refactor and the other is a broader dense/sparse model split. They are related, not the same concrete change and would not plausibly be merged as one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45332-2", + "nodes": [ + "pull_request:45332", + "pull_request:45333" + ], + "soft_pairs": [ + "pull_request:45332|pull_request:45333" + ], + "prepared_review_unit_hash": "553acd11d5260b7dfd6ee716dff322970eb40b4e65ee5c20697e415706fecebe", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5465, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5465, + "estimated_input_tokens": 1367, + "estimated_eval_tokens": 2990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related feature PRs for heterogeneous support, but they implement different layers: one adds per-layer config handling, the other adds model/cache integration. They share helper files, yet they are not the same underlying change.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "Pull request 45332 is the broader, more representative feature PR: it covers heterogeneous model support end-to-end and incorporates the downstream runtime changes that depend on per-layer config handling.", + "best_issue_reason": null, + "best_pr_reason": "PR 45332 is the better cluster representative because it is the more complete implementation and the one that would most naturally subsume the related config work, whereas 45333 is a narrower prerequisite/helper change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Related but distinct changes: 45333 adds per-layer configuration plumbing, while 45332 adds model/cache support and heterogeneous runtime behavior. They are not the same concrete bug/change and should not be merged as duplicates." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45415-2", + "nodes": [ + "pull_request:45415", + "pull_request:45425" + ], + "soft_pairs": [ + "pull_request:45415|pull_request:45425" + ], + "prepared_review_unit_hash": "c56df55e76430c8be3eaef468d0d3d05d7bc837aa874c16aef6a1f83a22f9a2d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5513, + "estimated_input_tokens": 1379, + "estimated_eval_tokens": 3014 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both items are typing-related PRs in transformers, but they address different scopes: one is a broad repo-wide type-checking pass, the other adds specific typing helpers and related adjustments. They are not the same underlying change.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the more focused and coherent typing change: it introduces shared typing helpers and updates a small set of type-sensitive call sites. It is a better representative than the broad sweep in 45415.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the best single PR in the cluster because it is narrower, more self-contained, and easier to treat as a distinct change. PR 45415 is a large multi-file type-checking cleanup rather than the same fix.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Do not merge: 45415 is a broad type-checking sweep across many files, while 45425 adds typing aliases/protocols and a few targeted fixes. They share a theme but not the same concrete bug or change, and they would not naturally combine 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": 22, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45513-2", + "nodes": [ + "pull_request:45513", + "pull_request:46084" + ], + "soft_pairs": [ + "pull_request:45513|pull_request:46084" + ], + "prepared_review_unit_hash": "a0ce8f836cf5ff9d64f7134dfb48590ee1000ada20d97c29ea62de7680dd1a80", + "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": "These are two unrelated pull requests fixing different cache-forward bugs in different model families: one in Qwen3.5/Qwen3 Next/Olmo Hybrid linear attention, the other in Mamba2 CUDA kernels. They should not be clustered as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46084 is the better canonical PR for this cluster only because it is narrowly scoped, has an explicit issue target (#46032), and the diff clearly addresses one concrete seq_len>1 cache path in Mamba2.", + "best_issue_reason": null, + "best_pr_reason": "PR #46084 is the strongest standalone PR: it targets a specific bug, has a clear issue reference, and the change is tightly limited to one code path with accompanying tests. PR #45513 is broader and spans multiple model files, making it less suitable as the canonical representative.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45513", + "right": "pull_request:46084", + "accept": false, + "reason": "Different underlying bugs in different architectures. #45513 fixes multi-token cached forward behavior for GDN linear attention in Qwen/Olmo models, while #46084 fixes use_cache with seq_len > 1 in Mamba2 CUDA kernels. 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": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45548-2", + "nodes": [ + "pull_request:45548", + "pull_request:45662" + ], + "soft_pairs": [ + "pull_request:45548|pull_request:45662" + ], + "prepared_review_unit_hash": "b3d8e14d3def9024d270bb325546187b11e78a650b513d268c2db44ccda8a5a3", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4508, + "estimated_input_tokens": 1127, + "estimated_eval_tokens": 2510 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are related through expert parallelism, but they address different bugs: #45548 fixes EP + DeepSpeed ZeRO-3 loading via accelerate launch by serializing distributed_config and adding an EP flag, while #45662 fixes EP + FSDP2 expert parameters being overwritten during distributed loading. They should not be merged as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45662 is the better representative of the cluster because it fixes the more concrete corruption path in distributed loading and spans the broader EP/FSDP2 code path, whereas #45548 is a narrower DeepSpeed ZeRO-3/accelerate configuration serialization fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45662 is the strongest standalone PR here: it targets a specific runtime bug where expert parameters can be silently overwritten, and its changes are tightly coupled to that failure mode.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45548", + "right": "pull_request:45662", + "accept": false, + "reason": "Different underlying bugs. #45548 is about preserving distributed_config and exposing EP state for DeepSpeed ZeRO-3 loading via accelerate; #45662 fixes expert weights being overwritten during FSDP2/distributed loading. Shared EP terminology and one common file are 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": 24, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45639-2", + "nodes": [ + "pull_request:45639", + "pull_request:45645" + ], + "soft_pairs": [ + "pull_request:45639|pull_request:45645" + ], + "prepared_review_unit_hash": "3f948ad5bc91207759e6db2058f97662838f178d67970f698c0651fbe96dcf34", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4590, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4590, + "estimated_input_tokens": 1148, + "estimated_eval_tokens": 2552 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same xdist-related captured_info collision in `testing_utils.py`; PR 45645 is the broader, more complete fix that also updates CI artifact collection, while 45639 is the narrower core-code change.", + "confidence": 0.89, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45645 is the better canonical PR because it covers the core xdist-safe output-file change in `src/transformers/testing_utils.py` and also updates the CI workflow to collect all split `captured_info*.txt` artifacts, making it the more complete end-to-end fix.", + "best_issue_reason": null, + "best_pr_reason": "PR 45645 is the best representative of the cluster since it addresses the same underlying bug as 45639 but includes the surrounding CI/log-handling pieces needed for the fix to work in practice.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45639", + "right": "pull_request:45645", + "accept": true, + "reason": "Accept: both PRs target the same captured_info/xdist collision problem in `testing_utils.py` and share the same issue target; 45645 extends the same fix rather than introducing a different change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45670-3", + "nodes": [ + "pull_request:45670", + "pull_request:46320", + "pull_request:46401" + ], + "soft_pairs": [ + "pull_request:45670|pull_request:46320", + "pull_request:46320|pull_request:46401" + ], + "prepared_review_unit_hash": "e722c74e86b11169ba4f3b942d3e31d2ac7d4050b2a8ddbd59d813600f0d10bc", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 7030, + "estimated_input_tokens": 1758, + "estimated_eval_tokens": 3772 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 7030, + "estimated_input_tokens": 1758, + "estimated_eval_tokens": 3772 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Only PRs 46320 and 46401 are superficially similar, but they target different model families and code paths, so this cluster does not contain a true duplicate change set. PR 45670 is unrelated to the attention-dispatch work.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46320 is the best representative of the cluster\u2019s only near-duplicate theme: a concrete attn_implementation dispatch implementation with code changes, conversion mapping updates, docs, and tests.", + "best_issue_reason": null, + "best_pr_reason": "PR 46320 is the strongest standalone representative because it is the most complete dispatch-support change and is closest in shape to the other similar PR, while still being clearly distinct from the unrelated glmasr mapping tweak.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45670", + "right": "pull_request:46320", + "accept": false, + "reason": "These are different changes: 45670 adds model mapping entries for glmasr/dia, while 46320 adds RecurrentGemma attention-dispatch support across model code, conversion mapping, docs, and tests. Not the same bug or change." + }, + { + "left": "pull_request:46320", + "right": "pull_request:46401", + "accept": false, + "reason": "Both are attn_implementation dispatch work, but for different architectures and code paths (RecurrentGemma vs BLIP-2/InstructBLIP/InstructBLIPVideo). They are not the same underlying fix and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 26, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45720-2", + "nodes": [ + "pull_request:45720", + "pull_request:46259" + ], + "soft_pairs": [ + "pull_request:45720|pull_request:46259" + ], + "prepared_review_unit_hash": "113785978077212c586b4c67f049040fd602df1024789239f35a1be136efb0fb", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7189, + "estimated_input_tokens": 1798, + "estimated_eval_tokens": 3852 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7189, + "estimated_input_tokens": 1798, + "estimated_eval_tokens": 3852 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two pull requests are not duplicates: one implements a linter rule for model forward signatures, while the other adds a different modeling rule about consumed image/video processor flags. They overlap only at the level of shared transformer model files.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45720 is the cleaner representative: it has a focused, concrete change set around forward signatures and return handling, with title and diff aligned.", + "best_issue_reason": null, + "best_pr_reason": "PR 45720 is the best representative of the cluster because its scope is narrower and more internally consistent than the other PR\u2019s broader rule/docs update.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45720", + "right": "pull_request:46259", + "accept": false, + "reason": "Different underlying changes: TRF003 forward-signature/return_dict cleanup versus a separate TRF016 rule for processor boolean flags. Shared files are incidental, not the same bug or 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": 27, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "cache", + "cache_hit": true, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45737-2", + "nodes": [ + "pull_request:45737", + "pull_request:45738" + ], + "soft_pairs": [ + "pull_request:45737|pull_request:45738" + ], + "prepared_review_unit_hash": "cc2ade73a49ec6a8883a7a24c433988645588bcd049e8aad2bce56b1b24daa1c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4131, + "estimated_input_tokens": 1033, + "estimated_eval_tokens": 2322 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs address the same MusicgenMelody cache initialization bug in `forward()`: replacing `EncoderDecoderCache(DynamicCache, DynamicCache)` with a single `DynamicCache`. PR #45737 is the more complete duplicate because it bundles the fix with a regression test, while #45738 is the fix-only variant.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45737 is the best canonical PR because it contains the same code fix as #45738 plus a regression test, making it the most complete representation of the bug fix.", + "best_issue_reason": null, + "best_pr_reason": "PR #45737 is the strongest overall candidate: it fixes the concrete cache bug and adds a targeted regression test, whereas #45738 only changes the implementation.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45737", + "right": "pull_request:45738", + "accept": true, + "reason": "Same underlying code-path bug in `musicgen_melody` cache setup; the diffs match on the core fix and the test-only addition in #45737 could plausibly live in the same PR as #45738." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The assessment is grounded in the packet: both PRs make the same `musicgen_melody` cache initialization change, and #45737 only adds a regression test on top of that fix. The soft-edge accept is conservative enough because the diffs are mergeable into one PR for the same bugfix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 28, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-45932-2", + "nodes": [ + "pull_request:45932", + "pull_request:46100" + ], + "soft_pairs": [ + "pull_request:45932|pull_request:46100" + ], + "prepared_review_unit_hash": "71f34ae2890f2e2718cb4e1085855949a3a7ef770a2429d972641675a0f1227f", + "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": "The two PRs are unrelated: one is a broad testing/infra and model-maintenance change, while the other is a narrow ViT test edit that adds intentional failing assertions. They do not appear to fix the same bug or implement the same change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:45932", + "right": "pull_request:46100", + "accept": false, + "reason": "Different underlying intent and code paths: PR 45932 touches workflow/secrets and multiple model files, while PR 46100 only changes one ViT test file and appears to be test-migration noise. They are not plausibly 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": 29, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46073-3", + "nodes": [ + "pull_request:46073", + "pull_request:46154", + "pull_request:46196" + ], + "soft_pairs": [ + "pull_request:46073|pull_request:46154", + "pull_request:46073|pull_request:46196", + "pull_request:46154|pull_request:46196" + ], + "prepared_review_unit_hash": "21ddcc6283bb2aa50e105116b3989f412b5194e7858a75114a4966bb8b1ec8b8", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8083, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 8083, + "estimated_input_tokens": 2021, + "estimated_eval_tokens": 4298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related in theme only: all add SDPA dispatch support, but each PR targets a different model family and distinct code paths, so they are not duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical duplicate PR: CTRL, T5, and the wav2vec2/SeamlessM4T work are separate model-specific SDPA implementations.", + "best_issue_reason": null, + "best_pr_reason": "PR 46196 is the strongest representative of the theme because it covers multiple related speech models and includes broader code/test updates, making it the most complete SDPA-dispatch change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46073", + "right": "pull_request:46154", + "accept": false, + "reason": "Different architectures and code paths: CTRL vs T5. Same feature theme, but not the same concrete bug/change and not mergeable into one PR." + }, + { + "left": "pull_request:46073", + "right": "pull_request:46196", + "accept": false, + "reason": "CTRL support is unrelated to wav2vec2/SeamlessM4T speech-model dispatch internals. Broad SDPA theme only, not a duplicate fix." + }, + { + "left": "pull_request:46154", + "right": "pull_request:46196", + "accept": false, + "reason": "T5 seq2seq attention dispatch and speech-model dispatch are separate implementations in different model families; they are 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": 30, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46126-2", + "nodes": [ + "pull_request:46126", + "pull_request:46269" + ], + "soft_pairs": [ + "pull_request:46126|pull_request:46269" + ], + "prepared_review_unit_hash": "e29b30697c2e410af6697160a7ca276bc1d89064aee8659693586e28dfe98db5", + "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": "Two open PRs overlap in distributed infrastructure files, but they target different levels of work: one is a focused EP + Trainer integration, the other is a broad distributed branch base sweep. They are not the same change and should not be merged as duplicates.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #46269 is the broader umbrella change: it spans many distributed subsystems and looks like the base branch integration point rather than a narrow feature patch, making it the more canonical representative of this cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR #46269 is the best single representative because it covers the larger distributed refactor/integration surface and subsumes the shared files without being narrowly scoped to one feature path.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46126", + "right": "pull_request:46269", + "accept": false, + "reason": "Overlap is only at the subsystem/file level. #46126 is a specific EP + Trainer integration, while #46269 is a broad distributed base PR touching many unrelated areas; they do not appear to fix the same concrete code-path problem and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 31, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46138-2", + "nodes": [ + "pull_request:46138", + "pull_request:46432" + ], + "soft_pairs": [ + "pull_request:46138|pull_request:46432" + ], + "prepared_review_unit_hash": "15ef53debc7b97ac1fc6e9fe62711fe7a118a0499503fbc1270b7f193ce7e49e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3863, + "estimated_input_tokens": 966, + "estimated_eval_tokens": 2188 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 3863, + "estimated_input_tokens": 966, + "estimated_eval_tokens": 2188 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two pull requests are not duplicates: one changes a workflow invocation style in self-comment CI, while the other hardens ssh-runner by pinning a downloaded script and clearing HF_TOKEN before SSH.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:46138", + "right": "pull_request:46432", + "accept": false, + "reason": "Different workflows and different underlying changes: stdin piping for a CI helper vs security/hardening updates for ssh-runner. They do not fix the same code-path problem and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 32, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46152-2", + "nodes": [ + "pull_request:46152", + "pull_request:46207" + ], + "soft_pairs": [ + "pull_request:46152|pull_request:46207" + ], + "prepared_review_unit_hash": "05f4f61fa02ba2c1df9d614e5a7193150b18ccc39e4c25c02b4ca4df708cc6cf", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4276, + "estimated_input_tokens": 1069, + "estimated_eval_tokens": 2394 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are about avoiding brittle file-open/source inspection, but they change different code paths in different modules and address different implementation details. They are related thematically, not duplicates.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 46207 is the more concrete exemplar: it tightens one specific source-reading heuristic in modeling_utils and includes tests, making it the cleaner representative of the cluster theme.", + "best_issue_reason": null, + "best_pr_reason": "PR 46207 is the best single PR in this set because it has the clearest narrowly-scoped fix and test coverage. PR 46152 is a separate utility-level refactor in import_utils, not the same concrete bug.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46152", + "right": "pull_request:46207", + "accept": false, + "reason": "Different files, different helpers, and different behaviors: one adds Traversable resolution for importing package source, the other replaces open() with inspect.getsource for attention-implementation detection. Shared motivation 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": 33, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46337-3", + "nodes": [ + "pull_request:46337", + "pull_request:46362", + "pull_request:46368" + ], + "soft_pairs": [ + "pull_request:46337|pull_request:46368", + "pull_request:46337|pull_request:46362", + "pull_request:46362|pull_request:46368" + ], + "prepared_review_unit_hash": "991b266e7fb2dacce86caec08ac65d505141e33a2cd08976917fe0c7f27c5a43", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5964, + "estimated_input_tokens": 1491, + "estimated_eval_tokens": 3238 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5964, + "estimated_input_tokens": 1491, + "estimated_eval_tokens": 3238 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate test-only pull requests for different model families and different failure modes. The only commonality is that they all touch model test code and mention quant/cache or device-map related behavior; they do not look like duplicates or one shared fix.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "pull_request:46337", + "right": "pull_request:46368", + "accept": false, + "reason": "Different model targets and fixes: Cohere2Moe adjusts model-test settings for offload/device-map splitting, while NemotronH skips an invalid quantized-cache test. No shared code path or mergeable single fix." + }, + { + "left": "pull_request:46337", + "right": "pull_request:46362", + "accept": false, + "reason": "Cohere2Moe test infrastructure tweak vs. DeepSeek V4 quant-cache skip. They address unrelated model-specific test failures and do not appear to be the same underlying bug." + }, + { + "left": "pull_request:46362", + "right": "pull_request:46368", + "accept": false, + "reason": "Both skip quant-cache tests, but for different architectures and different incompatibilities (DeepSeek V4 compressor cache vs. NemotronH hybrid cache). Same broad subsystem, 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": "pull_request", + "component_index": 34, + "component_count": 34, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-46376-2", + "nodes": [ + "pull_request:46376", + "pull_request:46417" + ], + "soft_pairs": [ + "pull_request:46376|pull_request:46417" + ], + "prepared_review_unit_hash": "0c64f31b0f74f743ffcc2e14cef82e92a1c2a39e0183d50d37b17ef2b7e8219b", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5960, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5960, + "estimated_input_tokens": 1490, + "estimated_eval_tokens": 3236 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two unrelated open PRs: one is a draft for multimodal continuous batching infrastructure, the other adds a new HunYuan VL model and its docs/tests. They do not look like duplicate fixes or mergeable variants of the same change.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the two PRs target different problems and codepaths, so there is no duplicate representative.", + "best_issue_reason": null, + "best_pr_reason": "PR #46417 is the better standalone representative because it is a complete, non-draft feature addition with docs, model code, and tests; #46376 is still a draft infrastructure refactor.", + "soft_edge_verdicts": [ + { + "left": "pull_request:46376", + "right": "pull_request:46417", + "accept": false, + "reason": "Different changes in different areas: continuous batching multimodal support versus adding a new HunYuan VL model. They do not fix the same concrete bug or share a mergeable codepath." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/manifest.json b/snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..b96e69975267c1f95fe7ae60a5d442b71cfedf92 --- /dev/null +++ b/snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260605t180026z", + "artifacts": { + "hybrid": "snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-06-05T18:11:22Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260605T180026Z", + "variant": "hybrid" +} diff --git a/snapshots/20260605T180026Z/manifest.json b/snapshots/20260605T180026Z/manifest.json index 8b884202ebf1445e18135dd4a4b2647fc4eb73ce..0091329562070998a4e7794f34692afd715e5766 100644 --- a/snapshots/20260605T180026Z/manifest.json +++ b/snapshots/20260605T180026Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 1458 }, "extracted_at": "2026-06-05T18:00:26Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260605t180026z", + "runs": { + "hybrid-model-20260605t180026z": { + "analysis_id": "hybrid-model-20260605t180026z", + "artifacts": { + "hybrid": "snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260605T180026Z/analysis-runs/hybrid-model-20260605t180026z/manifest.json", + "model": null, + "published_at": "2026-06-05T18:11:22Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260605T180026Z", "watermark": {