diff --git a/analysis/current/analysis-report-hybrid.json b/analysis/current/analysis-report-hybrid.json index 8d1faa3376111c73629504d27d70b41bf406a683..aeeb3aaa5ccc6b2c9c61bbcca6f8cce9e91703a3 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": "20260418T170534Z", - "generated_at": "2026-04-18T18:16:19Z", + "snapshot_id": "20260420T094431Z", + "generated_at": "2026-04-20T09:57:23Z", "evidence_quality": "full", "llm_enrichment": true, "meta_bugs": [ @@ -1353,7 +1353,7 @@ "status": "open", "confidence": 0.8, "canonical_issue_number": 43366, - "canonical_pr_number": 43757, + "canonical_pr_number": 45506, "issue_numbers": [ 43366 ], @@ -1417,7 +1417,7 @@ "left_pr_number": 43757, "right_pr_number": 45506, "code_similarity": 0.096, - "size_similarity": 0.149, + "size_similarity": 0.15, "file_overlap": 0.2, "area_overlap": 0.005, "patch_similarity": 0.031, @@ -1461,11 +1461,11 @@ { "left_pr_number": 45500, "right_pr_number": 45506, - "code_similarity": 1.0, - "size_similarity": 1.0, + "code_similarity": 0.997, + "size_similarity": 0.996, "file_overlap": 1.0, - "area_overlap": 1.0, - "patch_similarity": 1.0, + "area_overlap": 0.996, + "patch_similarity": 0.996, "shared_filenames": [ "docs/source/en/model_doc/gpt_oss.md", "src/transformers/integrations/ggml.py", @@ -1569,7 +1569,117 @@ ], [ 386, - 406 + 405 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45341-4", + "summary": "Cluster of 1 issues and 3 PRs centered on issue #45341.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45341, + "canonical_pr_number": 45351, + "issue_numbers": [ + 45341 + ], + "pr_numbers": [ + 45351, + 45509, + 45515 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45351, + "right_pr_number": 45509, + "code_similarity": 0.427, + "size_similarity": 0.308, + "file_overlap": 1.0, + "area_overlap": 0.148, + "patch_similarity": 0.093, + "shared_filenames": [ + "src/transformers/testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3207, + 3230 + ] + ], + "right_ranges": [ + [ + 3204, + 3210 + ] + ] + } + ] + }, + { + "left_pr_number": 45351, + "right_pr_number": 45515, + "code_similarity": 0.437, + "size_similarity": 0.359, + "file_overlap": 1.0, + "area_overlap": 0.148, + "patch_similarity": 0.091, + "shared_filenames": [ + "src/transformers/testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3207, + 3230 + ] + ], + "right_ranges": [ + [ + 3204, + 3210 + ] + ] + } + ] + }, + { + "left_pr_number": 45509, + "right_pr_number": 45515, + "code_similarity": 0.946, + "size_similarity": 0.857, + "file_overlap": 1.0, + "area_overlap": 1.0, + "patch_similarity": 0.833, + "shared_filenames": [ + "src/transformers/testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3204, + 3210 + ] + ], + "right_ranges": [ + [ + 3204, + 3210 ] ] } @@ -2042,10 +2152,10 @@ }, { "cluster_id": "cluster-43366-4", - "canonical_pr_number": 43757, + "canonical_pr_number": 45506, "duplicate_pr_numbers": [ - 45500, - 45506 + 43757, + 45500 ], "target_issue_number": 43366, "reason": "PRs in cluster-43366-4 are treated as duplicates because they converge on issue #43366 with closing_reference, shared_issue_target, soft_similarity evidence." @@ -2103,6 +2213,16 @@ ], "target_issue_number": 45081, "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45341-4", + "canonical_pr_number": 45351, + "duplicate_pr_numbers": [ + 45509, + 45515 + ], + "target_issue_number": 45341, + "reason": "PRs in cluster-45341-4 are treated as duplicates because they converge on issue #45341 with closing_reference, shared_issue_target, 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 de2aa5f979693f3fc192d6939ec514cdd96ac82f..10e71a151296dc202e65a8dbaf38c5ac3971b560 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": "20260418T170534Z", - "generated_at": "2026-04-18T18:16:19Z", + "snapshot_id": "20260420T094431Z", + "generated_at": "2026-04-20T09:57:23Z", "model": "gpt-5.4-mini?service_tier=flex", "reviews": [ { @@ -26,12 +26,12 @@ "issue:13244|issue:16618", "issue:15980|issue:16618" ], - "prepared_review_unit_hash": "dcd9f595a958e37b4890a6765530239dc1f6801a44b3ae39d5bee95b8f597333", + "prepared_review_unit_hash": "def90bc16496d6f679e8ab2e49a31cde60469c2df424e68eb6615a9ca14b9d26", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2477, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -39,7 +39,7 @@ "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 2479, + "serialized_chars": 2477, "estimated_input_tokens": 620, "estimated_eval_tokens": 1496 }, @@ -48,24 +48,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is weakly connected: two issues are about model loading/auth errors, while the Tapas tokenization issue is unrelated. None of the soft edges look like true duplicates.", - "confidence": 0.92, - "canonical_issue_reason": "issue #16618 is the most general and representative model-loading complaint in the set; it is broader than the private-model login error in #15980 and far closer to the apparent cluster theme than the Tapas tokenization issue.", + "summary": "No duplicate cluster here: one issue is about Tapas tokenization, one about private-model auth/download messaging, and one about a generic model-loading failure. The only soft similarities are at the subsystem level, not the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": "issue #16618 is the broadest model-loading report in the set and is the closest representative of the only related pair; the other issues describe different symptoms and causes.", "canonical_pr_reason": null, - "best_issue_reason": "#16618 is the best single issue to anchor this cluster because it is the broadest user-facing failure report and has the strongest discussion activity among the model-loading reports.", + "best_issue_reason": "#16618 is the most general and user-facing of the issues, so it is the best single representative if one must be chosen.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:13244", "right": "issue:16618", "accept": false, - "reason": "Different problems: Tapas tokenization vs. failing to load bert-base-uncased. They do not describe the same code path or bug." + "reason": "Tapas tokenization mismatch vs. a BERT model-loading failure are different bugs in different code paths; shared ML infrastructure is too broad to treat as duplicate." }, { "left": "issue:15980", "right": "issue:16618", "accept": false, - "reason": "#15980 is specifically about a bad auth/error message when downloading a private model, while #16618 is a general model-load failure for a public model. Related area, but not the same underlying issue." + "reason": "Both involve model download/load behavior, but #15980 is specifically an unauthorized private-model download error while #16618 is a generic 'can't load model' report for bert-base-uncased; not the same underlying issue." } ] }, @@ -87,7 +87,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:24643", "issue:30277", @@ -108,7 +108,6 @@ "issue:43425", "issue:43531", "issue:43541", - "issue:43572", "issue:43596", "issue:43638", "issue:43716", @@ -136,12 +135,14 @@ "issue:45161", "issue:45237", "issue:45362", - "issue:45464" + "issue:45464", + "issue:45507" ], "soft_pairs": [ "issue:44322|issue:45464", "issue:43716|issue:45237", "issue:43828|issue:45237", + "issue:45237|issue:45507", "issue:44292|issue:45237", "issue:30277|issue:35545", "issue:35545|issue:42915", @@ -165,12 +166,13 @@ "issue:43638|issue:45161", "issue:34928|issue:36331", "issue:43596|issue:45161", + "issue:43716|issue:45507", "issue:39290|issue:45081", "issue:43828|issue:45161", "issue:43425|issue:44292", - "issue:43541|issue:43716", - "issue:44918|issue:45161", "issue:34928|issue:41720", + "issue:44918|issue:45161", + "issue:43541|issue:43716", "issue:43828|issue:44292", "issue:43716|issue:44292", "issue:44292|issue:44918", @@ -185,363 +187,362 @@ "issue:34928|issue:36010", "issue:41762|issue:43596", "issue:43901|issue:44512", - "issue:43039|issue:44560", "issue:39290|issue:41720", + "issue:43039|issue:44560", "issue:43866|issue:44863", "issue:43854|issue:43866", "issue:43425|issue:45070", "issue:39290|issue:43531", "issue:41093|issue:43638", - "issue:43792|issue:45070", - "issue:45362|issue:45464", "issue:43404|issue:43866", + "issue:45362|issue:45464", + "issue:43792|issue:45070", "issue:43541|issue:43828", "issue:36010|issue:39290", "issue:44841|issue:45084", "issue:41093|issue:41762", - "issue:43572|issue:45070", "issue:43296|issue:43366" ], - "prepared_review_unit_hash": "66348aa2f72ea46bd5794182e4f1caa120f33db8ab630a81317b381a97b91da1", + "prepared_review_unit_hash": "94e914bca7ac44e87e08ddaf6a3ada324e9127ca80320efe843f8e204003e97c", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 62, - "serialized_chars": 46190, - "estimated_input_tokens": 11548, - "estimated_eval_tokens": 23352 + "soft_pair_count": 63, + "serialized_chars": 46472, + "estimated_input_tokens": 11618, + "estimated_eval_tokens": 23492 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated issue reports; the only clear duplicate-like pair is the DeepSpeed ZeRO-3/Bert index-out-of-bounds reports. Most other soft pairs share a broad subsystem or model family but describe different failures and should stay separate.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43596 is the clearest representative of the one true duplicate cluster here: it names the concrete ZeRO-3 + Bert index-out-of-bounds failure directly and matches the strongest near-duplicate wording with issue:43638.", + "summary": "This set is mostly a grab bag of unrelated model/runtime/docs issues. The only clear duplicate family is the DeepSpeed ZeRO-3 zero-size/index-out-of-bounds loading bug around Bert/Gemma3; the rest are distinct failures across different models, APIs, and doc cleanups.", + "confidence": 0.9, + "canonical_issue_reason": "issue:43596 is the best representative of the one real duplicate family here: it states the concrete ZeRO-3 zero-size index-out-of-bounds failure clearly, and it has a near-exact duplicate (43638) plus a closely related follow-up (41762) on the same loading path.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43596 is the best issue anchor because it is specific, concrete, and the most likely canonical target for the ZeRO-3/Bert duplicate pair; the rest of the cluster is too heterogeneous to serve as a better global canonical.", + "best_issue_reason": "issue:43596 is the strongest global issue representative because it captures the shared low-level DeepSpeed ZeRO-3 loading bug in a model-agnostic way while still being specific enough to anchor the small duplicate cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44322", "right": "issue:45464", "accept": false, - "reason": "Same model family, but different failures: config attribute error vs chat/streaming API failure." + "reason": "Different failures: Qwen3.5 config attribute access vs chat/completions streaming behavior on a different path." }, { "left": "issue:43716", "right": "issue:45237", "accept": false, - "reason": "Different models and different failure modes; only a loose GPU/runtime overlap." + "reason": "Both involve model/runtime problems, but one is an autocast dtype mismatch and the other is an AMD GPU execution failure for GPT-OSS." }, { "left": "issue:43828", "right": "issue:45237", "accept": false, - "reason": "Autocast dtype mismatch is unrelated to AMD GPU support failure." + "reason": "Different concrete bugs: Phi-tiny-MoE autocast dtype mismatch vs GPT-OSS AMD GPU failure." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model/code paths: GPT-OSS on AMD GPUs vs GraniteMoEHybrid calling an invalid method." }, { "left": "issue:44292", "right": "issue:45237", "accept": false, - "reason": "Qwen-3-8B-NVFP4 runtime error is not the same bug as GPT-OSS-on-AMD." + "reason": "Different models and symptoms: Qwen-3-8B-NVFP4 runtime error vs GPT-OSS AMD GPU failure." }, { "left": "issue:30277", "right": "issue:35545", "accept": false, - "reason": "DeepSpeed collective mismatch is unrelated to ModernBERT ONNX export." + "reason": "DeepSpeed collective mismatch on Jamba-v01 is unrelated to ModernBERT ONNX export." }, { "left": "issue:35545", "right": "issue:42915", "accept": false, - "reason": "ONNX export error and FineGrainedFP8Config failure are different code paths." + "reason": "Different bugs in different models: ONNX export failure vs FineGrainedFP8Config load failure." }, { "left": "issue:35545", "right": "issue:41720", "accept": false, - "reason": "Export-to-ONNX problem vs auto device mapping CUDA assert." + "reason": "ModernBERT ONNX export and Qwen3 auto-device-map cuda assert are unrelated." }, { "left": "issue:44509", "right": "issue:44512", "accept": false, - "reason": "Both are docs regressions, but they target different removed docs/topics." + "reason": "Both are docs cleanup around v5 removals, but they concern different removed topics and are not the same concrete change." }, { "left": "issue:43638", "right": "issue:44805", "accept": false, - "reason": "Both mention IndexError, but one is ZeRO-3/Bert loading and the other is a mask shape mismatch." + "reason": "Both mention index/shape problems, but one is a ZeRO-3 Bert loading failure and the other is a different mask/tensor mismatch symptom." }, { "left": "issue:35545", "right": "issue:36010", "accept": false, - "reason": "ModernBERT ONNX export is unrelated to the GenerationMixin import error." + "reason": "ONNX export failure and GenerationMixin import error are unrelated." }, { "left": "issue:43596", "right": "issue:44805", "accept": false, - "reason": "ZeRO-3/Bert out-of-bounds loading bug is not the same as the mask-shape IndexError." + "reason": "Similar index/shape language, but the specific error and code path differ; not enough evidence for the same bug." }, { "left": "issue:44805", "right": "issue:45161", "accept": false, - "reason": "Mask/tensor shape mismatch and GPT-OSS MoE tensor parallel failure are different issues." - }, - { - "left": "issue:44805", - "right": "issue:44918", - "accept": false, - "reason": "Different failures: mask indexing error vs unpacking embeddings in TRL SFT trainer." + "reason": "Mask/tensor shape mismatch is unrelated to GPT-OSS tensor-parallel failure." }, { "left": "issue:44661", "right": "issue:44805", "accept": false, - "reason": "Model-registration failure is unrelated to the mask shape IndexError." + "reason": "Tokenizer-mapping issue and mask/tensor mismatch are different problems." }, { "left": "issue:30277", "right": "issue:43638", "accept": false, - "reason": "Collective mismatch during ZeRO-3 and Bert index-out-of-bounds are distinct failures." + "reason": "Both touch DeepSpeed, but one is a collective mismatch and the other is a zero-size index error; different bugs." }, { "left": "issue:34634", "right": "issue:35545", "accept": false, - "reason": "BarkProcessor voice preset bug is unrelated to ModernBERT export." + "reason": "BarkProcessor voice preset regression is unrelated to ModernBERT export." }, { "left": "issue:43596", "right": "issue:43638", "accept": true, - "reason": "Near-identical ZeRO-3/Bert loading bug: same IndexError family, same underlying code path, and clearly duplicate phrasing." + "reason": "Same concrete ZeRO-3 zero-size index-out-of-bounds loading bug, with nearly identical symptom and code path." }, { "left": "issue:43975", "right": "issue:44457", "accept": false, - "reason": "Detokenization bug and LoRA merge/save inconsistency are different problems." + "reason": "Different issues: incorrect detokenization vs LoRA merge/save/reload output mismatch." }, { "left": "issue:43039", "right": "issue:45161", "accept": false, - "reason": "Liger cross_entropy call and GPT-OSS tensor-parallel failure are unrelated." + "reason": "Liger Kernel cross_entropy invocation and GPT-OSS tensor-parallel failure are different code paths." }, { "left": "issue:44292", "right": "issue:44457", "accept": false, - "reason": "Qwen-3 NVFP4 runtime error is unrelated to LoRA merge/save mismatch." + "reason": "Qwen-3-8B-NVFP4 runtime error and LoRA merge/reload mismatch are unrelated." }, { "left": "issue:34928", "right": "issue:44805", "accept": false, - "reason": "Activation-checkpointing/FSDP tensor-size mismatch is not the same as the mask-shape IndexError." + "reason": "Both are tensor-shape failures, but the contexts and symptoms differ: activation checkpointing/FSDP vs mask indexing." }, { "left": "issue:24643", "right": "issue:30277", "accept": false, - "reason": "DeepSpeed training weight-dimension error and collective mismatch are different bugs." + "reason": "Different DeepSpeed-related failures: weight dimensionality check vs collective mismatch." }, { "left": "issue:43975", "right": "issue:45237", "accept": false, - "reason": "Detokenization bug is unrelated to AMD GPU runtime failure." + "reason": "Detokenization bug is unrelated to GPT-OSS AMD GPU loading/runtime failure." }, { "left": "issue:43638", "right": "issue:45161", "accept": false, - "reason": "ZeRO-3/Bert loading bug is unrelated to GPT-OSS tensor-parallel support." + "reason": "ZeRO-3 Bert loading failure is unrelated to GPT-OSS tensor-parallel support." }, { "left": "issue:34928", "right": "issue:36331", "accept": false, - "reason": "Activation-checkpointing mismatch and trainer signature change are unrelated." + "reason": "Activation checkpointing/FSDP tensor mismatch and CustomTrainer kwarg mismatch are unrelated." }, { - "left": "issue:43596", - "right": "issue:45161", + "left": "issue:43716", + "right": "issue:45507", "accept": false, - "reason": "ZeRO-3/Bert loading bug is unrelated to GPT-OSS tensor-parallel support." + "reason": "Different failures: image preprocessor dtype mismatch vs GraniteMoEHybrid invalid method call." }, { "left": "issue:39290", "right": "issue:45081", "accept": false, - "reason": "Gemma3 sliding_window attribute issue is unrelated to the tokenizer patch crash." + "reason": "Gemma3 config attribute missing in vLLM and Mistral regex patch backend_tokenizer error are unrelated." }, { "left": "issue:43828", "right": "issue:45161", "accept": false, - "reason": "Autocast dtype mismatch is unrelated to GPT-OSS tensor-parallel behavior." + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to GPT-OSS tensor-parallel failure." }, { - "left": "issue:43425", - "right": "issue:44292", + "left": "issue:44292", + "right": "issue:44918", "accept": false, - "reason": "Torch version incompatibility is not the same as the Qwen-3 NVFP4 runtime error." + "reason": "Both mention Qwen3-family models, but one is an NVFP4 runtime error and the other is a TRL SFT trainer embedding-unpacking failure." }, { - "left": "issue:43541", - "right": "issue:43716", + "left": "issue:43039", + "right": "issue:43828", "accept": false, - "reason": "Mixtral grouped_mm tracing and Mistral-3 dtype mismatch are different failures." + "reason": "Different execution bugs: Liger Kernel cross_entropy path vs autocast dtype mismatch." }, { - "left": "issue:44918", - "right": "issue:45161", + "left": "issue:30277", + "right": "issue:36331", "accept": false, - "reason": "TRL SFT embedding unpacking bug is unrelated to GPT-OSS MoE tensor parallel." + "reason": "DeepSpeed collective mismatch and CustomTrainer compute_loss signature mismatch are unrelated." }, { - "left": "issue:34928", - "right": "issue:41720", + "left": "issue:43039", + "right": "issue:44805", "accept": false, - "reason": "FSDP checkpointing mismatch and auto device mapping CUDA assert are different bugs." + "reason": "Liger Kernel cross_entropy issue is unrelated to mask/tensor shape mismatch." }, { - "left": "issue:43827", - "right": "issue:44512", + "left": "issue:34928", + "right": "issue:36010", "accept": false, - "reason": "Both are docs cleanup, but they refer to different removed APIs/commands." + "reason": "FSDP activation checkpointing mismatch and GenerationMixin import error are unrelated." }, { - "left": "issue:43716", - "right": "issue:44560", - "accept": false, - "reason": "Mistral-3 dtype mismatch is unrelated to Qwen3-vl video StopIteration." + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Same underlying DeepSpeed ZeRO-3 zero-size index-out-of-bounds loading failure, just reported on a different model." }, { - "left": "issue:43716", - "right": "issue:44918", + "left": "issue:43827", + "right": "issue:44512", "accept": false, - "reason": "Different models and different failure modes." + "reason": "Both are docs issues caused by v5 removals, but they reference different removed features and should not be merged as the same change." }, { - "left": "issue:43039", - "right": "issue:43828", + "left": "issue:43901", + "right": "issue:44512", "accept": false, - "reason": "Liger Kernel cross_entropy issue and autocast dtype mismatch are not the same bug." + "reason": "TextClassificationPipeline docs mismatch and transformers run-command docs removal are separate documentation fixes." }, { - "left": "issue:30277", - "right": "issue:36331", + "left": "issue:39290", + "right": "issue:41720", "accept": false, - "reason": "DeepSpeed collective mismatch and CustomTrainer signature incompatibility are unrelated." + "reason": "Different bugs: Gemma3 sliding-window config error vs Qwen3 auto device mapping cuda assert." }, { "left": "issue:43039", - "right": "issue:44805", + "right": "issue:44560", "accept": false, - "reason": "Liger Kernel cross_entropy and mask-shape IndexError are unrelated." + "reason": "Liger Kernel cross_entropy regression is unrelated to Qwen3-vl video StopIteration." }, { - "left": "issue:34928", - "right": "issue:36010", + "left": "issue:43866", + "right": "issue:44863", "accept": false, - "reason": "Checkpointing/FSDP bug is unrelated to GenerationMixin import failure." + "reason": "Different model-loading failures: corrupted Ovis2 checkpoint vs NemotronH implementation unable to load checkpoints." }, { - "left": "issue:41762", - "right": "issue:43638", + "left": "issue:43854", + "right": "issue:43866", "accept": false, - "reason": "Gemma3 ZeRO-3 load failure and Bert ZeRO-3 index error are different model/code paths." + "reason": "Different model/test failures: GLM-4.7-Flash unit test loading issue vs Ovis2 checkpoint corruption." }, { - "left": "issue:43901", - "right": "issue:44512", + "left": "issue:45362", + "right": "issue:45464", "accept": false, - "reason": "Docs mismatch around return_all_scores is unrelated to the run-command docs removal." + "reason": "Both involve Qwen3.5 chat use, but one is a transformers chat crash and the other is a streaming chat/completions API failure; not the same bug." }, { - "left": "issue:43039", - "right": "issue:44560", + "left": "issue:43792", + "right": "issue:45070", "accept": false, - "reason": "Liger Kernel cross_entropy issue is unrelated to the Qwen3-vl video StopIteration." + "reason": "Whisper model runtime failure is unrelated to the pydantic PretrainedConfig field breakage." }, { - "left": "issue:39290", - "right": "issue:41720", + "left": "issue:43541", + "right": "issue:43828", "accept": false, - "reason": "Gemma3 sliding_window attribute issue and Qwen3 CUDA assert are different bugs." + "reason": "Grouped_mm torch-dynamo tracing failure and autocast dtype mismatch are different execution problems." }, { - "left": "issue:43866", - "right": "issue:44863", + "left": "issue:36010", + "right": "issue:39290", "accept": false, - "reason": "Checkpoint corruption and checkpoint-loading implementation failure are not the same problem." + "reason": "GenerationMixin import error and Gemma3 sliding_window_pattern error are unrelated." }, { - "left": "issue:43854", - "right": "issue:43866", + "left": "issue:44841", + "right": "issue:45084", "accept": false, - "reason": "GLM-4.7 unit test loading failure and Ovis2 checkpoint corruption are distinct." + "reason": "Processor failure for Voxtral-Mini-3B-2507 is unrelated to the non-template-node compile error." }, { - "left": "issue:43425", - "right": "issue:45070", + "left": "issue:41093", + "right": "issue:41762", "accept": false, - "reason": "Torch compatibility issue is unrelated to the PretrainedConfig pydantic-model breakage." + "reason": "Mask/index shape mismatch during one path is not the same as the ZeRO-3 zero-size index-out-of-bounds loading error." }, { - "left": "issue:39290", - "right": "issue:43531", + "left": "issue:43404", + "right": "issue:43866", "accept": false, - "reason": "Both mention sliding_window, but Gemma3 and Qwen3-MoE are different model/config bugs." + "reason": "Mistral3 lm_head tying bug is unrelated to the Ovis2 checkpoint corruption report." }, { "left": "issue:41093", "right": "issue:43638", "accept": false, - "reason": "Mask-shape mismatch and ZeRO-3/Bert index error are different manifestations, not the same concrete bug." + "reason": "Mask/tensor shape mismatch and ZeRO-3 Bert index-out-of-bounds are different bugs." }, { - "left": "issue:43792", + "left": "issue:43425", "right": "issue:45070", "accept": false, - "reason": "Whisper load failure is unrelated to the PretrainedConfig pydantic regression." + "reason": "Torch 2.10 incompatibility is unrelated to the pydantic PretrainedConfig regression." }, { - "left": "issue:45362", - "right": "issue:45464", + "left": "issue:39290", + "right": "issue:43531", "accept": false, - "reason": "Both involve Qwen3.5 chat paths, but one is a crash in chat usage and the other is a streaming inference API failure." + "reason": "Different Qwen/Gemma sliding-window problems with different model families and symptoms." }, { - "left": "issue:43404", - "right": "issue:43866", + "left": "issue:43296", + "right": "issue:43366", "accept": false, - "reason": "Mistral3 lm_head tying bug and Ovis2 checkpoint corruption are unrelated." + "reason": "PaddleOCR-VL loading failure and GGUF gpt-oss support request are unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The assessment is grounded in the packet. The chosen canonical issue (issue:43596) is well-supported by the near-identical wording with issue:43638, and the summary stays conservative about the rest of the cluster being heterogeneous. The soft-edge verdicts are also appropriately cautious; no pair is overstated as a duplicate except the one strongly supported ZeRO-3/Bert pair." + "feedback": "The summary is well grounded in the packet and stays conservative on the soft-edge calls. The only clearly supported duplicate family is the DeepSpeed ZeRO-3 index-out-of-bounds loading bug cluster (43596/43638/41762), and the other pairings are appropriately rejected as distinct issues." }, "evaluator_used": true, "retried": false, @@ -560,10 +561,9 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:36331", - "issue:38175", "issue:39186", "issue:39290", "issue:39692", @@ -577,15 +577,14 @@ "issue:43404", "issue:43531", "issue:43541", + "issue:43572", "issue:43596", + "issue:43643", "issue:43644", "issue:43645", - "issue:43742", "issue:43828", "issue:43866", - "issue:43872", "issue:43873", - "issue:43881", "issue:43883", "issue:43950", "issue:43975", @@ -599,6 +598,7 @@ "issue:44512", "issue:44530", "issue:44534", + "issue:44560", "issue:44589", "issue:44661", "issue:44863", @@ -609,23 +609,25 @@ "issue:45070", "issue:45161", "issue:45399", - "issue:45464" + "issue:45464", + "issue:45507" ], "soft_pairs": [ + "issue:43572|issue:45070", "issue:44387|issue:45005", "issue:43541|issue:45161", "issue:43828|issue:45464", "issue:44291|issue:45070", - "issue:43531|issue:45070", "issue:43366|issue:45464", + "issue:43531|issue:45070", "issue:44661|issue:45464", "issue:43541|issue:44589", - "issue:39290|issue:43742", + "issue:39290|issue:43643", "issue:43054|issue:43994", "issue:41093|issue:43596", "issue:42915|issue:43541", - "issue:39186|issue:44918", "issue:40990|issue:44960", + "issue:39186|issue:44918", "issue:43404|issue:45005", "issue:43645|issue:45399", "issue:44530|issue:45005", @@ -633,251 +635,257 @@ "issue:43873|issue:45005", "issue:42915|issue:45399", "issue:44360|issue:44512", + "issue:44918|issue:45507", "issue:43975|issue:44292", - "issue:44492|issue:44512", "issue:43950|issue:44534", + "issue:44492|issue:44512", "issue:43644|issue:43950", + "issue:44292|issue:45507", "issue:43828|issue:44928", "issue:39692|issue:43054", - "issue:41553|issue:43883", "issue:43866|issue:44403", - "issue:36331|issue:41762", - "issue:41553|issue:43881", - "issue:43872|issue:44589", - "issue:38175|issue:43054" + "issue:41553|issue:43883", + "issue:44560|issue:45507", + "issue:36331|issue:41762" ], - "prepared_review_unit_hash": "942bfe441e092cde0e0968c00a23394ad91680cc67c8e33ebd5e75e345b58734", + "prepared_review_unit_hash": "ea18976fa602f7739bdcee183c749da69a271efded6bc9d799f1a02715b00976", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37301, - "estimated_input_tokens": 9326, - "estimated_eval_tokens": 18908 + "soft_pair_count": 34, + "serialized_chars": 37619, + "estimated_input_tokens": 9405, + "estimated_eval_tokens": 19066 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated. The only clear duplicate set is the Transformers v5 non-persistent-buffer regression: issues 43644, 43950, and 44534 describe the same corruption/junk-fill problem from different angles. No PRs were provided.", - "confidence": 0.87, - "canonical_issue_reason": "Issue 43950 is the clearest and most specific statement of the shared bug: `from_pretrained()` corrupts non-persistent buffers in Transformers 5.x.", + "summary": "Two clear duplicates in this set: the non-persistent buffer regression around `from_pretrained()` / Transformers v5, and the broader duplicate wording of that same issue. The rest are same-surface-area only and not the same concrete bug.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43950 is the clearest representative of the non-persistent-buffer regression: it names the failing path (`from_pretrained`) and the specific corruption of `register_buffer(persistent=False)`.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43950 best captures the underlying regression in a concrete, actionable way and is the strongest dedup target among the provided issues.", + "best_issue_reason": "issue:43950 best captures the underlying bug with enough detail to anchor the cluster; issue:43644 is the same regression but phrased more generally.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Both mention config/field problems after v5, but one is `pad_token_idx` in StableLmConfig and the other is a pydantic `PretrainedConfig` field issue; different bugs." + }, { "left": "issue:44387", "right": "issue:45005", "accept": false, - "reason": "Different problems: int4 quantization OOM vs tied-weights behavior in translation models." + "reason": "Both touch v5/weights/quantization themes, but one is an int4 OOM/memory regression and the other is tied weights in translation models." }, { "left": "issue:43541", "right": "issue:45161", "accept": false, - "reason": "Both involve MoE models, but one is a torch-dynamo grouped_mm tracing error and the other is a tensor-parallel support issue." + "reason": "Both involve MoE/inference, but the concrete failures differ: grouped_mm tracing error vs tensor-parallel support not working for GPT-OSS." }, { "left": "issue:43828", "right": "issue:45464", "accept": false, - "reason": "Autocast dtype mismatch and streaming chat/completions failure are distinct failure modes." + "reason": "Different models and failure modes: dtype mismatch under autocast vs chat/completions streaming failure." }, { "left": "issue:44291", "right": "issue:45070", "accept": false, - "reason": "Unexpected `_is_hf_initialized` TypeError is unrelated to the pydantic `PretrainedConfig` field regression." + "reason": "`init_empty_weights` unexpected argument error is unrelated to the pydantic `PretrainedConfig` field regression." }, { - "left": "issue:43531", - "right": "issue:45070", + "left": "issue:43366", + "right": "issue:45464", "accept": false, - "reason": "Qwen3-MoE sliding-window issue is not the same as the pydantic model field problem." + "reason": "GPT-OSS GGUF support request is unrelated to Qwen3.5 streaming inference failures." }, { - "left": "issue:43366", - "right": "issue:45464", + "left": "issue:43531", + "right": "issue:45070", "accept": false, - "reason": "GGUF architecture support request and Qwen3.5 streaming inference failure are different issues." + "reason": "Sliding-window behavior in Qwen3-MoE is a different model-specific bug from the config-field/pydantic issue." }, { "left": "issue:44661", "right": "issue:45464", "accept": false, - "reason": "Tokenizer mapping failure is unrelated to the chat/completions streaming bug." + "reason": "Tokenizer mapping failure when adding a model is unrelated to Qwen3.5 streaming inference." }, { "left": "issue:43541", "right": "issue:44589", "accept": false, - "reason": "Different code paths: grouped_mm tracing error vs missing Float8 storage type." + "reason": "Grouped_mm tracing failure and missing Float8 storage type are different bugs." }, { "left": "issue:39290", - "right": "issue:43742", + "right": "issue:43643", "accept": false, - "reason": "Gemma3 sliding-window attribute error and MobileLLM key error are unrelated loading failures." + "reason": "Both are config-field issues, but one is Gemma3/vLLM missing `sliding_window_pattern` and the other is `trust_remote_code` returning incomplete config objects." }, { "left": "issue:43054", "right": "issue:43994", "accept": false, - "reason": "Both mention SigLIP2, but one is an embedding-quality complaint and the other is a nonsensical-output/pipeline issue; not enough evidence they are the same bug." + "reason": "Both are SigLIP2-related, but one reports poor text embeddings and the other nonsensical AutoModel/pipeline results; not the same concrete defect." }, { "left": "issue:41093", "right": "issue:43596", "accept": false, - "reason": "Both are index/shape errors, but they involve different shapes and model/loading contexts." + "reason": "Both are index errors around loading, but the shapes and code paths differ (`mask` mismatch vs zero-sized index during ZeRO-3/BertModel load)." }, { "left": "issue:42915", "right": "issue:43541", "accept": false, - "reason": "Qwen3Moe FP8 failure and Mixtral grouped_mm tracing error are different problems." + "reason": "Different issues: FineGrainedFP8Config failure for Qwen3-MoE vs float32 grouped_mm tracing failure for Mixtral." }, { - "left": "issue:39186", - "right": "issue:44918", + "left": "issue:40990", + "right": "issue:44960", "accept": false, - "reason": "FSDP weight-dimension error is unrelated to TRL embedding unpacking failure." + "reason": "Unrelated model/perplexity report versus a GLM5 issue with no concrete overlap." }, { - "left": "issue:40990", - "right": "issue:44960", + "left": "issue:39186", + "right": "issue:44918", "accept": false, - "reason": "High perplexity on gpt-oss and a GLM5 issue are not the same bug." + "reason": "FSDP 2-D weight error and TRL SFT unpacking failure are distinct training/runtime bugs." }, { "left": "issue:43404", "right": "issue:45005", "accept": false, - "reason": "Both mention tied weights, but one is a specific Mistral3 model bug and the other is a broader v5 translation-model regression." + "reason": "Both mention tied weights, but one is a specific Mistral3 lm_head tie bug and the other is a broader translation-model v5 issue." }, { "left": "issue:43645", "right": "issue:45399", "accept": false, - "reason": "Custom-model Jupyter notebook breakage is unrelated to flash-attn2 fallback gating." + "reason": "Custom model initialization in Jupyter is unrelated to the flash-attn2 fallback/check ordering problem." }, { "left": "issue:44530", "right": "issue:45005", "accept": false, - "reason": "PagedAttentionCache crash and tied-weights regression are different subsystems and symptoms." + "reason": "PagedAttentionCache/group-type crash on Qwen3.5 is unrelated to tied-weight translation-model regressions." }, { "left": "issue:44863", "right": "issue:45005", "accept": false, - "reason": "NemotronH checkpoint loading failure is unrelated to tied-weights handling." + "reason": "Checkpoint-loading failure for NemotronH is not the same bug as tied weights on translation models." }, { "left": "issue:43873", "right": "issue:45005", "accept": false, - "reason": "Quantization/offloading behavior does not match the tied-weights regression." + "reason": "Quantization/offloading behavior and tied-weight handling are different issues." }, { "left": "issue:42915", "right": "issue:45399", "accept": false, - "reason": "Qwen3Moe FP8 support and flash-attn2 fallback checks are unrelated." + "reason": "Qwen3-MoE FP8 loading and flash-attn2 fallback gating are unrelated." }, { "left": "issue:44360", "right": "issue:44512", "accept": false, - "reason": "DSA indexer ReLU discussion is unrelated to documentation command removal." + "reason": "A DSA indexer ReLU discussion is not the same as docs still mentioning the removed run command." }, { - "left": "issue:43975", - "right": "issue:44292", + "left": "issue:44918", + "right": "issue:45507", "accept": false, - "reason": "Detokenization regression and NVFP4 loading failure are different issues." + "reason": "TRL input-embedding unpacking and GraniteMoEHybrid invalid method calls are different model/integration bugs." }, { - "left": "issue:44492", - "right": "issue:44512", + "left": "issue:43975", + "right": "issue:44292", "accept": false, - "reason": "A typo in cache-strategy docs is not the same as outdated run-command docs." + "reason": "Detokenization bug in deepseek-coder v5 is unrelated to Qwen-3-8B-NVFP4 runtime errors." }, { "left": "issue:43950", "right": "issue:44534", "accept": true, - "reason": "Same underlying regression: Transformers v5 corrupts/fills non-persistent buffers with junk." + "reason": "Same regression: Transformers v5 corrupts non-persistent buffers; one title names the concrete `from_pretrained()` path, the other states the same junk-buffer symptom." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A typo in cache strategies is unrelated to docs referencing the removed run command." }, { "left": "issue:43644", "right": "issue:43950", "accept": true, - "reason": "Both report the same non-persistent-buffer corruption in Transformers 5.x, with 43950 adding the `from_pretrained()` detail." + "reason": "Same underlying bug: v5 incorrectly fills/corrupts non-persistent buffers. `43950` is the clearer, more specific report." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime failure and GraniteMoEHybrid invalid method calls are unrelated." }, { "left": "issue:43828", "right": "issue:44928", "accept": false, - "reason": "Different numerical issues: autocast dtype mismatch vs NaN gradient explosion from 3D position_ids/SDPA fallback." + "reason": "Different problems in different models: autocast dtype mismatch vs Qwen3.5 RLHF gradient explosion/SDPA fallback." }, { "left": "issue:39692", "right": "issue:43054", "accept": false, - "reason": "SigLIP2 docs-example mismatches/quantization errors are not the same as the embedding-quality regression." - }, - { - "left": "issue:41553", - "right": "issue:43883", - "accept": false, - "reason": "Bad Voxtral tokenizer error messaging is unrelated to Molmo tied-weights-key errors." + "reason": "SigLIP2 docs example errors (model/processor mismatch and quantization failure) are not the same as a text-embedding quality regression." }, { "left": "issue:43866", "right": "issue:44403", "accept": false, - "reason": "Checkpoint corruption and generic loading noise are not the same bug." - }, - { - "left": "issue:36331", - "right": "issue:41762", - "accept": false, - "reason": "CustomTrainer keyword-argument mismatch is unrelated to DeepSpeed ZeRO-3 loading index errors." + "reason": "Corrupted Ovis2 checkpoint and generic loading noise are unrelated." }, { "left": "issue:41553", - "right": "issue:43881", + "right": "issue:43883", "accept": false, - "reason": "Voxtral AutoTokenizer messaging issue is unrelated to glm-4v-9b loading failure." + "reason": "Bad AutoTokenizer error message for Voxtral and missing `all_tied_weights_keys` on MolmoForCausalLM are different issues." }, { - "left": "issue:43872", - "right": "issue:44589", + "left": "issue:44560", + "right": "issue:45507", "accept": false, - "reason": "bitsandbytes `_is_hf_initialized` incompatibility is unrelated to Float8 storage lookup failure." + "reason": "Qwen3-vl video StopIteration and GraniteMoEHybrid invalid method calls are unrelated." }, { - "left": "issue:38175", - "right": "issue:43054", + "left": "issue:36331", + "right": "issue:41762", "accept": false, - "reason": "Both mention SigLIP2, but zero probabilities and weaker embeddings are not clearly the same defect without stronger evidence." + "reason": "CustomTrainer `compute_loss` kwarg mismatch and Gemma3 ZeRO-3 loading index error are distinct bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The only accepted duplicate cluster, 43644/43950/44534, is well supported by the issue titles and descriptions, and the other pair verdicts do not overclaim similarity. No PR pair claims were made." + "feedback": "The summary is grounded in the packet, and the two accepted soft-edge pairs are conservative and well-supported by the issue titles: 43950/44534 and 43644/43950 clearly describe the same non-persistent-buffer regression. The other verdicts stay cautious and do not overclaim duplicates." }, "evaluator_used": true, "retried": false, @@ -896,9 +904,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:36331", + "issue:38175", "issue:41093", "issue:41553", "issue:41762", @@ -931,12 +940,10 @@ "issue:44387", "issue:44403", "issue:44451", - "issue:44479", "issue:44488", "issue:44509", - "issue:44560", + "issue:44589", "issue:44661", - "issue:44805", "issue:44863", "issue:44960", "issue:44991", @@ -945,289 +952,279 @@ "issue:45305", "issue:45313", "issue:45356", - "issue:45399" + "issue:45399", + "issue:45520" ], "soft_pairs": [ + "issue:43872|issue:44589", + "issue:41553|issue:43881", + "issue:38175|issue:43054", "issue:42947|issue:44387", - "issue:43881|issue:43883", "issue:43054|issue:43493", + "issue:43881|issue:43883", "issue:44403|issue:44451", "issue:42947|issue:43856", - "issue:44403|issue:44488", + "issue:43975|issue:45520", "issue:43643|issue:45020", + "issue:44403|issue:44488", "issue:43782|issue:44960", "issue:42491|issue:44368", "issue:41762|issue:42915", "issue:43975|issue:44451", "issue:45020|issue:45356", - "issue:43824|issue:43872", "issue:44661|issue:45020", - "issue:43975|issue:44488", + "issue:43824|issue:43872", "issue:43054|issue:43278", "issue:44991|issue:45399", "issue:43792|issue:44220", - "issue:45081|issue:45399", "issue:43827|issue:44509", + "issue:45081|issue:45399", "issue:43404|issue:43950", "issue:42915|issue:45020", - "issue:41553|issue:44991", "issue:43257|issue:43866", "issue:43742|issue:45399", "issue:43742|issue:45020", - "issue:43872|issue:44291", + "issue:41553|issue:44991", "issue:43940|issue:45313", + "issue:43872|issue:44291", + "issue:43854|issue:44863", "issue:42915|issue:43872", "issue:41553|issue:44488", - "issue:43854|issue:44863", "issue:43257|issue:45305", - "issue:36331|issue:41093", - "issue:41093|issue:44805", - "issue:44479|issue:44560" + "issue:36331|issue:41093" ], - "prepared_review_unit_hash": "a8affc8763a22ac561dca7fe62715b54e67ed14b3ef985f26db2524da6d2ccd7", + "prepared_review_unit_hash": "19323840d6f22a77861d4ba4e10d79dcbed2906cad12b132531e2c79ec9c9edb", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 35, - "serialized_chars": 38367, - "estimated_input_tokens": 9592, - "estimated_eval_tokens": 19440 + "soft_pair_count": 36, + "serialized_chars": 38577, + "estimated_input_tokens": 9645, + "estimated_eval_tokens": 19546 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: many unrelated Transformers 5.x loading/regression reports, with a few true near-duplicates around the same concrete bug. The clearest matches are SigLIP2 implementation mismatch, docs still referencing removed pipelines, the `_is_hf_initialized` loading bug, and the Qwen video regression.", - "confidence": 0.61, - "canonical_issue_reason": "issue:45020 is the best umbrella anchor for the cluster because it is the broadest open report about recent-version regressions breaking remote_code models, which matches the dominant theme of the set.", + "summary": "Mostly heterogeneous issues with a few true duplicates. Clear duplicate pairs are the two `_is_hf_initialized` loading errors and the two v5 docs pipeline-removal issues; one Whisper audio-loading pair also looks like the same underlying bug. Most other soft pairs are only loosely related by model family or subsystem and should stay separate.", + "confidence": 0.81, + "canonical_issue_reason": "No single issue cleanly represents the whole set because it spans unrelated bugs. If an umbrella must be chosen, issue 45020 is the broadest current regression around remote_code-based model loading, but it does not cover most of the other items.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45020 is the most representative issue for this cluster overall: broad scope, active, and closest to the common loading/regression pattern seen across the items.", + "best_issue_reason": "Issue 45020 is the best global issue candidate only as a broad remote_code-loading regression anchor; otherwise the cluster is too mixed for a true canonical issue.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:42947", - "right": "issue:44387", + "left": "issue:43872", + "right": "issue:44589", "accept": false, - "reason": "Different bugs: LoRA/gradient-checkpointing effectiveness vs int4 memory/OOM behavior." + "reason": "Both are TypeErrors during loading, but they involve different failure points: `_is_hf_initialized` vs missing `Float8_e4m3fnStorage`." }, { - "left": "issue:43881", - "right": "issue:43883", + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different bugs: a misleading AutoTokenizer error for Voxtral vs a concrete load failure for GLM-4v-9b." + }, + { + "left": "issue:38175", + "right": "issue:43054", "accept": false, - "reason": "Different model-specific loading failures with unrelated root causes." + "reason": "Same model family, but one is zero probabilities and the other is worse text embeddings; not the same concrete bug." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Different training problems: LoRA gradient checkpointing vs int4 quantization memory growth." }, { "left": "issue:43054", "right": "issue:43493", - "accept": true, - "reason": "Both point to the same SigLIP2 implementation discrepancy vs the expected behavior/reference implementation." + "accept": false, + "reason": "Both concern SigLIP2, but one is a quality regression and the other is an HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Unrelated load failures in different models and code paths." }, { "left": "issue:44403", "right": "issue:44451", "accept": false, - "reason": "One is load-time noise, the other is a hard model-load failure for a specific checkpoint." + "reason": "Docs/noise complaint vs a model-loading failure for ScandiBERT." }, { "left": "issue:42947", "right": "issue:43856", "accept": false, - "reason": "Both mention training efficiency/memory, but the concrete problems and code paths differ." + "reason": "Both involve training efficiency, but they are different issues: LoRA checkpointing vs Qwen3 MoE memory usage." }, { - "left": "issue:44403", - "right": "issue:44488", + "left": "issue:43975", + "right": "issue:45520", "accept": false, - "reason": "Different reports: generic loading noise versus a specific model failing to load." + "reason": "Tokenizer detokenization regression is unrelated to the Python 3.13 `flash_attn` import error." }, { "left": "issue:43643", "right": "issue:45020", "accept": false, - "reason": "Related theme, but not the same concrete bug; one is missing fields in AutoConfig, the other is a broad remote_code regression report." + "reason": "Related to remote_code, but one is a specific missing-fields bug in AutoConfig and the other is a broader regression claim." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Different symptoms and targets; one is docs/noise, the other is a model load failure." }, { "left": "issue:43782", "right": "issue:44960", "accept": false, - "reason": "Different models and failure modes; no clear shared code-path bug." + "reason": "Different models and failure modes." }, { "left": "issue:42491", "right": "issue:44368", "accept": false, - "reason": "LoRA compatibility regression is not the same as a tied-embeddings warning/config issue." + "reason": "Different problems: hf4.x/hf5.x LoRA compatibility vs a tie_word_embeddings warning." }, { "left": "issue:41762", "right": "issue:42915", "accept": false, - "reason": "Different DeepSpeed/ZeRO-3 failures affecting different models and code paths." + "reason": "Different model families and different DeepSpeed/FP8 failure modes." }, { "left": "issue:43975", "right": "issue:44451", "accept": false, - "reason": "Different model/tokenizer problems with no concrete overlap." + "reason": "Tokenizer detokenization bug vs unrelated model loading failure." }, { "left": "issue:45020", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer-related regressions, but one is broad remote_code breakage and the other is a Kimi-K2.5 codec/mistral-regex regression." + "reason": "Remote_code regression vs Kimi tokenizer codec/fix_mistral_regex regression; not the same bug." }, { - "left": "issue:43824", - "right": "issue:43872", + "left": "issue:44661", + "right": "issue:45020", "accept": false, - "reason": "Import/export issue for a model class versus a bitsandbytes constructor incompatibility." + "reason": "One is a new-model-like script edge case, the other is a broad remote_code loading regression." }, { - "left": "issue:44661", - "right": "issue:45020", - "accept": false, - "reason": "Tokenizer-mapping edge case is a different bug from the broader remote_code loading regression." - }, - { - "left": "issue:43975", - "right": "issue:44488", + "left": "issue:43824", + "right": "issue:43872", "accept": false, - "reason": "Different checkpoints and symptoms; not the same underlying defect." + "reason": "ImportError for a missing class vs a TypeError in bitsandbytes initialization." }, { "left": "issue:43054", "right": "issue:43278", "accept": false, - "reason": "Both touch embedding-related behavior, but one is SigLIP2 quality/discrepancy and the other is a dtype change during eval." + "reason": "Both mention embeddings, but one is SigLIP2 quality and the other is a dtype change during evaluation." }, { "left": "issue:44991", "right": "issue:45399", "accept": false, - "reason": "Tokenizer load failure and flash-attn fallback logic are unrelated." + "reason": "Tokenizer loading regression vs flash-attn fallback logic." }, { "left": "issue:43792", "right": "issue:44220", - "accept": false, - "reason": "Whisper runtime failure and `_torch_extract_fbank_features()` may be related, but the available evidence is too weak to call them the same bug." - }, - { - "left": "issue:45081", - "right": "issue:45399", - "accept": false, - "reason": "Mistral regex patch crash versus flash-attn fallback gating are unrelated." + "accept": true, + "reason": "Likely the same Whisper audio feature-extraction bug; the generic 'can't run' report matches the `_torch_extract_fbank_features()` failure." }, { "left": "issue:43827", "right": "issue:44509", "accept": true, - "reason": "Same docs cleanup: both say pipeline task docs still reference removals from v5." + "reason": "Same docs gap: both say removed v5 pipeline tasks are still referenced." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer regex crash vs flash-attn fallback being blocked are different code paths." }, { "left": "issue:43404", "right": "issue:43950", "accept": false, - "reason": "Different state-management bugs: untied lm_head weights versus non-persistent buffer corruption." + "reason": "Tied lm_head weights vs non-persistent buffer corruption are unrelated issues." }, { "left": "issue:42915", "right": "issue:45020", "accept": false, - "reason": "Specific Qwen3 MoE/DeepSpeed weight-conversion failure is not the same as broad remote_code breakage." - }, - { - "left": "issue:41553", - "right": "issue:44991", - "accept": false, - "reason": "Bad error messaging for one tokenizer/model versus a tokenizer loading failure for another model." + "reason": "Different Qwen3/DeepSpeed/FP8 issue vs broad remote_code loading regression." }, { "left": "issue:43257", "right": "issue:43866", "accept": false, - "reason": "Different Qwen3 MoE loading/training issues with no shared concrete fix path." + "reason": "Different Qwen3 MoE problems: weight conversion/loading vs a corrupted Ovis2 checkpoint." }, { "left": "issue:43742", "right": "issue:45399", "accept": false, - "reason": "MobileLLM loading KeyError and flash-attn fallback gating are unrelated." + "reason": "MobileLLM key error is unrelated to flash-attn fallback blocking." }, { "left": "issue:43742", "right": "issue:45020", "accept": false, - "reason": "A specific model KeyError is not the same as the broad remote_code regression." + "reason": "A specific model-key error, not the broader remote_code regression." }, { - "left": "issue:43872", - "right": "issue:44291", - "accept": true, - "reason": "Same underlying `_is_hf_initialized` incompatibility during model loading/init_empty_weights." + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer-related failures in different models and code paths." }, { "left": "issue:43940", "right": "issue:45313", "accept": false, - "reason": "Both are ZeRO-3 load failures, but they involve different models and likely different missing-weight code paths." + "reason": "Both are ZeRO-3 weight-loading failures, but for different Qwen model variants with different missing-weight patterns." }, { - "left": "issue:42915", - "right": "issue:43872", - "accept": false, - "reason": "DeepSpeed/Qwen3 MoE weight loading and bitsandbytes `_is_hf_initialized` incompatibility are different bugs." - }, - { - "left": "issue:41553", - "right": "issue:44488", - "accept": false, - "reason": "Different model/tokenizer loading problems with no concrete overlap." + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying bug: loading with `init_empty_weights`/bitsandbytes trips on unexpected `_is_hf_initialized` handling." }, { "left": "issue:43854", "right": "issue:44863", "accept": false, - "reason": "Separate model-specific loading issues for different architectures/checkpoints." - }, - { - "left": "issue:43257", - "right": "issue:45305", - "accept": false, - "reason": "Different DeepSpeed-related reports: MoE weight conversion versus gradient averaging with loss kwargs." + "reason": "Different models and failures: GLM-4.7-Flash test loading vs NemotronH checkpoint loading." }, { "left": "issue:36331", "right": "issue:41093", "accept": false, - "reason": "Different error classes and code paths: trainer signature mismatch versus tensor-mask shape mismatch." - }, - { - "left": "issue:41093", - "right": "issue:44805", - "accept": false, - "reason": "Very similar error text, but the available information is insufficient to confirm the same underlying bug." - }, - { - "left": "issue:44479", - "right": "issue:44560", - "accept": true, - "reason": "Same Qwen VL video-input regression family in v5.3.0, with one issue being a specific manifestation of the broader breakage." + "reason": "Unrelated errors: custom trainer loss signature vs a mask shape mismatch." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative. The canonical anchor at issue:45020 is reasonable as a broad remote_code regression umbrella, and the accepted soft pairs are supported by the issue titles/bodies (SigLIP2 discrepancy, removed pipeline docs, _is_hf_initialized loading bug, and the Qwen video regression family). The rejected pairs are phrased cautiously and do not overclaim shared root causes." + "feedback": "The clustering is grounded in the packet and the verdicts stay conservative. The accepted pairs are well-supported by the titles and issue descriptions: the two `_is_hf_initialized`/`init_empty_weights` TypeErrors, the two v5 docs pipeline-removal issues, and the Whisper audio-feature extraction reports. The remaining soft pairs are appropriately rejected as only loosely related by subsystem or model family." }, "evaluator_used": true, "retried": false, @@ -1246,13 +1243,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:24643", "issue:30064", "issue:36010", "issue:38175", "issue:39692", + "issue:41093", "issue:42491", "issue:42617", "issue:42915", @@ -1269,14 +1267,13 @@ "issue:43720", "issue:43824", "issue:43828", - "issue:43856", "issue:43872", - "issue:43873", "issue:43950", "issue:43957", "issue:43994", "issue:44075", "issue:44451", + "issue:44479", "issue:44488", "issue:44530", "issue:44534", @@ -1298,244 +1295,251 @@ "issue:45357" ], "soft_pairs": [ + "issue:41093|issue:44805", + "issue:44479|issue:44560", "issue:44530|issue:44863", "issue:43872|issue:45084", "issue:43366|issue:43828", "issue:24643|issue:30064", "issue:43366|issue:43531", "issue:43550|issue:44910", - "issue:44560|issue:44918", - "issue:43366|issue:44589", "issue:43645|issue:43646", "issue:43950|issue:45325", + "issue:43366|issue:44589", + "issue:44560|issue:44918", "issue:43824|issue:43957", "issue:42915|issue:43957", "issue:42617|issue:43366", "issue:38175|issue:43994", - "issue:44560|issue:44805", "issue:43646|issue:43950", + "issue:44560|issue:44805", "issue:36010|issue:42915", "issue:43638|issue:44661", - "issue:44877|issue:45030", "issue:43596|issue:44661", + "issue:44877|issue:45030", "issue:43317|issue:44756", "issue:43644|issue:44534", "issue:38175|issue:39692", "issue:45313|issue:45357", - "issue:44960|issue:44977", - "issue:44451|issue:44488", - "issue:43708|issue:44075", "issue:42491|issue:43720", - "issue:43856|issue:43873" + "issue:44451|issue:44488", + "issue:44960|issue:44977", + "issue:43708|issue:44075" ], - "prepared_review_unit_hash": "4ea509db17b7ac46dd2dd8575bb8795e72bbee9f0883fed612b62d57c03c7611", + "prepared_review_unit_hash": "54d380f1f9e6dded91bfc4ac0da3a645400ef84064f76bb1688f82de61c6b480", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36234, - "estimated_input_tokens": 9059, - "estimated_eval_tokens": 18374 + "soft_pair_count": 30, + "serialized_chars": 36649, + "estimated_input_tokens": 9163, + "estimated_eval_tokens": 18582 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: many pairs are clearly different bugs or model-specific regressions. Only a few soft edges look like true duplicates, notably the two Transformers 5 custom-model-init reports, the non-persistent-buffer regression reports, and likely the two SigLIP2 result-regression reports.", - "confidence": 0.91, - "canonical_issue_reason": "issue:43646 is the cleanest canonical among the true duplicate pockets: it states the core custom-model initialization regression broadly, while issue:43645 is a narrower notebook-specific manifestation.", + "summary": "This set is not one coherent cluster; it breaks into a few small duplicate pairs and many unrelated issues. The clearest duplicates are around the mask-shape regression, custom model initialization in Transformers 5.0.0, SigLIP2 output corruption, and non-persistent buffer corruption.", + "confidence": 0.82, + "canonical_issue_reason": "No single canonical issue fits the whole set because the items span several unrelated bugs; the best-defined duplicate subclusters are 41093/44805, 43645/43646, 38175/43994, and 43644/44534.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43646 is the best issue representative here because the cluster has no single shared bug theme; most items are unrelated, and 43646 is the broadest of the clear duplicate candidates.", + "best_issue_reason": "No global best issue among this heterogeneous set; 43644 is a strong representative for the non-persistent-buffer regression, but it does not cover the other duplicate subclusters.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same mask-shape mismatch regression; 44805 is the same failure pattern with the same tensor-size mismatch." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-related, but they describe different model families and different failure modes; not clearly the same concrete bug." + }, { "left": "issue:44530", "right": "issue:44863", "accept": false, - "reason": "Different models and failures: Qwen3.5 flash-attention/paged attention crash vs NemotronH checkpoint loading." + "reason": "Qwen3.5 cache/linear_attention crash vs NemotronH checkpoint loading are unrelated." }, { "left": "issue:43872", "right": "issue:45084", "accept": false, - "reason": "Unrelated code paths: bitsandbytes Int8Params construction vs template-node compilation error." + "reason": "bitsandbytes initialization incompatibility and template-node compile error are different code paths." }, { "left": "issue:43366", "right": "issue:43828", "accept": false, - "reason": "Different subsystems and symptoms: GGUF/gpt-oss support vs autocast dtype mismatch in Phi-tiny-MoE." + "reason": "GGUF gpt-oss support and Phi-tiny-MoE autocast dtype mismatch are unrelated." }, { "left": "issue:24643", "right": "issue:30064", "accept": false, - "reason": "Training with DeepSpeed vs image processor void-segmentation preprocessing are unrelated bugs." + "reason": "DeepSpeed weight-dimension error and void segmentation-map processing failure are unrelated." }, { "left": "issue:43366", "right": "issue:43531", "accept": false, - "reason": "Different Qwen/GGUF issues; no shared concrete code-path." + "reason": "gpt-oss GGUF support and Qwen3-MoE sliding_window bug are different issues." }, { "left": "issue:43550", "right": "issue:44910", "accept": false, - "reason": "Both touch attention kernels, but the models and failure modes differ (Bamba torch.compile/SDPA vs Qwen3.5 flash-attention position_ids)." + "reason": "Different models and different failure mechanisms: SDPA/torch.compile vs flash-attention position_ids handling." }, { - "left": "issue:44560", - "right": "issue:44918", + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Both report Transformers 5.0.0 breaking custom model initialization; the notebook-specific report is a variant of the same regression." + }, + { + "left": "issue:43950", + "right": "issue:45325", "accept": false, - "reason": "Different failures in different pipelines: Qwen3-vl video stopiteration vs Qwen3.5 embedding unpacking with TRL." + "reason": "Non-persistent buffer corruption and Qwen2.5-VL rope-index scaling are unrelated." }, { "left": "issue:43366", "right": "issue:44589", "accept": false, - "reason": "GGUF architecture support and missing Float8 storage are unrelated." + "reason": "GGUF architecture support and missing Float8 storage type are unrelated." }, { - "left": "issue:43645", - "right": "issue:43646", - "accept": true, - "reason": "Same underlying Transformers 5 custom-model initialization regression; 43645 is the notebook-specific variant of 43646." - }, - { - "left": "issue:43950", - "right": "issue:45325", + "left": "issue:44560", + "right": "issue:44918", "accept": false, - "reason": "Non-persistent buffer corruption vs Qwen2.5-VL rope-index scaling are different bugs." + "reason": "Video StopIteration in Qwen3-VL embedding and TRL SFT input-embedding unpacking are different bugs." }, { "left": "issue:43824", "right": "issue:43957", "accept": false, - "reason": "Different loading/import problems: Qwen2_5_VL import error vs meta-device loading regressions." + "reason": "Importing Qwen2_5_VL class and meta-device model loading failures are unrelated." }, { "left": "issue:42915", "right": "issue:43957", "accept": false, - "reason": "Qwen3Moe FP8 failure and meta-device loading breakage are separate issues." + "reason": "Qwen3Moe FP8 config failure and meta-device loading regressions are different problems." }, { "left": "issue:42617", "right": "issue:43366", "accept": false, - "reason": "3d_parallel.py runtime failure is unrelated to GGUF gpt-oss support." + "reason": "3d_parallel.py execution failure and GGUF gpt-oss support are unrelated." }, { "left": "issue:38175", "right": "issue:43994", "accept": true, - "reason": "Same SigLIP2 model/result regression: zero probabilities and nonsensical outputs on the same checkpoint look like the same underlying bug." + "reason": "Same SigLIP2 checkpoint/output regression; the second report rephrases the same bad model behavior in AutoModel/pipeline terms." }, { - "left": "issue:44560", - "right": "issue:44805", + "left": "issue:43646", + "right": "issue:43950", "accept": false, - "reason": "Qwen3-vl video StopIteration and a tensor mask shape IndexError are unrelated." + "reason": "Custom model initialization regression and non-persistent buffer corruption are both Transformers 5 issues but not the same bug." }, { - "left": "issue:43646", - "right": "issue:43950", + "left": "issue:44560", + "right": "issue:44805", "accept": false, - "reason": "Both concern loading/custom-model regressions, but the concrete bug paths differ: init handling vs non-persistent buffer corruption." + "reason": "Different failures: video StopIteration versus token/mask shape mismatch." }, { "left": "issue:36010", "right": "issue:42915", "accept": false, - "reason": "ImportError for GenerationMixin and Qwen3Moe FP8 config failure are unrelated." + "reason": "GenerationMixin import error and Qwen3Moe FP8 failure are unrelated." }, { "left": "issue:43638", "right": "issue:44661", "accept": false, - "reason": "Different bugs: DeepSpeed ZeRO-3 BertModel init vs add-new-model-like/tokenizer-mapping handling." + "reason": "DeepSpeed ZeRO-3 Bert index error and TOKENIZER_MAPPING_NAMES add-new-model-like failure are different code paths." }, { - "left": "issue:44877", - "right": "issue:45030", + "left": "issue:43596", + "right": "issue:44661", "accept": false, - "reason": "Both are config-loading issues, but the affected model families and validation failures differ." + "reason": "These are distinct issues: ZeRO-3/Bert initialization versus model registration in tokenizer mapping." }, { - "left": "issue:43596", - "right": "issue:44661", + "left": "issue:44877", + "right": "issue:45030", "accept": false, - "reason": "DeepSpeed ZeRO-3 empty-index error is unrelated to tokenizer-mapping model registration." + "reason": "Both involve config loading/validation, but they concern different model families and different validation failures." }, { "left": "issue:43317", "right": "issue:44756", "accept": false, - "reason": "Offload/dequantized-model loading and Strix Halo mmap/OOM are separate memory-management problems." + "reason": "Dequantized-model offload failure and mmap OOM workaround are not the same bug." }, { "left": "issue:43644", "right": "issue:44534", "accept": true, - "reason": "Same non-persistent-buffer regression; 44534 is essentially the later, shorter restatement of 43644." + "reason": "Both describe the same Transformers v5 regression where non-persistent buffers get filled with junk." }, { "left": "issue:38175", "right": "issue:39692", "accept": false, - "reason": "Both mention SigLIP2, but one is a result regression and the other is a documentation-example/processor-mismatch issue." + "reason": "Both mention SigLIP2, but one is output probabilities while the other is a docs-example/model-processor/quantization issue." }, { "left": "issue:45313", "right": "issue:45357", "accept": false, - "reason": "Different Qwen3.5 problems: ZeRO-3 weight loading vs save_pretrained visual-encoder key serialization." + "reason": "ZeRO-3 language_model loading failure and save_pretrained visual-encoder-key regression are different bugs." }, { - "left": "issue:44960", - "right": "issue:44977", + "left": "issue:42491", + "right": "issue:43720", "accept": false, - "reason": "GLM5 tracking issue vs Qwen3.5 flash-attention generation bug are not the same underlying change." + "reason": "Qwen3_MoE LoRA compatibility and BitNet packed-weight loading are unrelated." }, { "left": "issue:44451", "right": "issue:44488", "accept": false, - "reason": "Both are model-load failures, but the titles indicate different checkpoints and no shared concrete failure mode." - }, - { - "left": "issue:43708", - "right": "issue:44075", - "accept": false, - "reason": "Trainer checkpoint-resume max_steps bug and optimizer-argument handling are unrelated." + "reason": "Both are model-loading failures, but the titles point to different checkpoints and no shared concrete failure mode." }, { - "left": "issue:42491", - "right": "issue:43720", + "left": "issue:44960", + "right": "issue:44977", "accept": false, - "reason": "LoRA/hf4.x to hf5.x compatibility and BitNet packed-weight unloading are different code paths." + "reason": "GLM5 and Qwen3.5 flash-attention generation failures are unrelated." }, { - "left": "issue:43856", - "right": "issue:43873", + "left": "issue:43708", + "right": "issue:44075", "accept": false, - "reason": "Qwen3 MoE memory efficiency and quantization offload behavior are not the same bug." + "reason": "Trainer max_steps resume bug and SGD args not being used are different training/optimizer issues." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The chosen canonical/best issue rationale fits the clear duplicate pocket around Transformers 5 custom-model initialization, and the other accepted soft edges are plausible duplicates rather than overclaims. The rejected pairs are also conservatively explained from the titles alone. No mergeability issues apply here since these are issue pairs, not PR pairs." + "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted soft edges are supported by closely matching titles and issue descriptions, and the rejected pairs are appropriately cautious where the evidence only suggests broad thematic similarity rather than the same bug. The cluster-level conclusion that this is heterogeneous with a few small duplicate subclusters is also well supported." }, "evaluator_used": true, "retried": false, @@ -1554,7 +1558,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:36010", "issue:39404", @@ -1606,321 +1610,328 @@ "issue:45406" ], "soft_pairs": [ - "issue:41720|issue:42915", "issue:43854|issue:43931", - "issue:44488|issue:44991", - "issue:39692|issue:43994", - "issue:43856|issue:44387", + "issue:43856|issue:43873", + "issue:41720|issue:42915", "issue:45362|issue:45406", - "issue:44661|issue:44683", "issue:42491|issue:45276", - "issue:42915|issue:44756", + "issue:39692|issue:43994", + "issue:44488|issue:44991", "issue:43257|issue:43854", + "issue:44661|issue:44683", + "issue:43856|issue:44387", + "issue:42915|issue:44756", "issue:43644|issue:43645", "issue:42491|issue:43931", - "issue:45276|issue:45335", "issue:42491|issue:43856", + "issue:45276|issue:45335", "issue:43404|issue:43454", "issue:36010|issue:39404", "issue:44164|issue:44291", - "issue:43873|issue:44387", "issue:44402|issue:44488", "issue:41720|issue:44155", - "issue:44387|issue:44756", + "issue:43873|issue:44387", "issue:39692|issue:44960", + "issue:44387|issue:44756", "issue:43425|issue:43957", - "issue:42947|issue:43381", - "issue:43854|issue:45216", "issue:44292|issue:44912", + "issue:43854|issue:45216", + "issue:42947|issue:43381", "issue:44291|issue:44756", "issue:45276|issue:45292", "issue:44743|issue:44960", - "issue:43957|issue:44756", "issue:43381|issue:44928", + "issue:43957|issue:44756", "issue:44164|issue:44756", "issue:43646|issue:43708", - "issue:44479|issue:44877", "issue:45292|issue:45335", - "issue:43645|issue:44164", + "issue:44479|issue:44877", "issue:44402|issue:44451", - "issue:43425|issue:43828", + "issue:43645|issue:44164", "issue:43257|issue:44863", + "issue:43425|issue:43828", "issue:43828|issue:43957" ], - "prepared_review_unit_hash": "d54519f4c6b566d9d20203c15445a83566407387b3a5f1f1ca1337a3d781a1fd", + "prepared_review_unit_hash": "2eaaa3da7a1f901a99a53ab665d8bd4cb3e06591f523c030d412fd2e9f202b22", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 40, - "serialized_chars": 39511, - "estimated_input_tokens": 9878, - "estimated_eval_tokens": 20012 + "soft_pair_count": 41, + "serialized_chars": 39811, + "estimated_input_tokens": 9953, + "estimated_eval_tokens": 20162 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous. The only clear duplicate family is the `resize_token_embeddings` regressions affecting output/decoder embeddings across related model classes; the rest are distinct model-loading, quantization, checkpointing, or serving bugs.", - "confidence": 0.73, - "canonical_issue_reason": "issue:45292 is the broadest and most central of the resize-token-embedding reports: it names the shared symptom directly and is generic enough to subsume the more model-specific variants.", + "summary": "This cluster is mostly heterogeneous; only the resize_token_embeddings subgroup looks like true duplicates. Most other pairs are topical similarities rather than the same underlying bug.", + "confidence": 0.71, + "canonical_issue_reason": "No single canonical issue fits the whole cluster. For the only clear duplicate subgroup, issue 45292 is the best canonical candidate because it is the most generic description of the resize_token_embeddings regression and aligns with 45276/45335.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45292 is the best representative issue because it is the most general statement of the shared embedding-resize bug, while the others in that mini-family are narrower model-specific manifestations.", + "best_issue_reason": "Issue 45292 is the broadest, most reusable representative of the strongest duplicate set and is the best anchor for triage cleanup.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different model-loading failures for different models; same broad area, not the same bug." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both concern quantization/memory behavior, but the reported failures and likely code paths differ." + }, { "left": "issue:41720", "right": "issue:42915", "accept": false, - "reason": "Different bugs: one is auto device-mapping CUDA assert on Qwen3, the other is a Qwen3-MoE HF5.x compatibility/loadability regression." + "reason": "Qwen3 device-mapping crash vs Qwen3 MoE LoRA compatibility issue; not the same concrete problem." }, { - "left": "issue:43854", - "right": "issue:43931", + "left": "issue:45362", + "right": "issue:45406", "accept": false, - "reason": "Different models and failure modes: GLM-4.7-Flash unit-test loading vs Qwen3-VL weight-shape mismatch." + "reason": "Different model/processor bugs: chat crash vs processor missing _tokenizer in serve." }, { - "left": "issue:44488", - "right": "issue:44991", + "left": "issue:42491", + "right": "issue:45276", "accept": false, - "reason": "Both are tokenizer/model loading failures, but the affected models and likely causes differ; too broad to treat as the same bug." + "reason": "Both involve Qwen3.5/Qwen3 MoE, but one is a hf4.x/hf5.x compatibility issue and the other is token embedding resizing." }, { "left": "issue:39692", "right": "issue:43994", "accept": false, - "reason": "Both involve SigLIP2, but one is a docs/example mismatch and quantization issue, the other is incorrect outputs from AutoModel/pipeline." + "reason": "Both mention SigLIP2, but one is a docs example error and the other is an inference/output regression." }, { - "left": "issue:43856", - "right": "issue:44387", + "left": "issue:44488", + "right": "issue:44991", "accept": false, - "reason": "Both mention memory/quantization, but one is Qwen3 MoE training memory use and the other is int4 quantization reserved-memory/OOM regression." + "reason": "Tokenizer-loading failures for different models; similar symptom, not clearly the same root cause." }, { - "left": "issue:45362", - "right": "issue:45406", + "left": "issue:43257", + "right": "issue:43854", "accept": false, - "reason": "Different code paths and components: Qwen3.5 chat crash vs Gemma4Processor missing `_tokenizer` in serve." + "reason": "Qwen3 MoE weight conversion vs GLM-4.7 loading; unrelated bugs." }, { "left": "issue:44661", "right": "issue:44683", "accept": false, - "reason": "Unrelated areas: tokenizer-mapping model registration vs compiled flex_attention on newer torch." + "reason": "Tokenizer-mapping failure vs torch flex_attention compilation failure; unrelated." }, { - "left": "issue:42491", - "right": "issue:45276", + "left": "issue:43856", + "right": "issue:44387", "accept": false, - "reason": "Different regressions: Qwen3-MoE LoRA compatibility on hf5.x vs Gemma4 token-embedding resize behavior." + "reason": "Both are memory/quantization complaints, but the specific failure modes are different." }, { "left": "issue:42915", "right": "issue:44756", "accept": false, - "reason": "Different issues: Qwen3-MoE FP8 failure vs mmap/OOM on Strix Halo." - }, - { - "left": "issue:43257", - "right": "issue:43854", - "accept": false, - "reason": "Different bugs: Qwen3 MoE weights not converted under accelerate+deepspeed vs GLM-4.7-Flash loading in tests." + "reason": "Qwen3 MoE HF-version incompatibility vs mmap/OOM on Strix Halo; unrelated." }, { "left": "issue:43644", "right": "issue:43645", "accept": false, - "reason": "Both are Transformers 5.0.0 regressions, but one is junk non-persistent buffers and the other is custom-model init in notebooks." + "reason": "Both are Transformers 5.0 regressions, but one is buffer initialization and the other is custom-model/Jupyter initialization." }, { "left": "issue:42491", "right": "issue:43931", "accept": false, - "reason": "Different model families and failures: Qwen3-MoE LoRA reuse vs Qwen3-VL weight-shape mismatch." - }, - { - "left": "issue:45276", - "right": "issue:45335", - "accept": true, - "reason": "Same underlying resize-token-embedding bug family: both report that `resize_token_embeddings` does not propagate to model-specific embedding/output layers." + "reason": "Different issues: MoE checkpoint compatibility vs Qwen3-VL shape mismatch." }, { "left": "issue:42491", "right": "issue:43856", "accept": false, - "reason": "Different issues: Qwen3-MoE LoRA compatibility vs Qwen3 MoE training memory usage." + "reason": "Different Qwen3 MoE problems; no shared concrete bug." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Near-identical resize_token_embeddings regression affecting model-specific tied embeddings; same underlying bug family." }, { "left": "issue:43404", "right": "issue:43454", "accept": false, - "reason": "Similar theme of untied lm_head weights, but different model implementations and no evidence of the same concrete code path." + "reason": "Both are missing lm_head tying, but they are different model implementations and not proven to share one fixable code path." }, { "left": "issue:36010", "right": "issue:39404", "accept": false, - "reason": "Unrelated: import-time GenerationMixin import error vs Whisper pipeline `return_language` regression." + "reason": "ImportError for GenerationMixin vs Whisper pipeline return_language regression; unrelated." }, { "left": "issue:44164", "right": "issue:44291", "accept": false, - "reason": "Different save/load failures: `extra_state` handling vs `init_empty_weights` unexpected `_is_hf_initialized` argument." - }, - { - "left": "issue:43873", - "right": "issue:44387", - "accept": false, - "reason": "Both are quantization/offloading-related, but one is offloading not working and the other is int4 reserved-memory growth causing OOM." + "reason": "save/from_pretrained extra_state handling vs init_empty_weights argument mismatch; separate serialization/loading bugs." }, { "left": "issue:44402", "right": "issue:44488", "accept": false, - "reason": "Both are loading-related, but one is tokenizer vocab-size mismatch and the other is a specific model load failure; not the same bug." + "reason": "Tokenizer vocab-size inconsistency vs a specific model-loading failure; not the same defect." }, { "left": "issue:41720", "right": "issue:44155", "accept": false, - "reason": "Different code paths: Qwen3 auto device mapping CUDA assert vs AudioFlamingo3 batched inference token/embedding leak." + "reason": "CUDA assert in Qwen3 device mapping vs AudioFlamingo3 batched inference leakage; unrelated." }, { - "left": "issue:44387", - "right": "issue:44756", + "left": "issue:43873", + "right": "issue:44387", "accept": false, - "reason": "Different regressions: quantization memory growth vs disabling mmap to avoid OOM." + "reason": "General quantization/offloading complaint vs int4 reserved-memory OOM regression; not the same code-path problem." }, { "left": "issue:39692", "right": "issue:44960", "accept": false, - "reason": "Unrelated: SigLIP2 docs/example/model issues vs a separate GLM5 issue." + "reason": "SigLIP2 example/model mismatch vs GLM5 issue; unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Int4 CUDA reserved-memory regression vs Strix Halo mmap OOM; different mechanisms." }, { "left": "issue:43425", "right": "issue:43957", "accept": false, - "reason": "Different problems: Torch 2.10 compatibility vs meta-device model loading breakage." + "reason": "Torch version incompatibility vs meta-device loading bug; not the same issue." }, { - "left": "issue:42947", - "right": "issue:43381", + "left": "issue:44292", + "right": "issue:44912", "accept": false, - "reason": "Different gradient-checkpointing bugs: ineffective LoRA checkpointing vs checkpointing unusable in eval mode." + "reason": "Generic model-loading/initialization bug vs MXFP4 quantization fallback; unrelated." }, { "left": "issue:43854", "right": "issue:45216", "accept": false, - "reason": "Different model-loading/save bugs: GLM-4.7-Flash test loading vs Qwen3.5 save_pretrained regression." + "reason": "Different model families and different failure modes: load correctness vs save_pretrained regression." }, { - "left": "issue:44292", - "right": "issue:44912", + "left": "issue:42947", + "right": "issue:43381", "accept": false, - "reason": "Both are quantization-related, but the models and failure modes differ: Qwen-3-8B-NVFP4 vs git-oss-20b MXFP4 fallback." + "reason": "Both mention gradient checkpointing, but one is ineffective with PEFT LoRA and the other is eval-mode incompatibility." }, { "left": "issue:44291", "right": "issue:44756", "accept": false, - "reason": "Different issues: `init_empty_weights` loading failure vs mmap/OOM workaround on Strix Halo." + "reason": "init_empty_weights TypeError vs mmap/OOM on Strix Halo; unrelated." }, { "left": "issue:45276", "right": "issue:45292", - "accept": true, - "reason": "Same resize-token-embedding regression family; `output_embeddings` not updating is the shared symptom across the Gemma4-specific and generic reports." + "accept": false, + "reason": "Same resize_token_embeddings theme, but the evidence only shows similarity; not enough to treat as the same concrete bug without more detail." }, { "left": "issue:44743", "right": "issue:44960", "accept": false, - "reason": "Different subjects: recurrent-state reset in modular_qwen3_5 vs a GLM5 issue." + "reason": "Qwen3.5 cache reset bug vs GLM5 model issue; unrelated." }, { - "left": "issue:43957", - "right": "issue:44756", + "left": "issue:43381", + "right": "issue:44928", "accept": false, - "reason": "Different model-loading/OOM issues with no shared code path." + "reason": "Gradient checkpointing eval-mode issue vs RLHF NaN from 3D position_ids/SDPA fallback; unrelated." }, { - "left": "issue:43381", - "right": "issue:44928", + "left": "issue:43957", + "right": "issue:44756", "accept": false, - "reason": "Unrelated: gradient checkpointing in eval mode vs RLHF NaN/SDPA fallback issue." + "reason": "Meta-device loading regression vs mmap/OOM workaround; different problems." }, { "left": "issue:44164", "right": "issue:44756", "accept": false, - "reason": "Different bugs: extra_state serialization vs disabling mmap for OOM." + "reason": "Serialization extra_state bug vs Strix Halo mmap OOM; unrelated." }, { "left": "issue:43646", "right": "issue:43708", "accept": false, - "reason": "Different Transformers 5.0 regressions: custom model init vs Trainer resume/max_steps calculation." - }, - { - "left": "issue:44479", - "right": "issue:44877", - "accept": false, - "reason": "Different regressions: video-input handling vs strict config loading for granite_speech." + "reason": "Custom model initialization regression vs Trainer resume_from_checkpoint max_steps calculation; unrelated." }, { "left": "issue:45292", "right": "issue:45335", "accept": true, - "reason": "Same underlying resize-token-embedding bug family: generic `output_embeddings` mismatch and t5gemma `decoder.embed_tokens` mismatch are manifestations of the same class of failure." + "reason": "Same resize_token_embeddings regression expressed generically and for a specific model class; clearly the same underlying bug family." }, { - "left": "issue:43645", - "right": "issue:44164", + "left": "issue:44479", + "right": "issue:44877", "accept": false, - "reason": "Different save/load/runtime-init regressions; no shared underlying bug." + "reason": "Video input regression for Qwen models vs strict config preventing granite_speech loading; unrelated." }, { "left": "issue:44402", "right": "issue:44451", "accept": false, - "reason": "Different loading problems: vocab-size mismatch vs inability to load a specific model." + "reason": "Tokenizer vocab-size mismatch vs a model-specific loading failure; not the same bug." }, { - "left": "issue:43425", - "right": "issue:43828", + "left": "issue:43645", + "right": "issue:44164", "accept": false, - "reason": "Unrelated: torch version compatibility vs autocast dtype mismatch on Phi-tiny-MoE." + "reason": "Notebook/custom-model initialization regression vs extra_state save/load failure; separate issues." }, { "left": "issue:43257", "right": "issue:44863", "accept": false, - "reason": "Different issues: Qwen3 MoE loading/conversion vs NemotronH checkpoint loading." + "reason": "Qwen3 MoE deepspeed loading bug vs NemotronH checkpoint loading failure; unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility vs autocast dtype mismatch; different code paths." }, { "left": "issue:43828", "right": "issue:43957", "accept": false, - "reason": "Different bugs: autocast dtype mismatch vs meta-device model loading breakage." + "reason": "Autocast dtype mismatch vs meta-device model loading breakage; unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is well-grounded in the packet and stays conservative about the soft edges. The only accepted pairs are the clearly related `resize_token_embeddings` regressions, which are supported by the issue titles and similarity signals. No overstatement stands out." + "feedback": "The summary is grounded in the packet: this set is clearly heterogeneous, and the resize_token_embeddings trio is the only convincingly related subgroup. The two accepted soft edges stay conservative enough given the titles and high similarity scores, and the chosen canonical anchor (45292) is a reasonable generic representative for that subgroup." }, "evaluator_used": true, "retried": false, @@ -1939,7 +1950,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:32090", "issue:34567", @@ -1993,55 +2004,55 @@ "soft_pairs": [ "issue:44589|issue:44756", "issue:43906|issue:44530", - "issue:43425|issue:43716", "issue:43931|issue:44863", "issue:39404|issue:44912", "issue:43611|issue:43645", + "issue:43425|issue:43716", "issue:44530|issue:45084", + "issue:43299|issue:43931", "issue:44265|issue:44898", "issue:43901|issue:44509", - "issue:43299|issue:43931", "issue:44479|issue:44805", - "issue:44589|issue:45084", "issue:43596|issue:44683", - "issue:43425|issue:44918", - "issue:43856|issue:44756", - "issue:40990|issue:42947", "issue:43994|issue:44079", + "issue:44589|issue:45084", + "issue:43856|issue:44756", "issue:44671|issue:44991", "issue:39404|issue:44671", + "issue:40990|issue:42947", + "issue:43425|issue:44918", "issue:36032|issue:43452", "issue:44991|issue:45081", - "issue:43582|issue:43632", "issue:32090|issue:39404", + "issue:43582|issue:43632", + "issue:43957|issue:44164", "issue:44488|issue:45081", + "issue:44164|issue:44479", + "issue:40990|issue:43421", "issue:43425|issue:45161", - "issue:43957|issue:44164", "issue:43883|issue:44488", - "issue:40990|issue:43421", - "issue:44164|issue:44479", - "issue:42947|issue:44756", - "issue:43716|issue:45084", - "issue:44488|issue:44987", "issue:43611|issue:43644", "issue:44206|issue:44220", - "issue:43883|issue:44991", - "issue:34567|issue:40444" + "issue:44488|issue:44987", + "issue:42947|issue:44756", + "issue:43716|issue:45084", + "issue:34567|issue:40444", + "issue:43883|issue:44991" ], - "prepared_review_unit_hash": "13e560d18f1033839e91cd445a4132ed107b3b711717a3dc82a9ace240f324ad", + "prepared_review_unit_hash": "0968bb0c42ea639d854aa4e76d33c73574c94fb39c09a6d13d629b8978feb498", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 38477, + "serialized_chars": 38479, "estimated_input_tokens": 9620, "estimated_eval_tokens": 19496 }, @@ -2050,228 +2061,216 @@ "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a noisy cluster of mostly unrelated Transformers issues: v5 loading regressions, tokenizer/model-specific bugs, docs problems, audio/video/model-output edge cases, and a few training/runtime failures. I did not find any pair that clearly shares the same underlying bug or change, so I rejected all soft edges.", - "confidence": 0.22, - "canonical_issue_reason": "Issue 43611 is the clearest representative of the dominant theme in this cluster: a concrete, broadly scoped Transformers v5 model-loading regression. Even so, the set is heterogeneous rather than a true duplicate cluster.", + "summary": "This cluster is heterogeneous: most items are unrelated regressions or model-specific bugs, with a few loose topical similarities around Transformers v5 loading/tokenizer regressions, but no clear duplicate core.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue fits the whole set; the items span multiple independent bugs (trainer, tokenizer, vision, audio, MoE, loading, docs).", "canonical_pr_reason": null, - "best_issue_reason": "43611 is the best global anchor because it is specific, well-defined, and broadly representative of the v5 loading regressions appearing across the set.", + "best_issue_reason": "Issue 43611 is the closest broad representative because it captures the largest shared theme here (Transformers v5 model-loading regressions), but it is still not a true duplicate umbrella for the rest of the cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44589", "right": "issue:44756", "accept": false, - "reason": "Different failures: float8 storage lookup vs Apple Silicon mmap/OOM behavior." + "reason": "Different failures: float8 storage lookup vs mmap/OOM on Strix Halo; no shared code-path bug." }, { "left": "issue:43906", "right": "issue:44530", "accept": false, - "reason": "An isolated reproduction of another issue is not the same bug as a Qwen3.5 cache/group-type crash." - }, - { - "left": "issue:43425", - "right": "issue:43716", - "accept": false, - "reason": "Torch version incompatibility is unrelated to a Mistral image-preprocessor dtype mismatch." + "reason": "Unrelated: isolated reproduction of a prior issue vs Qwen3.5 PagedAttentionCache linear_attention crash." }, { "left": "issue:43931", "right": "issue:44863", "accept": false, - "reason": "Different models and different load errors: Qwen3-VL weight-shape mismatch vs NemotronH checkpoint loading." + "reason": "Both are model-loading problems, but for different architectures and failure modes (shape mismatch vs checkpoint loading support)." }, { "left": "issue:39404", "right": "issue:44912", "accept": false, - "reason": "Whisper pipeline return_language regression is unrelated to MXFP4 quantization fallback on git-oss-20b." + "reason": "Whisper pipeline return_language regression is unrelated to GPT-OSS MXFP4 quantization fallback." }, { "left": "issue:43611", "right": "issue:43645", "accept": false, - "reason": "Both are v5 loading regressions, but one is base_model_prefix handling and the other is custom-model initialization in notebooks; not the same defect." + "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is custom-model initialization in notebooks; different code paths." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch version incompatibility is not the same bug as a Mistral image preprocessor/model dtype mismatch." }, { "left": "issue:44530", "right": "issue:45084", "accept": false, - "reason": "Qwen3.5 PagedAttentionCache crash and template compilation failure are different code paths." + "reason": "Different subsystems and errors: Qwen3.5 cache group-type crash vs template compilation failure." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both involve Qwen3-VL loading, but one is a v5.dev0 MoE loading break and the other is a checkpoint shape mismatch; not clearly the same bug." }, { "left": "issue:44265", "right": "issue:44898", "accept": false, - "reason": "torch.export/torch_compilable_check failure is unrelated to Perceiver interpolation at non-default resolution." + "reason": "torch.export/torch_compilable_check is unrelated to Perceiver positional encoding interpolation failure." }, { "left": "issue:43901", "right": "issue:44509", "accept": false, - "reason": "Both are docs issues, but they cover different deprecated/removed pipeline documentation." - }, - { - "left": "issue:43299", - "right": "issue:43931", - "accept": false, - "reason": "Both involve Qwen/VL loading, but they are different model families and different breakages." + "reason": "Both are docs issues, but they concern different pipeline/task documentation changes." }, { "left": "issue:44479", "right": "issue:44805", "accept": false, - "reason": "A video-input regression is not the same as a mask/index shape mismatch." + "reason": "Video-input regression for Qwen VL models is unrelated to a generic mask shape IndexError." }, { - "left": "issue:44589", - "right": "issue:45084", + "left": "issue:43596", + "right": "issue:44683", "accept": false, - "reason": "Float8 storage lookup and non-template-node compilation are unrelated errors." + "reason": "Deepspeed ZeRO-3/BertModel index error is unrelated to compiled flex_attention failing on newer torch." }, { - "left": "issue:43596", - "right": "issue:44683", + "left": "issue:43994", + "right": "issue:44079", "accept": false, - "reason": "DeepSpeed ZeRO-3/Bert init failure is unrelated to compiled flex_attention breaking on torch>=2.9." + "reason": "SigLIP2 bad outputs and ModelOutput key assignment bugs are different problems." }, { - "left": "issue:43425", - "right": "issue:44918", + "left": "issue:44589", + "right": "issue:45084", "accept": false, - "reason": "Torch 2.10 compatibility and TRL SFT embedding unpacking are different problems." + "reason": "Float8 storage lookup failure is unrelated to non-template-node compilation errors." }, { "left": "issue:43856", "right": "issue:44756", "accept": false, - "reason": "Qwen3 MoE training memory usage is unrelated to Strix Halo mmap/OOM behavior." + "reason": "MoE memory usage regression is unrelated to Strix Halo mmap/OOM behavior." }, { - "left": "issue:40990", - "right": "issue:42947", + "left": "issue:44671", + "right": "issue:44991", "accept": false, - "reason": "High perplexity on a benchmark and gradient checkpointing ineffectiveness are not the same underlying bug." + "reason": "CamemBERT prediction quality regression and tokenizer loading failure are distinct issues." }, { - "left": "issue:43994", - "right": "issue:44079", + "left": "issue:39404", + "right": "issue:44671", "accept": false, - "reason": "SigLIP2 bad outputs and ModelOutput key assignment are different issues." + "reason": "Whisper pipeline return_language regression is unrelated to CamemBERT masked-LM output issues." }, { - "left": "issue:44671", - "right": "issue:44991", + "left": "issue:40990", + "right": "issue:42947", "accept": false, - "reason": "CamemBERT prediction errors and tokenizer-loading regressions do not share the same concrete bug." + "reason": "High perplexity on GPT-OSS-20B and ineffective gradient checkpointing with LoRA are different training/eval bugs." }, { - "left": "issue:39404", - "right": "issue:44671", + "left": "issue:43425", + "right": "issue:44918", "accept": false, - "reason": "Whisper pipeline behavior and CamemBERT MLM predictions are unrelated." + "reason": "Torch version compatibility is not the same as Qwen3.5 SFT embedding unpacking failure." }, { "left": "issue:36032", "right": "issue:43452", "accept": false, - "reason": "T5Tokenizer special-token conflict is unrelated to gguf_file breakage in AutoTokenizer/AutoModel loading." + "reason": "T5Tokenizer add_special_tokens conflict is different from gguf_file breaking from_pretrained for tokenizer/model loading." }, { "left": "issue:44991", "right": "issue:45081", "accept": false, - "reason": "Tokenizer load regression for est-roberta is unrelated to the Mistral regex patch/backend_tokenizer crash." - }, - { - "left": "issue:43582", - "right": "issue:43632", - "accept": false, - "reason": "Apple Silicon caching allocator TypeError is unrelated to the _is_hf_initialized flag regression." + "reason": "Both are tokenizer-load regressions, but for different models and different failure mechanisms." }, { "left": "issue:32090", "right": "issue:39404", "accept": false, - "reason": "Trainer gpu_broadcast_one NoneType error and Whisper return_language pipeline regression are unrelated." + "reason": "Trainer GPU broadcast NoneType error is unrelated to Whisper pipeline return_language behavior." }, { - "left": "issue:44488", - "right": "issue:45081", - "accept": false, - "reason": "sleng-bert loading failure is not the same as the Mistral tokenizer regex-patching crash." - }, - { - "left": "issue:43425", - "right": "issue:45161", + "left": "issue:43582", + "right": "issue:43632", "accept": false, - "reason": "Torch compatibility is unrelated to GPT-OSS MoE tensor-parallel loading behavior." + "reason": "Apple Silicon caching_allocator_warmup TypeError is unrelated to the _is_hf_initialized flag regression." }, { "left": "issue:43957", "right": "issue:44164", "accept": false, - "reason": "Meta-device model loading and extra_state save/from_pretrained handling are different bugs." - }, - { - "left": "issue:43883", - "right": "issue:44488", - "accept": false, - "reason": "Missing all_tied_weights_keys is unrelated to loading cjvt/sleng-bert." + "reason": "Meta-device loading issues and save/from_pretrained extra_state handling are separate serialization/load bugs." }, { - "left": "issue:34567", - "right": "issue:40444", + "left": "issue:44488", + "right": "issue:45081", "accept": false, - "reason": "TrainerState token counter updates and Qwen2.5-VL iterable dataset failures are unrelated." + "reason": "Loading cjvt/sleng-bert and Mistral regex patch crashes are different tokenizer-load failures." }, { "left": "issue:44164", "right": "issue:44479", "accept": false, - "reason": "save/from_pretrained extra_state handling is unrelated to the v5.3.0 video-input regression." + "reason": "extra_state save/load failure is unrelated to Qwen VL video-input regression." }, { - "left": "issue:42947", - "right": "issue:44756", + "left": "issue:40990", + "right": "issue:43421", "accept": false, - "reason": "LoRA/gradient-checkpointing inefficiency is unrelated to Strix Halo mmap/OOM." + "reason": "Perplexity regression on a model is unrelated to tokenizer post-processor update behavior." }, { - "left": "issue:43716", - "right": "issue:45084", + "left": "issue:43425", + "right": "issue:45161", "accept": false, - "reason": "Mistral image-preprocessor dtype mismatch is unrelated to template compilation errors." + "reason": "Torch 2.10 compatibility and GPT-OSS MoE tensor-parallel loading are different issues." }, { - "left": "issue:44488", - "right": "issue:44987", + "left": "issue:43883", + "right": "issue:44488", "accept": false, - "reason": "sleng-bert loading failure and transformers>=5.1.0 loading physical-intelligence/fast are different model-loading bugs." + "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is unrelated to sleng-bert tokenizer loading." }, { "left": "issue:43611", "right": "issue:43644", "accept": false, - "reason": "Both are v5 regressions, but one concerns base_model_prefix loading and the other non-persistent buffer contents; not the same issue." + "reason": "Both are v5 loading regressions, but one is base_model_prefix handling and the other is non-persistent buffer initialization; different defects." }, { "left": "issue:44206", "right": "issue:44220", "accept": false, - "reason": "Both touch audio feature extraction, but the titles do not identify the same concrete failure." + "reason": "Audio feature extraction issues are related in area, but one is an unsupported center arg regression and the other is a generic _torch_extract_fbank_features problem; not clearly the same bug." }, { - "left": "issue:43883", - "right": "issue:44991", + "left": "issue:44488", + "right": "issue:44987", "accept": false, - "reason": "Missing all_tied_weights_keys is unrelated to tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Sleng-bert tokenizer loading and physical-intelligence/fast model loading are different failures." }, { - "left": "issue:34567", - "right": "issue:40444", + "left": "issue:42947", + "right": "issue:44756", + "accept": false, + "reason": "LoRA gradient checkpointing inefficiency is unrelated to Strix Halo mmap/OOM mitigation." + }, + { + "left": "issue:43883", + "right": "issue:44991", "accept": false, - "reason": "num_input_tokens_seen not updating is unrelated to multi-image iterable-dataset finetuning failures." + "reason": "Molmo tied-weights attribute error and est-roberta tokenizer loading failure are unrelated." } ] }, @@ -2293,11 +2292,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:36032", "issue:39404", - "issue:40990", "issue:41553", "issue:42915", "issue:43262", @@ -2333,7 +2331,6 @@ "issue:44610", "issue:44683", "issue:44756", - "issue:44811", "issue:44849", "issue:44987", "issue:44991", @@ -2342,282 +2339,283 @@ "issue:45127", "issue:45216", "issue:45325", - "issue:45357" + "issue:45357", + "issue:45507" ], "soft_pairs": [ - "issue:43541|issue:45084", "issue:43644|issue:43646", "issue:43643|issue:44530", "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:43452|issue:43482", "issue:43317|issue:43482", "issue:36032|issue:42915", - "issue:43452|issue:43482", "issue:43262|issue:43377", "issue:39404|issue:43825", "issue:42915|issue:44530", "issue:44849|issue:45325", "issue:36032|issue:43645", - "issue:44987|issue:44991", "issue:43404|issue:44466", + "issue:45084|issue:45507", + "issue:44987|issue:44991", "issue:43646|issue:44534", - "issue:44403|issue:45081", "issue:43994|issue:44610", - "issue:43645|issue:44534", "issue:43377|issue:43819", - "issue:44451|issue:45081", + "issue:43645|issue:44534", + "issue:44403|issue:45081", "issue:43322|issue:43540", - "issue:44220|issue:44683", "issue:45216|issue:45357", - "issue:41553|issue:43582", - "issue:42915|issue:45084", + "issue:44451|issue:45081", + "issue:44220|issue:44683", "issue:44265|issue:44610", "issue:43632|issue:44756", + "issue:42915|issue:45084", + "issue:41553|issue:43582", + "issue:43404|issue:45127", "issue:44987|issue:45081", "issue:43296|issue:43531", - "issue:43404|issue:45127", "issue:44062|issue:44987", "issue:44479|issue:45325", - "issue:40990|issue:44811", "issue:42915|issue:44291" ], - "prepared_review_unit_hash": "7d8ac6884c07f89919854387b1ba16955f3d457641233c1315f7c57d39492674", + "prepared_review_unit_hash": "9f12d1f54967670c7702848b5669903a2c7df7511503a1e6a7440a1844b99c9a", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 34, - "serialized_chars": 37984, - "estimated_input_tokens": 9496, - "estimated_eval_tokens": 19248 + "serialized_chars": 37380, + "estimated_input_tokens": 9345, + "estimated_eval_tokens": 18946 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous issue reports across unrelated model families and regressions. The only clear duplicate-looking pair is the Qwen3.5 `save_pretrained` regression, where the later report reads like a follow-up to the same incorrect checkpoint serialization bug.", - "confidence": 0.94, - "canonical_issue_reason": "Use issue:45216 as the canonical issue: it is the cleaner, more general regression report for Qwen3.5 `save_pretrained`, and issue:45357 appears to be a later follow-up on the same incorrect visual encoder key serialization problem.", + "summary": "The cluster is mostly heterogeneous; most soft links are false positives across unrelated model-loading, tokenizer, audio, and vision regressions. The only clear duplicate-looking pair is the Qwen3.5 `save_pretrained` checkpoint serialization regression.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45216 is the broadest and clearest representative of the only strong duplicate pair in this set; it describes the underlying Qwen3.5 `save_pretrained` checkpoint corruption regression, while issue:45357 is a narrower follow-up about the same visual-encoder key problem.", "canonical_pr_reason": null, - "best_issue_reason": "Issue:45216 is the best representative issue because it is specific, regression-focused, and directly matches the one strong duplicate pair in the cluster.", + "best_issue_reason": "issue:45216 is the best global issue because it captures the core bug at a higher level and matches the accepted duplicate link more directly than the follow-up.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:43541", - "right": "issue:45084", - "accept": false, - "reason": "Different failures: torch dynamo grouped_mm tracing vs template-node compilation. Same broad area, not the same bug." - }, { "left": "issue:43644", "right": "issue:43646", "accept": false, - "reason": "Both are Transformers v5 init regressions, but one is about non-persistent buffers and the other about custom model initialization; not the same concrete defect." + "reason": "Both mention Transformers v5 regressions, but one is about non-persistent buffers and the other about custom model initialization; different bugs." }, { "left": "issue:43643", "right": "issue:44530", "accept": false, - "reason": "Unrelated code paths: AutoConfig with trust_remote_code vs PagedAttentionCache linear_attention crash on Qwen3.5." + "reason": "`trust_remote_code` config field loss is unrelated to Qwen3.5 PagedAttentionCache/group-type failures." }, { "left": "issue:44451", "right": "issue:44987", "accept": false, - "reason": "Both are model-loading failures, but for different models and likely different root causes." + "reason": "Both are loading failures, but they concern different models and likely different root causes." }, { - "left": "issue:43317", - "right": "issue:43482", + "left": "issue:43541", + "right": "issue:45084", "accept": false, - "reason": "Dequantized device_map offload failure is unrelated to Qwen2.5-GGUF loading in v5." + "reason": "Torch Dynamo tracing on Mixtral and template compilation errors are unrelated code paths." }, { - "left": "issue:36032", - "right": "issue:42915", + "left": "issue:43452", + "right": "issue:43482", "accept": false, - "reason": "Tokenizer add_special_tokens conflict vs Qwen3Moe FineGrainedFP8Config failure are unrelated." + "reason": "Both involve loading, but one is generic gguf_file behavior and the other is a specific Qwen2.5-GGUF/v5 failure; not clearly the same bug." }, { - "left": "issue:43452", + "left": "issue:43317", "right": "issue:43482", "accept": false, - "reason": "Both mention GGUF/loading, but one is a broad gguf_file regression and the other is a specific Qwen2.5-GGUF v5 failure." + "reason": "Device_map offload/dequantization and GGUF tokenizer/model loading are different failure modes." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer method-name conflict and Qwen3Moe FP8 config failure are unrelated." }, { "left": "issue:43262", "right": "issue:43377", "accept": false, - "reason": "Audio chat-template sampling-rate default bug is unrelated to MIMI batching/padding-mask outputs." + "reason": "Audio chat-template sample-rate defaulting and MIMI batching/padding-mask mismatch are separate audio bugs." }, { "left": "issue:39404", "right": "issue:43825", "accept": false, - "reason": "Whisper return_language regression and pipeline translation error messaging are different issues." + "reason": "Whisper pipeline language-return behavior and a generic pipeline translation error message are not the same issue." }, { "left": "issue:42915", "right": "issue:44530", "accept": false, - "reason": "FineGrainedFP8Config load failure and linear_attention PagedAttentionCache crash are different bugs." + "reason": "FineGrainedFP8Config failure and linear_attention group-type crash affect different subsystems." }, { "left": "issue:44849", "right": "issue:45325", "accept": false, - "reason": "Qwen3.5 hidden-states bug is unrelated to Qwen2.5-VL rope index scaling." + "reason": "Qwen3.5 hidden-states output bug and Qwen2.5-VL rope-index scaling bug are distinct model-specific regressions." }, { "left": "issue:36032", "right": "issue:43645", "accept": false, - "reason": "Tokenizer method conflict is unrelated to notebook custom model initialization." + "reason": "Tokenizer special-token conflict and custom model initialization in notebooks are unrelated." }, { - "left": "issue:44987", - "right": "issue:44991", + "left": "issue:43404", + "right": "issue:44466", "accept": false, - "reason": "Both are tokenizer/model loading regressions on different checkpoints; no evidence of the same underlying bug." + "reason": "Missing tied lm_head weights and inconsistent serialization of tied weights are related thematically but not the same concrete bug." }, { - "left": "issue:43404", - "right": "issue:44466", + "left": "issue:45084", + "right": "issue:45507", "accept": false, - "reason": "Tied lm_head weight bug is related to, but distinct from, inconsistent serialization across devices." + "reason": "Template-node compilation failure and GraniteMoEHybrid invalid-method call are different problems." }, { - "left": "issue:43646", - "right": "issue:44534", + "left": "issue:44987", + "right": "issue:44991", "accept": false, - "reason": "Custom model initialization failure vs non-persistent buffer junk are different v5 regressions." + "reason": "Both are loading failures on v5, but they involve different models and likely different tokenizer/config issues." }, { - "left": "issue:44403", - "right": "issue:45081", + "left": "issue:43646", + "right": "issue:44534", "accept": false, - "reason": "Noise on loading a transformer is unrelated to Mistral tokenizer regex patching failure." + "reason": "Custom model initialization and non-persistent buffer junk are unrelated Transformers v5 regressions." }, { "left": "issue:43994", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 nonsensical outputs and OmDet-Turbo processor size mismatch are unrelated." + "reason": "SigLIP2 output quality and OmDet-Turbo input-size mismatch are not the same bug." }, { - "left": "issue:43645", - "right": "issue:44534", + "left": "issue:43377", + "right": "issue:43819", "accept": false, - "reason": "Notebook initialization issue is not the same as junk-filled non-persistent buffers." + "reason": "MIMI padding-mask mismatch and DAC STE mismatch are different audio-model correctness issues." }, { - "left": "issue:43377", - "right": "issue:43819", + "left": "issue:43645", + "right": "issue:44534", "accept": false, - "reason": "MIMI padding-mask batching mismatch and DAC STE mismatch are different audio-model bugs." + "reason": "Notebook custom-model initialization does not match the non-persistent-buffer serialization bug." }, { - "left": "issue:44451", + "left": "issue:44403", "right": "issue:45081", "accept": false, - "reason": "ScandiBERT loading failure and Mistral regex patch crash are unrelated." + "reason": "Generic loading noise and Mistral regex patch crash are unrelated." }, { "left": "issue:43322", "right": "issue:43540", "accept": false, - "reason": "Llava Next loading segfault and Qwen3OmniMoe video-input ValueError are different code paths." - }, - { - "left": "issue:44220", - "right": "issue:44683", - "accept": false, - "reason": "_torch_extract_fbank_features issue is unrelated to compiled flex_attention on torch 2.9." + "reason": "Llava Next segfault and Qwen3OmniMoe video-processing ValueError are distinct multimodal failures." }, { "left": "issue:45216", "right": "issue:45357", "accept": true, - "reason": "Both report the same Qwen3.5 `save_pretrained` checkpoint serialization bug, with the later issue describing the same incorrect visual encoder key handling." + "reason": "Both describe the same Qwen3.5 `save_pretrained` checkpoint corruption / incorrect visual-encoder key serialization regression." }, { - "left": "issue:41553", - "right": "issue:43582", + "left": "issue:44451", + "right": "issue:45081", "accept": false, - "reason": "AutoTokenizer error messaging for Voxtral is unrelated to Apple Silicon caching_allocator_warmup TypeError." + "reason": "A model-loading failure for one checkpoint is not the same as a tokenizer regex-patching crash." }, { - "left": "issue:42915", - "right": "issue:45084", + "left": "issue:44220", + "right": "issue:44683", "accept": false, - "reason": "Qwen3Moe FP8 config load failure and template-node compile error are unrelated." + "reason": "FBank feature extraction and compiled flex_attention on torch>=2.9 are unrelated." }, { "left": "issue:44265", "right": "issue:44610", "accept": false, - "reason": "torch.export export failure with torch_compilable_check is unrelated to OmDet-Turbo processor size mismatch." + "reason": "torch.export with `torch_compilable_check` and OmDet-Turbo preprocessing are separate issues." }, { "left": "issue:43632", "right": "issue:44756", "accept": false, - "reason": "Transformers v5 buffer/init behavior is unrelated to disabling mmap on Strix Halo." + "reason": "`_is_hf_initialized` flag behavior and Strix Halo mmap OOM are different failures." }, { - "left": "issue:44987", - "right": "issue:45081", + "left": "issue:42915", + "right": "issue:45084", "accept": false, - "reason": "Different failures: physical-intelligence/fast load regression vs Mistral tokenizer regex patch crash." + "reason": "Qwen3Moe FP8 loading and template compilation errors have no shared root cause." }, { - "left": "issue:43296", - "right": "issue:43531", + "left": "issue:41553", + "right": "issue:43582", "accept": false, - "reason": "PaddleOCR-VL load failure and Qwen3-MoE sliding_window issue are unrelated." + "reason": "Bad AutoTokenizer error message for Voxtral and AppleSilicon warmup TypeError are unrelated." }, { "left": "issue:43404", "right": "issue:45127", "accept": false, - "reason": "Both involve tied weights, but one is missing lm_head tying in a model class and the other is LoRA merge collapse after vocab extension." + "reason": "Tied lm_head weight bug in Mistral3 is not the same as LoRA merge collapse after vocab extension." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different model families and different failure points; no clear common bug." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL loading and Qwen3-MoE sliding_window behavior are unrelated." }, { "left": "issue:44062", "right": "issue:44987", "accept": false, - "reason": "AddedToken special-arg conflict is unrelated to loading physical-intelligence/fast." + "reason": "AddedToken special-argument conflict is a tokenizer construction bug, not the same as model loading failures for physical-intelligence/fast." }, { "left": "issue:44479", "right": "issue:45325", "accept": false, - "reason": "Video-input regression on Qwen models and Qwen2.5-VL rope index scaling are different bugs." - }, - { - "left": "issue:40990", - "right": "issue:44811", - "accept": false, - "reason": "High perplexity on gpt-oss and Whisper batch_decode skip_special_tokens are unrelated." + "reason": "Both involve Qwen VL temporal/video behavior, but one is a broad video-input regression and the other is a still-image rope-index scaling bug; too different to merge." }, { "left": "issue:42915", "right": "issue:44291", "accept": false, - "reason": "FineGrainedFP8Config failure and unexpected _is_hf_initialized argument during init_empty_weights are different regressions." + "reason": "Qwen3Moe FP8 loading and `_is_hf_initialized` init_empty_weights TypeError are separate regressions." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet, and the only accepted soft-edge pair (issue:45216 / issue:45357) is supported by the titles and timing. The other soft-edge rejections are conservative and do not overclaim duplicates." + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft pair, issue:45216 vs issue:45357, is well supported by the titles as the same Qwen3.5 save_pretrained serialization regression. The other pair verdicts generally avoid overclaiming and are phrased cautiously enough from the available evidence." }, "evaluator_used": true, "retried": false, @@ -2636,12 +2634,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:33666", "issue:35707", "issue:36032", "issue:40444", + "issue:40990", "issue:41720", "issue:42175", "issue:42915", @@ -2656,6 +2655,7 @@ "issue:43611", "issue:43632", "issue:43644", + "issue:43646", "issue:43742", "issue:43749", "issue:43792", @@ -2665,22 +2665,19 @@ "issue:44079", "issue:44265", "issue:44291", - "issue:44387", "issue:44403", "issue:44488", "issue:44534", "issue:44568", "issue:44589", - "issue:44617", "issue:44683", - "issue:44821", + "issue:44811", "issue:44829", "issue:44877", "issue:44991", "issue:45020", "issue:45072", "issue:45081", - "issue:45084", "issue:45325", "issue:45357", "issue:45381", @@ -2688,285 +2685,277 @@ "issue:45491" ], "soft_pairs": [ + "issue:40990|issue:44811", + "issue:43611|issue:44534", "issue:43317|issue:43856", - "issue:42915|issue:44589", "issue:36032|issue:43742", - "issue:43632|issue:44534", - "issue:43611|issue:44534", "issue:40444|issue:45325", - "issue:43531|issue:44877", - "issue:44821|issue:45084", + "issue:42915|issue:44589", + "issue:43632|issue:44534", + "issue:43531|issue:44877", "issue:44683|issue:45020", - "issue:43421|issue:44568", "issue:43818|issue:45357", + "issue:43421|issue:44568", + "issue:36032|issue:41720", + "issue:44829|issue:45479", "issue:43425|issue:44403", "issue:43792|issue:44589", - "issue:36032|issue:41720", "issue:44403|issue:44991", - "issue:44829|issue:45479", - "issue:43377|issue:45491", "issue:33666|issue:40444", - "issue:35707|issue:40444", "issue:44079|issue:44534", - "issue:44291|issue:44589", + "issue:43377|issue:45491", + "issue:35707|issue:40444", "issue:43482|issue:43525", + "issue:44291|issue:44589", "issue:43749|issue:44079", - "issue:43881|issue:44488", "issue:43644|issue:44079", "issue:44079|issue:45072", + "issue:44265|issue:45072", "issue:43632|issue:45072", + "issue:40444|issue:45381", "issue:43531|issue:44291", "issue:42175|issue:43296", - "issue:44265|issue:45072", "issue:41720|issue:45081", - "issue:40444|issue:45381", - "issue:44387|issue:44617" + "issue:45325|issue:45381", + "issue:43611|issue:43646", + "issue:43881|issue:44488" ], - "prepared_review_unit_hash": "a58ec789521f222652a0defed91b568f05dfcde3d6f4b60d0178d2af1e5a298e", + "prepared_review_unit_hash": "3b8671bbcc9ed5b1cbb502bacd4bd19c2e7e75c952b0ede05ebf656a10a0683d", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37475, - "estimated_input_tokens": 9369, - "estimated_eval_tokens": 18994 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 37276, + "estimated_input_tokens": 9319, + "estimated_eval_tokens": 18894 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a grab-bag of distinct Transformers/model-specific regression reports. I only found two clear duplicate pairs: the `_is_hf_initialized` loading regression and the non-persistent-buffer junk regression. The rest are too symptom- or model-specific to merge safely.", - "confidence": 0.89, - "canonical_issue_reason": "Issue 43632 is the best canonical target: it captures the underlying `_is_hf_initialized` regression in Transformers v5, and 44291 is a narrower manifestation of the same loading/init bug.", + "summary": "This cluster is mostly a grab-bag of distinct issues, with a dominant theme of Transformers v5 loading/regression reports. Only the two Qwen2.5-VL position-id reports look like the same underlying bug; the rest are too model- or subsystem-specific to merge as duplicates.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 43611 is the best representative of the cluster\u2019s main theme: broad Transformers v5 loading regressions. It is more central and reusable as an umbrella than the highly specific model-failure reports, even though this cluster is not a clean duplicate set.", "canonical_pr_reason": null, - "best_issue_reason": "43632 is the clearest broad regression report in the set, with the widest applicability and the strongest duplicate match among the issues.", + "best_issue_reason": "43611 is the strongest global issue candidate because it captures the repeated v5 model-loading breakage pattern seen across several nearby reports, without being tied to one narrow model or symptom.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43317", - "right": "issue:43856", + "left": "issue:40990", + "right": "issue:44811", "accept": false, - "reason": "Related loading/memory theme, but one is device_map/offload failure and the other is Qwen3 MoE training memory usage; different code paths." + "reason": "Different failures: language-model perplexity on GPT-OSS vs Whisper batch_decode special-token handling." }, { - "left": "issue:42915", - "right": "issue:44589", + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Both are Transformers v5 regressions, but they describe different code paths: base_model_prefix loading vs non-persistent buffer initialization." + }, + { + "left": "issue:43317", + "right": "issue:43856", "accept": false, - "reason": "Both mention float8/Qwen3Moe-ish failures, but the reported errors and likely code paths differ." + "reason": "Unrelated: device_map auto/offload loading failure vs Qwen3 MoE training memory usage." }, { "left": "issue:36032", "right": "issue:43742", "accept": false, - "reason": "Tokenizer method-name conflict vs model-specific load KeyError; not the same bug." + "reason": "Different models and symptoms: T5 tokenizer class conflict vs MobileLLM loading KeyError." }, { - "left": "issue:43632", - "right": "issue:44534", + "left": "issue:40444", + "right": "issue:45325", "accept": false, - "reason": "Both are Transformers v5 regressions, but `_is_hf_initialized` breakage is a different issue from non-persistent-buffer initialization." + "reason": "Same model family, but different bugs: multi-image iterable dataset failure vs still-image temporal position scaling." }, { - "left": "issue:43611", - "right": "issue:44534", + "left": "issue:42915", + "right": "issue:44589", "accept": false, - "reason": "Base-model-prefix loading failure and non-persistent-buffer junk are separate v5 regressions." + "reason": "Both touch float8/fp8 territory, but the reported failures are different: FineGrainedFP8Config vs missing Float8 storage object." }, { - "left": "issue:40444", - "right": "issue:45325", + "left": "issue:43632", + "right": "issue:44534", "accept": false, - "reason": "Both concern Qwen2.5-VL vision inputs, but one is multi-image finetuning failure and the other is video/still-image position-id scaling." + "reason": "Both are v5 regressions, but one is about the `_is_hf_initialized` flag and the other about non-persistent buffers." }, { "left": "issue:43531", "right": "issue:44877", "accept": false, - "reason": "Qwen3-MoE sliding-window bug vs strict config loading for granite_speech; unrelated failures." + "reason": "Different issues: Qwen3-MoE sliding window behavior vs strict config blocking granite_speech loading." }, { - "left": "issue:44821", - "right": "issue:45084", + "left": "issue:44683", + "right": "issue:45020", "accept": false, - "reason": "AutoImageProcessor URL loading and template-node compilation are unrelated." + "reason": "Different subsystems: compiled flex_attention on torch>=2.9 vs remote_code model loading regressions." }, { - "left": "issue:44683", - "right": "issue:45020", + "left": "issue:43818", + "right": "issue:45357", "accept": false, - "reason": "Compiled flex_attention on torch>=2.9 and remote_code model breakage are different regressions." + "reason": "Different multimodal problems: Video-LLaVA tower/weight sharing vs Qwen3.5 save_pretrained key names." }, { "left": "issue:43421", "right": "issue:44568", "accept": false, - "reason": "Runtime post-processor sync bug vs add_special_tokens not inserting BOS/EOS for one tokenizer; not the same change." + "reason": "Related tokenization area, but not the same bug: runtime special-token updates vs mDeBERTa add_special_tokens output behavior." }, { - "left": "issue:43818", - "right": "issue:45357", + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Completely unrelated: tokenizer attribute conflict vs Qwen3 device-map cudaErrorAssert." + }, + { + "left": "issue:44829", + "right": "issue:45479", "accept": false, - "reason": "Video-tower temporal-attention/weight-sharing bug vs save_pretrained key layout regression; different issues." + "reason": "Both are training pathologies, but the concrete failures differ: flash_attention_3 degeneration vs zero-loss single-label classification." }, { "left": "issue:43425", "right": "issue:44403", "accept": false, - "reason": "Torch compatibility complaint vs generic loading noise; no shared defect." + "reason": "Unrelated: torch 2.10 compatibility issue vs noisy model-loading output." }, { "left": "issue:43792", "right": "issue:44589", "accept": false, - "reason": "Whisper runtime failure and missing Float8 storage type are unrelated." - }, - { - "left": "issue:36032", - "right": "issue:41720", - "accept": false, - "reason": "Tokenizer conflict and Qwen3 auto device-map CUDA assert are unrelated." + "reason": "Different failures and code paths: Whisper runtime issue vs Float8 storage lookup error." }, { "left": "issue:44403", "right": "issue:44991", "accept": false, - "reason": "Generic loading noise vs a specific tokenizer-loading failure for one model." + "reason": "Different symptoms: logging/noise during loading vs tokenizer loading failure for a specific model." }, { - "left": "issue:44829", - "right": "issue:45479", + "left": "issue:33666", + "right": "issue:40444", "accept": false, - "reason": "Flash-attention-3 training degeneration and num_labels=1 zero-loss behavior have different causes." + "reason": "Different Qwen2-VL family problems: multi-GPU training vs iterable-dataset multi-image finetuning." }, { - "left": "issue:43377", - "right": "issue:45491", + "left": "issue:44079", + "right": "issue:44534", "accept": false, - "reason": "MIMI batching/padding-mask bug vs Gemma3 sliding-window NaNs; different model paths." + "reason": "Unrelated bugs: ModelOutput key assignment vs non-persistent buffer initialization." }, { - "left": "issue:33666", - "right": "issue:40444", + "left": "issue:43377", + "right": "issue:45491", "accept": false, - "reason": "Both are Qwen2-VL/Qwen2.5-VL training problems, but the concrete failure modes differ." + "reason": "Both involve batching/padding, but they affect different models and likely different implementations." }, { "left": "issue:35707", "right": "issue:40444", "accept": false, - "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to multi-image finetuning failure." + "reason": "Different generation/mechanism bug vs multimodal finetuning failure." }, { - "left": "issue:44079", - "right": "issue:44534", + "left": "issue:43482", + "right": "issue:43525", "accept": false, - "reason": "ModelOutput key handling and non-persistent-buffer initialization are separate bugs." + "reason": "Different loading failures: Qwen2.5-GGUF/Transformers v5 vs missing pad_token_id on Llama4Config." }, { "left": "issue:44291", "right": "issue:44589", "accept": false, - "reason": "_is_hf_initialized/init_empty_weights regression is unrelated to the Float8 storage lookup error." - }, - { - "left": "issue:43482", - "right": "issue:43525", - "accept": false, - "reason": "Qwen2.5-GGUF/v5 load failure and Llama4Config pad_token_id error are different models and errors." + "reason": "Different initialization failures: init_empty_weights/_is_hf_initialized vs Float8 storage object lookup." }, { "left": "issue:43749", "right": "issue:44079", "accept": false, - "reason": "FSDP CPU RAM efficient loading and ModelOutput key assignment are unrelated." - }, - { - "left": "issue:43881", - "right": "issue:44488", - "accept": false, - "reason": "glm-4v-9b load failure and sleng-bert tokenizer load failure are different model-specific issues." + "reason": "Different areas: FSDP CPU RAM efficient loading vs ModelOutput key assignment." }, { "left": "issue:43644", "right": "issue:44079", "accept": false, - "reason": "Non-persistent-buffer junk and ModelOutput key assignment are unrelated." + "reason": "Different bug classes: non-persistent buffers filled with junk vs ModelOutput key assignment." }, { "left": "issue:44079", "right": "issue:45072", "accept": false, - "reason": "ModelOutput key handling and bfloat16 dtype mismatch are different bugs." + "reason": "Unrelated: ModelOutput key handling vs bfloat16 dtype mismatch in CI." }, { - "left": "issue:43632", + "left": "issue:44265", "right": "issue:45072", "accept": false, - "reason": "_is_hf_initialized loading regression and bfloat16 dtype mismatch are unrelated." + "reason": "Different problems: torch.export with torch_compilable_check vs dtype mismatch during inference." }, { - "left": "issue:43531", - "right": "issue:44291", + "left": "issue:43632", + "right": "issue:45072", "accept": false, - "reason": "Qwen3-MoE sliding-window issue and init_empty_weights/_is_hf_initialized TypeError are unrelated." + "reason": "Different Transformers v5 regressions: `_is_hf_initialized` handling vs bfloat16 dtype mismatch." }, { - "left": "issue:42175", - "right": "issue:43296", + "left": "issue:40444", + "right": "issue:45381", "accept": false, - "reason": "Missing TensorFlow backend in a torch install vs PaddleOCR-VL load failure; different package/setup issues." + "reason": "Same model family, but one is multi-image finetuning failure and the other is video position-id mismatch." }, { - "left": "issue:44265", - "right": "issue:45072", + "left": "issue:43531", + "right": "issue:44291", "accept": false, - "reason": "torch.export/torch_compilable_check failure and bfloat16 inference dtype mismatch are different code paths." + "reason": "Different issues: Qwen3-MoE sliding window behavior vs unexpected `_is_hf_initialized` arg in init_empty_weights." }, { "left": "issue:41720", "right": "issue:45081", "accept": false, - "reason": "Qwen3 auto device-map CUDA assert and Mistral regex patch crash are unrelated." + "reason": "Unrelated: Qwen3 device-map CUDA assert vs Mistral tokenizer regex patch crash." }, { - "left": "issue:40444", + "left": "issue:45325", "right": "issue:45381", - "accept": false, - "reason": "Both touch Qwen2.5-VL positions, but one is multi-image finetuning failure and the other is video position-id scaling." + "accept": true, + "reason": "Both describe the same Qwen2.5-VL position-id / rope-index bug in transformers 5.3.0, just from still-image vs video inputs." }, { - "left": "issue:44387", - "right": "issue:44617", + "left": "issue:43611", + "right": "issue:43646", "accept": false, - "reason": "int4 quantization memory regression and Sam3Video OOM are too generic to treat as the same bug." - }, - { - "left": "issue:43632", - "right": "issue:44291", - "accept": true, - "reason": "Same `_is_hf_initialized` loading regression in Transformers v5; one report is the broader flag breakage, the other is a specific init_empty_weights TypeError." + "reason": "Related only at a high level; the concrete failures differ between base_model_prefix loading and custom model initialization." }, { - "left": "issue:43644", - "right": "issue:44534", - "accept": true, - "reason": "Exact same regression: Transformers v5 fills non-persistent buffers with junk; the titles are just rephrasings of the same bug." + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Different model-loading failures for different models and causes." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The cluster read is grounded in the packet: it is mostly a mix of distinct issues, with a real recurring Transformers v5 loading/regression theme. The soft-edge verdicts are conservative overall, and the single accepted Qwen2.5-VL pair is narrowly tied to the same position-id/rope-index path rather than an obviously broader leap." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -2981,7 +2970,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:34634", "issue:35707", @@ -3000,9 +2989,7 @@ "issue:43526", "issue:43550", "issue:43606", - "issue:43611", "issue:43632", - "issue:43646", "issue:43701", "issue:43720", "issue:43746", @@ -3019,9 +3006,11 @@ "issue:44186", "issue:44265", "issue:44291", + "issue:44387", "issue:44466", "issue:44534", "issue:44610", + "issue:44617", "issue:44683", "issue:44877", "issue:44898", @@ -3029,284 +3018,269 @@ "issue:44991", "issue:45072", "issue:45325", - "issue:45357", - "issue:45381" + "issue:45357" ], "soft_pairs": [ - "issue:43828|issue:44683", - "issue:45325|issue:45381", - "issue:43611|issue:43646", "issue:43994|issue:44898", - "issue:42915|issue:43482", + "issue:43828|issue:44683", "issue:43454|issue:44466", + "issue:44387|issue:44617", + "issue:42915|issue:43482", "issue:43792|issue:43957", "issue:43526|issue:43883", "issue:43701|issue:45325", - "issue:43881|issue:44991", - "issue:43317|issue:43873", "issue:43262|issue:43746", + "issue:43317|issue:43873", + "issue:34634|issue:35707", "issue:44291|issue:44877", + "issue:43881|issue:44991", "issue:43606|issue:44265", - "issue:34634|issue:35707", "issue:44186|issue:44265", - "issue:43632|issue:44164", "issue:43606|issue:43819", + "issue:43632|issue:44164", "issue:35707|issue:41720", - "issue:44610|issue:45325", "issue:43377|issue:44898", + "issue:44610|issue:45325", "issue:43377|issue:44610", "issue:43454|issue:43550", "issue:43454|issue:45072", + "issue:43749|issue:44534", + "issue:39692|issue:40444", "issue:43296|issue:44991", + "issue:44466|issue:45357", "issue:43296|issue:44877", - "issue:39692|issue:40444", "issue:43792|issue:44683", - "issue:43749|issue:44534", - "issue:44466|issue:45357", - "issue:41720|issue:43322", "issue:43454|issue:43720", + "issue:41720|issue:43322", "issue:41720|issue:44977", - "issue:43526|issue:43881", "issue:43404|issue:45357" ], - "prepared_review_unit_hash": "c2e3fd6039a7fcc10b36d68f9ad759488cb5d654b7d361823d3ad0f68ddb758e", + "prepared_review_unit_hash": "904bfbb2d4032d88f1578f794503f2bc0f4e837c1ab9b099fb6d219b842516e1", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 35, - "serialized_chars": 38176, - "estimated_input_tokens": 9544, - "estimated_eval_tokens": 19344 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36982, + "estimated_input_tokens": 9246, + "estimated_eval_tokens": 18748 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No soft-edge pair looks like the same underlying bug/change. The set is a heterogeneous mix of unrelated model-loading, tokenizer/processor, offloading, quantization, and serialization regressions.", - "confidence": 0.96, - "canonical_issue_reason": "No single canonical issue is a safe representative here: the items span many unrelated models and failure modes, with only superficial overlap in broad areas like loading/offloading or tied weights.", + "summary": "The set is heterogeneous; the soft links mostly connect superficially similar bug themes (loading, offloading, weight tying, generation) rather than the same concrete defect. I found no safe duplicate merges among the candidate pairs.", + "confidence": 0.66, + "canonical_issue_reason": "No single true canonical issue emerges from these pairs. If one representative must be chosen, issue 43749 is the strongest standalone bug report with the most discussion and duplicate-triage value around loading/offload regressions.", "canonical_pr_reason": null, - "best_issue_reason": "If one had to be picked, issue 43606 is the closest to a recurring theme (CPU offload/device mismatch), but it still does not represent the cluster well and should not be treated as canonical.", + "best_issue_reason": "Issue 43873 is the broadest open report and the closest umbrella for ongoing offloading/quantization failures, though it still does not subsume the whole set cleanly.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43828", - "right": "issue:44683", + "left": "issue:43994", + "right": "issue:44898", "accept": false, - "reason": "One is an autocast dtype mismatch in Phi-MoE; the other is a torch>=2.9 flex_attention compilation failure." + "reason": "Different models and failure modes: SigLIP2 pipeline/nonsensical outputs vs Perceiver resolution/interpolation behavior." }, { - "left": "issue:45325", - "right": "issue:45381", + "left": "issue:43828", + "right": "issue:44683", "accept": false, - "reason": "Both are Qwen2.5-VL vision-position bugs, but one is still-image temporal scaling and the other is video input position IDs; different code paths." + "reason": "One is an autocast dtype mismatch in Phi-tiny-MoE; the other is a torch>=2.9 flex_attention compilation regression." }, { - "left": "issue:43611", - "right": "issue:43646", + "left": "issue:43454", + "right": "issue:44466", "accept": false, - "reason": "Both are Transformers v5 regressions, but one is base_model_prefix loading and the other is custom model initialization; not the same defect." + "reason": "Both touch tied weights, but one is an init/tie_weights bug causing bad generation while the other is device-dependent save_pretrained serialization." }, { - "left": "issue:43994", - "right": "issue:44898", + "left": "issue:44387", + "right": "issue:44617", "accept": false, - "reason": "Different models and failures: SigLIP2 nonsense outputs vs Perceiver non-default-resolution classification." + "reason": "Both mention OOM, but one is int4 memory growth in Transformers 5.x and the other is a Sam3Video runtime OOM; different concrete causes." }, { "left": "issue:42915", "right": "issue:43482", "accept": false, - "reason": "Qwen3Moe FP8 failure and Qwen2.5-GGUF loading failure are unrelated." - }, - { - "left": "issue:43454", - "right": "issue:44466", - "accept": false, - "reason": "Both involve lm_head/tied weights, but one is incorrect tying causing bad generation and the other is inconsistent serialization by device." + "reason": "Different artifacts and failures: FineGrainedFP8Config load failure vs GGUF loading failure in transformers v5." }, { "left": "issue:43792", "right": "issue:43957", "accept": false, - "reason": "Whisper runtime failure and meta-device loading breakage are different model/loading issues." + "reason": "Whisper runtime failure is unrelated to the meta-device loading regression affecting other models." }, { "left": "issue:43526", "right": "issue:43883", "accept": false, - "reason": "BeitImageProcessorFast label reduction and Molmo tied-weights attribute errors are unrelated." + "reason": "BEiT image processor label reduction bug is unrelated to the Molmo missing all_tied_weights_keys AttributeError." }, { "left": "issue:43701", "right": "issue:45325", "accept": false, - "reason": "resume_from_checkpoint key mismatch is unrelated to Qwen2.5-VL rope/position-id scaling." + "reason": "Checkpoint resume key mismatch and Qwen2.5-VL rope-index scaling are different code paths." }, { - "left": "issue:43881", - "right": "issue:44991", + "left": "issue:43262", + "right": "issue:43746", "accept": false, - "reason": "GLM-4V loading failure and est-roberta tokenizer loading failure are different model/component regressions." + "reason": "Audio sampling-rate defaulting in apply_chat_template is unrelated to GraniteSpeech PEFT checkpoint loading." }, { "left": "issue:43317", "right": "issue:43873", "accept": false, - "reason": "Both touch offloading/quantization, but one is dequantized model loading with device_map=auto and the other is a broader offloading expectation issue; not clearly the same bug." + "reason": "Both are offloading/quantization themed, but one is a dequantized gpu+cpu offload load failure and the other is a broader quantization offloading complaint; not the same concrete bug." }, { - "left": "issue:43262", - "right": "issue:43746", + "left": "issue:34634", + "right": "issue:35707", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and GraniteSpeech PEFT checkpoint loading are unrelated." + "reason": "BarkProcessor voice_preset and progressive generation with inputs_embeds/past_key_values are unrelated failures." }, { "left": "issue:44291", "right": "issue:44877", "accept": false, - "reason": "Unexpected _is_hf_initialized argument during init_empty_weights and strict granite_speech config loading are different problems." + "reason": "Unexpected _is_hf_initialized argument handling is distinct from strict config rejection for granite_speech." }, { - "left": "issue:43606", - "right": "issue:44265", + "left": "issue:43881", + "right": "issue:44991", "accept": false, - "reason": "CPU offload device mismatch in Bark is unrelated to torch.export failures with torch_compilable_check." + "reason": "glm-4v-9b loading failure and EMBEDDIA tokenizer loading failure are separate model-specific issues." }, { - "left": "issue:34634", - "right": "issue:35707", + "left": "issue:43606", + "right": "issue:44265", "accept": false, - "reason": "BarkProcessor voice_preset and progressive generation with inputs_embeds/past_key_values are unrelated." + "reason": "CPU-offload device mismatch for Bark-small is unrelated to torch.export failures involving torch_compilable_check." }, { "left": "issue:44186", "right": "issue:44265", "accept": false, - "reason": "Tokenizer crashes on NER padding/truncation and torch.export failures are different subsystems." + "reason": "LayoutLMv2 tokenizer crashes on NER/padding are unrelated to torch.export/torch_compilable_check." }, { - "left": "issue:43632", - "right": "issue:44164", + "left": "issue:43606", + "right": "issue:43819", "accept": false, - "reason": "_is_hf_initialized flag breakage and save/from_pretrained extra_state handling are separate loading/saving regressions." + "reason": "Device mismatch in Bark offload and DAC.from_latents vs forward mismatch are different bugs." }, { - "left": "issue:43606", - "right": "issue:43819", + "left": "issue:43632", + "right": "issue:44164", "accept": false, - "reason": "Bark CPU offload device mismatch and DAC.from_latents forward mismatch are unrelated." + "reason": "_is_hf_initialized regression and extra_state save/from_pretrained handling are related to loading internals but not the same defect." }, { "left": "issue:35707", "right": "issue:41720", "accept": false, - "reason": "Progressive generation inputs_embeds/past_key_values and Qwen3 auto device mapping CUDA assert are unrelated." + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen3 auto device-map CUDA asserts." }, { - "left": "issue:44610", - "right": "issue:45325", + "left": "issue:43377", + "right": "issue:44898", "accept": false, - "reason": "OmDet-Turbo processor resolution mismatch is not the same as Qwen2.5-VL still-image temporal position scaling." + "reason": "MIMI batching/padding-mask mismatch is unrelated to Perceiver non-default-resolution failures." }, { - "left": "issue:43377", - "right": "issue:44898", + "left": "issue:44610", + "right": "issue:45325", "accept": false, - "reason": "MIMI batch-vs-single padding-mask bug and Perceiver resolution bug are unrelated." + "reason": "OmDet-Turbo processor size mismatch and Qwen2.5-VL temporal position scaling are different processor/model bugs." }, { "left": "issue:43377", "right": "issue:44610", "accept": false, - "reason": "MIMI encoder padding-mask behavior and OmDet-Turbo processor size mismatch are different issues." + "reason": "Batched-vs-single MIMI encoder outputs and OmDet-Turbo image-size mismatch are separate issues." }, { "left": "issue:43454", "right": "issue:43550", "accept": false, - "reason": "AyaVision tied-weights bug and Bamba torch.compile SDPA failure are unrelated." + "reason": "lm_head tying/serialization bug is unrelated to Bamba torch.compile SDPA failure." }, { "left": "issue:43454", "right": "issue:45072", "accept": false, - "reason": "AyaVision weight tying/serialization issue is unrelated to dtype mismatches in SwitchTransformers/TimmWrapperModel." - }, - { - "left": "issue:43296", - "right": "issue:44991", - "accept": false, - "reason": "PaddleOCR-VL load failure and est-roberta tokenizer loading failure are different model-specific regressions." + "reason": "Weight-tying bug and bfloat16 dtype mismatch are different failure modes." }, { - "left": "issue:43296", - "right": "issue:44877", + "left": "issue:43749", + "right": "issue:44534", "accept": false, - "reason": "PaddleOCR-VL loading in vLLM and granite_speech strict config loading are unrelated." + "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk filling are both v5/loading topics but not the same code-path bug." }, { "left": "issue:39692", "right": "issue:40444", "accept": false, - "reason": "SigLIP2 documentation errors and Qwen2.5-VL multi-image iterable fine-tuning failure are different problems." + "reason": "SigLIP2 doc example errors and Qwen2.5-VL iterable-dataset finetuning failure are unrelated." }, { - "left": "issue:43792", - "right": "issue:44683", + "left": "issue:43296", + "right": "issue:44991", "accept": false, - "reason": "Whisper model runtime failure and compiled flex_attention failure are unrelated." + "reason": "PaddleOCR-VL load failure in vLLM is unrelated to EMBEDDIA tokenizer loading in transformers v5." }, { - "left": "issue:43749", - "right": "issue:44534", + "left": "issue:44466", + "right": "issue:45357", "accept": false, - "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk in v5 are distinct regressions." + "reason": "Both involve serialization, but one is tied lm_head.weight serialization and the other is incorrect visual encoder keys in Qwen3.5 save_pretrained." }, { - "left": "issue:44466", - "right": "issue:45357", + "left": "issue:43296", + "right": "issue:44877", "accept": false, - "reason": "Both concern serialization/tied weights, but one is device-dependent lm_head serialization and the other is wrong visual encoder keys for Qwen3.5." + "reason": "PaddleOCR-VL/vLLM loading failure is unrelated to strict config handling for granite_speech." }, { - "left": "issue:41720", - "right": "issue:43322", + "left": "issue:43792", + "right": "issue:44683", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and Llava Next segfault on load are different failures." + "reason": "Whisper loading/runtime failure and compiled flex_attention failure are unrelated." }, { "left": "issue:43454", "right": "issue:43720", "accept": false, - "reason": "AyaVision tied weights and BitNet packed-weight unpacking during accelerate loading are unrelated." + "reason": "Tied lm_head weight bug and BitNet packed-weight unpacking during accelerate loading are different defects." }, { "left": "issue:41720", - "right": "issue:44977", + "right": "issue:43322", "accept": false, - "reason": "Qwen3 auto device mapping CUDA assert and Qwen3.5 flash-attention generation issue are unrelated." + "reason": "Qwen3 auto device-map CUDA assert and Llava Next segmentation fault are separate model-loading failures." }, { - "left": "issue:43526", - "right": "issue:43881", + "left": "issue:41720", + "right": "issue:44977", "accept": false, - "reason": "BeitImageProcessorFast label reduction and GLM-4V model loading failure are unrelated." + "reason": "Auto device mapping CUDA assert and flash-attention generation failure for Qwen3.5 are different issues." }, { "left": "issue:43404", "right": "issue:45357", "accept": false, - "reason": "Mistral3 lm_head tying and Qwen3.5 visual encoder save_pretrained keys are different model-specific bugs." + "reason": "Mistral3 lm_head tying bug is unrelated to Qwen3.5 visual-encoder save_pretrained key regression." } ] }, @@ -3328,12 +3302,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ - "issue:30064", "issue:33357", "issue:34567", - "issue:34634", "issue:35707", "issue:36010", "issue:36331", @@ -3347,6 +3319,7 @@ "issue:43322", "issue:43388", "issue:43482", + "issue:43526", "issue:43531", "issue:43540", "issue:43550", @@ -3357,6 +3330,7 @@ "issue:43720", "issue:43782", "issue:43873", + "issue:43881", "issue:44186", "issue:44292", "issue:44361", @@ -3384,268 +3358,275 @@ "issue:44186|issue:45072", "issue:34567|issue:40990", "issue:43873|issue:44756", + "issue:43526|issue:43881", "issue:43577|issue:43708", "issue:42617|issue:42915", "issue:44877|issue:44991", - "issue:44479|issue:44918", - "issue:44610|issue:44898", "issue:43550|issue:43720", + "issue:44610|issue:44898", + "issue:44479|issue:44918", "issue:44186|issue:44610", - "issue:42994|issue:43482", "issue:36331|issue:42915", "issue:36331|issue:42617", + "issue:42994|issue:43482", + "issue:43316|issue:44913", "issue:44671|issue:44743", "issue:42915|issue:43317", - "issue:43316|issue:44913", "issue:43550|issue:43606", + "issue:41720|issue:42617", "issue:43531|issue:43572", "issue:44292|issue:44479", - "issue:41720|issue:42617", "issue:41720|issue:45161", - "issue:43540|issue:43782", "issue:43388|issue:45059", "issue:44186|issue:44361", + "issue:43540|issue:43782", "issue:41720|issue:43873", - "issue:42915|issue:43322", "issue:45406|issue:45464", + "issue:42915|issue:43322", "issue:36010|issue:42617", + "issue:36331|issue:41720", "issue:44514|issue:44545", - "issue:30064|issue:34634", "issue:33357|issue:35707", "issue:43606|issue:43720" ], - "prepared_review_unit_hash": "e30f0445ae8f9e3c23bfc32bb24b4f64ca79e66c3e5ac185f5bf233160a32eb0", + "prepared_review_unit_hash": "1668ae07276aef46dcd4c82c43be2b98a6c326ccff952fe5b8939494fd96b3a8", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37358, - "estimated_input_tokens": 9340, - "estimated_eval_tokens": 18936 + "soft_pair_count": 34, + "serialized_chars": 37685, + "estimated_input_tokens": 9422, + "estimated_eval_tokens": 19100 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly unrelated issue reports. The only clear duplicate pair is the two identical `Qwen2_5_VLProcessor.apply_chat_template` crashes; the rest are distinct bugs in different models, trainers, or serving paths.", - "confidence": 0.91, - "canonical_issue_reason": "issue:44514 is the best issue representative because it is the earlier, more complete report of the exact same batched `padding=False` crash as issue:44545.", + "summary": "This cluster is mostly a set of unrelated bug reports across different models, processors, and trainer/config paths. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash (issues 44514 and 44545).", + "confidence": 0.96, + "canonical_issue_reason": "issue:44545 is the clearest representative of the only confirmed duplicate bug: the same Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False, with the cleaner title and later corroborating report.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44514 is the strongest issue candidate overall because it is the clearest exact duplicate target in the set and has the richer original report.", + "best_issue_reason": "issue:44545 is the best canonical issue candidate because it precisely describes the exact duplicate bug and is the more specific, readable report.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44423", "right": "issue:45406", "accept": false, - "reason": "Both are serve-time AttributeErrors, but they hit different components (`str.to` in multimodal batching vs missing `_tokenizer` on `Gemma4Processor`) and do not look like the same bug." + "reason": "Both are serve/multimodal crashes, but the failing paths differ: continuous-batching Qwen3.5 AttributeError vs Gemma4Processor missing _tokenizer." }, { "left": "issue:44186", "right": "issue:45072", "accept": false, - "reason": "Different failures in different areas: LayoutLMv2 tokenizer crashes vs dtype mismatches in SwitchTransformers/TimmWrapperModel inference." + "reason": "Different bugs: LayoutLMv2Tokenizer NER/padding crash versus dtype mismatch in SwitchTransformers/TimmWrapperModel inference." }, { "left": "issue:34567", "right": "issue:40990", "accept": false, - "reason": "`num_input_tokens_seen` not updating in TrainerState is unrelated to GPT-OSS perplexity evaluation quality." + "reason": "TrainerState token-count tracking is unrelated to GPT-OSS perplexity evaluation." }, { "left": "issue:43873", "right": "issue:44756", "accept": false, - "reason": "Both mention loading/offloading, but one is quantization offload behavior and the other is Strix Halo mmap/OOM mitigation; different bugs." + "reason": "Both mention offloading/memory, but one is quantization offloading behavior and the other is mmap OOM on Strix Halo; not the same bug." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug is unrelated to glm-4v-9b loading failure." }, { "left": "issue:43577", "right": "issue:43708", "accept": false, - "reason": "Blip2 dtype persistence and Trainer `max_steps` recomputation are separate code paths with different symptoms." + "reason": "Different regression classes: Blip2 dtype propagation vs Trainer resume/max_steps calculation." }, { "left": "issue:42617", "right": "issue:42915", "accept": false, - "reason": "Different model/runtime failures: `3d_parallel.py` startup vs Qwen3MoE + FineGrainedFP8Config." + "reason": "Both are model-run failures, but one is a 3d_parallel.py runner issue and the other is Qwen3Moe FineGrainedFP8Config loading." }, { "left": "issue:44877", "right": "issue:44991", "accept": false, - "reason": "Strict config loading for `granite_speech` and tokenizer loading for `est-roberta` are unrelated model/config bugs." + "reason": "Both concern config/tokenizer loading after v5, but one is strict model config handling and the other is a tokenizer-loading regression for a specific model." }, { - "left": "issue:44479", - "right": "issue:44918", + "left": "issue:43550", + "right": "issue:43720", "accept": false, - "reason": "Video-input regression across Qwen VL models is separate from the TRL SFT input-embedding unpacking failure." + "reason": "Different code paths: Bamba torch.compile with SDPA versus BitNet packed-weight unpacking during accelerate loading." }, { "left": "issue:44610", "right": "issue:44898", "accept": false, - "reason": "Both concern image sizes, but one is a processor/model size mismatch while the other is a Perceiver interpolation failure; not the same concrete bug." + "reason": "Different vision-model bugs: OmDet processor size mismatch versus Perceiver non-default-resolution interpolation failure." }, { - "left": "issue:43550", - "right": "issue:43720", + "left": "issue:44479", + "right": "issue:44918", "accept": false, - "reason": "SDPA/torch.compile failure in Bamba is unrelated to packed-weight unpacking during accelerate loading in BitNet." + "reason": "Both involve Qwen-family regressions, but one is video-input handling and the other is trl SFT unpacking of embeddings." }, { "left": "issue:44186", "right": "issue:44610", "accept": false, - "reason": "Tokenizer NER/padding crashes and image-resolution processor mismatch are different issues." - }, - { - "left": "issue:42994", - "right": "issue:43482", - "accept": false, - "reason": "Quantized model saving failure and GGUF loading failure are different operations and likely different root causes." + "reason": "Different processors/models and failures: LayoutLMv2Tokenizer crash versus OmDet image-size mismatch." }, { "left": "issue:36331", "right": "issue:42915", "accept": false, - "reason": "CustomTrainer `compute_loss` signature breakage is unrelated to Qwen3MoE FP8 loading." + "reason": "Trainer custom compute_loss signature mismatch is unrelated to Qwen3Moe FineGrainedFP8Config loading." }, { "left": "issue:36331", "right": "issue:42617", "accept": false, - "reason": "Trainer API keyword mismatch and `3d_parallel.py` failure are distinct problems." + "reason": "CustomTrainer keyword-argument regression is unrelated to the 3d_parallel.py execution issue." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure is not the same as Qwen2.5-GGUF loading failure under transformers v5." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Different config serialization bugs: Gemma3TextConfig API mismatch vs GPTNeoX rotary_pct not persisting on reload." }, { "left": "issue:44671", "right": "issue:44743", "accept": false, - "reason": "CamemBERT masked-LM prediction regression and Qwen3.5 recurrent-state reset are separate model behaviors." + "reason": "CamemBERT masked-LM prediction regression is unrelated to Qwen3.5 recurrent-state cache reset." }, { "left": "issue:42915", "right": "issue:43317", "accept": false, - "reason": "FP8 config failure for Qwen3MoE and dequantized model offload failure are both loading-related but not the same code-path problem." + "reason": "FineGrainedFP8Config model load failure is unrelated to dequantized model loading with device_map=auto offload." }, { - "left": "issue:43316", - "right": "issue:44913", + "left": "issue:43550", + "right": "issue:43606", "accept": false, - "reason": "Gemma3TextConfig API discrepancy and GPTNeoX `rotary_pct` reload regression are unrelated config issues." + "reason": "Both are runtime/inference failures, but the model families and failures differ: Bamba SDPA compile issue vs bark-small CPU offload device mismatch." }, { - "left": "issue:43550", - "right": "issue:43606", + "left": "issue:41720", + "right": "issue:42617", "accept": false, - "reason": "Both are device/compile/offload themed, but the affected models and failure modes differ materially." + "reason": "Qwen3 auto device mapping cuda assert is a different failure from the 3d_parallel.py runner problem." }, { "left": "issue:43531", "right": "issue:43572", "accept": false, - "reason": "Qwen3-MoE sliding-window handling and StableLmConfig missing `pad_token_idx` are unrelated." + "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig pad_token_idx missing are unrelated configuration bugs." }, { "left": "issue:44292", "right": "issue:44479", "accept": false, - "reason": "Qwen-3-8B-NVFP4 run failure is not the same as the VL video-input regression affecting multiple Qwen VL variants." - }, - { - "left": "issue:41720", - "right": "issue:42617", - "accept": false, - "reason": "CUDA assert during auto device mapping and `3d_parallel.py` startup failure are different issues." + "reason": "Qwen-3-8B-NVFP4 runtime error is unrelated to the Qwen2_5_vl / qwen3_vl video-input regression." }, { "left": "issue:41720", "right": "issue:45161", "accept": false, - "reason": "Both involve parallelism/device placement, but the model, failure point, and likely fix are different." - }, - { - "left": "issue:43540", - "right": "issue:43782", - "accept": false, - "reason": "Video input processing error in Qwen3OmniMoe is unrelated to `from_pretrained(weight_only=True)` loading failure in Qwen3VL." + "reason": "Same broad GPT/Qwen MoE area, but the concrete failures differ: auto device mapping cuda assert versus tensor-parallel-only not working." }, { "left": "issue:43388", "right": "issue:45059", "accept": false, - "reason": "Metric gathering truncation with tuple labels is unrelated to SAM3 text/box behavior." + "reason": "Metric-label dropping in evaluation is unrelated to SAM3 PCS text/bounding-box behavior." }, { "left": "issue:44186", "right": "issue:44361", "accept": false, - "reason": "Different tokenizer bugs affecting different models and task types." + "reason": "Different tokenizer crashes for different models: LayoutLMv2Tokenizer NER/padding issue versus MLukeTokenizer task AttributeError." }, { - "left": "issue:41720", - "right": "issue:43873", + "left": "issue:43540", + "right": "issue:43782", "accept": false, - "reason": "CUDA assert on Qwen3 auto device mapping is not the same as quantization offloading misbehavior." + "reason": "Qwen3OmniMoe video-input ValueError and Qwen3VL weight_only loading error are distinct model-loading/processing bugs." }, { - "left": "issue:42915", - "right": "issue:43322", + "left": "issue:41720", + "right": "issue:43873", "accept": false, - "reason": "Qwen3MoE FP8 failure and Llava Next segmentation fault are unrelated model-loading crashes." + "reason": "Both touch Qwen model execution/offloading, but one is a cuda assert with auto device mapping and the other is quantization offloading behavior." }, { "left": "issue:45406", "right": "issue:45464", "accept": false, - "reason": "Both hit serving paths, but they involve different models and different errors (`_tokenizer` missing vs streaming inference failure)." + "reason": "Both are serve/API crashes, but Gemma4Processor missing _tokenizer is unrelated to Qwen3.5 streaming inference failure." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segmentation fault are different models and different crash modes." }, { "left": "issue:36010", "right": "issue:42617", "accept": false, - "reason": "GenerationMixin import regression and `3d_parallel.py` runtime failure are unrelated." + "reason": "GenerationMixin import error is unrelated to running 3d_parallel.py." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer custom-loss signature mismatch is unrelated to Qwen3 auto device mapping cuda errors." }, { "left": "issue:44514", "right": "issue:44545", "accept": true, - "reason": "These are the same bug: identical `Qwen2_5_VLProcessor.apply_chat_template` crash on batched input with `padding=False`, just reported twice." - }, - { - "left": "issue:30064", - "right": "issue:34634", - "accept": false, - "reason": "ImageProcessor segmentation-map failure and BarkProcessor `voice_preset` bug are unrelated." + "reason": "Exact same bug: Qwen2_5_VLProcessor.apply_chat_template crashes on batched input when padding=False; these are duplicate reports of the same path." }, { "left": "issue:33357", "right": "issue:35707", "accept": false, - "reason": "MacOS bus error with CLIP and progressive generation with `inputs_embeds`/`past_key_values` are distinct failures." + "reason": "MacOS bus error on CLIP model loading is unrelated to progressive generation with inputs_embeds and past_key_values." }, { "left": "issue:43606", "right": "issue:43720", "accept": false, - "reason": "CPU offload device mismatch in Bark and packed-weight unpacking in BitNet are different bugs." + "reason": "Different offload/loader bugs: bark-small CPU offload device mismatch versus BitNet packed-weights unpacking during accelerate loading." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the only clearly duplicate-looking pair is issue:44514 / issue:44545, whose titles match closely and whose metadata supports 44514 as the earlier, longer report. The remaining soft-edge rejections are conservative and stay within what can be inferred from titles/metadata alone." + "feedback": "The summary is grounded in the packet, and the duplicate call on 44514/44545 is well supported. The other pair verdicts are conservative and the distinctions are specific to the titles/body context provided." }, "evaluator_used": true, "retried": false, @@ -3664,8 +3645,9 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ + "issue:30064", "issue:34567", "issue:34634", "issue:35707", @@ -3675,7 +3657,6 @@ "issue:43381", "issue:43421", "issue:43531", - "issue:43540", "issue:43550", "issue:43596", "issue:43606", @@ -3684,9 +3665,9 @@ "issue:43746", "issue:43782", "issue:43844", + "issue:43873", "issue:43927", "issue:43994", - "issue:44008", "issue:44062", "issue:44077", "issue:44186", @@ -3717,254 +3698,258 @@ ], "soft_pairs": [ "issue:41720|issue:43746", + "issue:30064|issue:34634", "issue:43550|issue:43653", - "issue:44492|issue:44509", - "issue:44625|issue:45479", "issue:44368|issue:45357", - "issue:43421|issue:43927", + "issue:44625|issue:45479", + "issue:44492|issue:44509", "issue:43606|issue:43653", - "issue:44918|issue:44936", - "issue:43782|issue:45381", - "issue:44451|issue:44991", + "issue:43421|issue:43927", "issue:34634|issue:41720", - "issue:43381|issue:43844", "issue:44898|issue:45072", + "issue:43782|issue:45381", + "issue:44451|issue:44991", "issue:44062|issue:44451", "issue:34567|issue:35707", + "issue:43381|issue:43844", + "issue:42886|issue:44843", + "issue:44077|issue:44625", "issue:43421|issue:43653", + "issue:44918|issue:44936", "issue:43531|issue:44589", - "issue:42886|issue:44843", "issue:44987|issue:45042", - "issue:44077|issue:44625", "issue:43994|issue:44671", - "issue:43421|issue:44186", "issue:39692|issue:44610", + "issue:43421|issue:44186", "issue:43596|issue:44220", "issue:43531|issue:43645", "issue:45081|issue:45356", - "issue:45183|issue:45479", - "issue:43540|issue:44008", "issue:43746|issue:45381", + "issue:45183|issue:45479", + "issue:39692|issue:43873", "issue:44220|issue:44661" ], - "prepared_review_unit_hash": "cfc136971dfa8660ec33698be37825ed78ab76a32bd6bfda43d0188dda5c373e", + "prepared_review_unit_hash": "1a20c543544ef84e329bc5ca266c0f232f714acd92d0ca23402ae0569d84ea8e", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 30, - "serialized_chars": 37047, - "estimated_input_tokens": 9262, - "estimated_eval_tokens": 18780 + "soft_pair_count": 31, + "serialized_chars": 37251, + "estimated_input_tokens": 9313, + "estimated_eval_tokens": 18882 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is mostly heterogeneous, with one clear duplicate-like pair around HF_HUB_OFFLINE tokenizer loading regression. Most other pairs share only broad subsystem similarity and should not be merged.", - "confidence": 0.78, - "canonical_issue_reason": "Issue 44843 is the best canonical issue because it is the most specific and root-cause-oriented report for the offline tokenizer regression; issue 42886 reads like the broader symptom case.", + "summary": "This cluster is highly heterogeneous: most items are unrelated tokenizer, model-loading, training, vision, or docs bugs that only share broad library vocabulary. None of the soft edges looks like the same underlying issue or a mergeable PR pair.", + "confidence": 0.96, + "canonical_issue_reason": "No clear canonical issue: the items span many unrelated code paths, and the soft edges are only superficial title similarity rather than one underlying bug.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44843 is the strongest representative of the only plausible duplicate pair in this set and gives the clearest actionable diagnosis.", + "best_issue_reason": "Weak fit at best: issue 43421 is the broadest tokenizer-internals report, but it still does not represent the rest of the cluster well.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41720", "right": "issue:43746", "accept": false, - "reason": "Different models and failures: Qwen3 device-mapping CUDA assert vs GraniteSpeech PEFT local checkpoint loading." + "reason": "Different failures: Qwen3 device-map CUDA assert vs GraniteSpeech PEFT local-checkpoint loading." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Unrelated domains: image segmentation-map processing vs Bark voice preset handling." }, { "left": "issue:43550", "right": "issue:43653", "accept": false, - "reason": "Unrelated bugs: torch.compile/SDPA failure for Bamba-9B-v2 vs BigBirdTokenizer special-token registration." + "reason": "Different subsystems: torch.compile/SDPA model bug vs tokenizer special-token registration/decode issue." }, { - "left": "issue:44492", - "right": "issue:44509", + "left": "issue:44368", + "right": "issue:45357", "accept": false, - "reason": "A typo/documentation issue is not the same as stale v5 pipeline task docs." + "reason": "Both involve Qwen3.5, but one is a tied-embeddings warning and the other is a save_pretrained serialization bug; not the same code path." }, { "left": "issue:44625", "right": "issue:45479", "accept": false, - "reason": "Both involve classification config, but one is num_labels propagation in Qwen3.5 and the other is zero-loss behavior across multiple models; not the same bug." + "reason": "Both touch classification config, but one is num_labels propagation and the other is zero-loss behavior from problem_type handling." }, { - "left": "issue:44368", - "right": "issue:45357", + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A typo in cache-strategy docs is unrelated to removed pipeline-task documentation." + }, + { + "left": "issue:43606", + "right": "issue:43653", "accept": false, - "reason": "Different Qwen3.5 problems: a warning about tie_word_embeddings vs incorrect visual encoder keys on save/load." + "reason": "CPU offload device mismatch and BigBirdTokenizer special-token decode are separate bugs." }, { "left": "issue:43421", "right": "issue:43927", "accept": false, - "reason": "Tokenizer backend post-processor updates and config persistence/custom token IDs are different code paths and symptoms." + "reason": "Both concern tokenization/config mutation, but one is runtime post-processor refresh and the other is save/load custom token IDs." }, { - "left": "issue:43606", - "right": "issue:43653", + "left": "issue:34634", + "right": "issue:41720", "accept": false, - "reason": "CPU offload device mismatch for Bark is unrelated to BigBirdTokenizer special-token decode behavior." + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping are unrelated." }, { - "left": "issue:44918", - "right": "issue:44936", + "left": "issue:44898", + "right": "issue:45072", "accept": false, - "reason": "Both mention trainers, but one is SFT unpacking with inputs_embeds/past_key_values and the other is evaluate-after-train lifecycle failure." + "reason": "Different model bugs: Perceiver image-size handling vs dtype mismatch in SwitchTransformers/TimmWrapperModel." }, { "left": "issue:43782", "right": "issue:45381", "accept": false, - "reason": "Different multimodal Qwen failures: weight_only load error vs wrong video position ids." + "reason": "Both are multimodal Qwen bugs, but one is from_pretrained/weight_only loading and the other is video vision_position_ids." }, { "left": "issue:44451", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer loading regressions, but they affect different models and likely different failure causes." + "reason": "Both are tokenizer-loading regressions, but for different models and different failure mechanisms." }, { - "left": "issue:34634", - "right": "issue:41720", + "left": "issue:44062", + "right": "issue:44451", "accept": false, - "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping are unrelated." + "reason": "AddedToken special-arg duplication is not the same bug as loading ScandiBERT." }, { - "left": "issue:43381", - "right": "issue:43844", + "left": "issue:34567", + "right": "issue:35707", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction is not the same as gradient explosion under ZeRO-3." + "reason": "Trainer token accounting and progressive generation with inputs_embeds/past_key_values are unrelated." }, { - "left": "issue:44898", - "right": "issue:45072", + "left": "issue:43381", + "right": "issue:43844", "accept": false, - "reason": "Different inference bugs in different model families: Perceiver resolution handling vs dtype mismatch in SwitchTransformers/TimmWrapperModel." + "reason": "Gradient checkpointing in eval mode and abnormal gradients under ZeRO-3 are distinct training issues." }, { - "left": "issue:44062", - "right": "issue:44451", + "left": "issue:42886", + "right": "issue:44843", "accept": false, - "reason": "AddedToken argument duplication is unrelated to ScandiBERT load failure." + "reason": "Both break HF_HUB_OFFLINE, but one is cache loading and the other is an unconditional model_info call in Mistral regex patching." }, { - "left": "issue:34567", - "right": "issue:35707", + "left": "issue:44077", + "right": "issue:44625", "accept": false, - "reason": "TrainerState token-count tracking and progressive generation with inputs_embeds/past_key_values are unrelated." + "reason": "PatchTSMixer post_init policy and Qwen3.5 num_labels propagation are unrelated." }, { "left": "issue:43421", "right": "issue:43653", "accept": false, - "reason": "Both involve tokenizer special tokens, but one is runtime post-processor syncing and the other is missing mask-token registration; not the same bug." + "reason": "Both mention special tokens, but one is runtime post-processor syncing and the other is a model-specific tokenizer registration bug." }, { - "left": "issue:43531", - "right": "issue:44589", + "left": "issue:44918", + "right": "issue:44936", "accept": false, - "reason": "Qwen3-MoE sliding-window behavior is unrelated to the Float8 storage object error." + "reason": "Trainer bugs, but different ones: SFT inputs_embeds unpacking vs evaluate-after-train failure." }, { - "left": "issue:42886", - "right": "issue:44843", - "accept": true, - "reason": "Both describe HF_HUB_OFFLINE tokenizer-loading failures in 4.57.3 and fit the same offline Mistral-regex/model_info regression." + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Sliding-window Qwen3-MoE behavior and Float8 storage lookup failure are unrelated." }, { "left": "issue:44987", "right": "issue:45042", "accept": false, - "reason": "Different regression classes: model load failure for physical-intelligence/fast vs image-processor torchvision dependency." - }, - { - "left": "issue:44077", - "right": "issue:44625", - "accept": false, - "reason": "Optional post_init policy for patchtsmixer is unrelated to Qwen3.5 num_labels propagation." + "reason": "Loading a specific model and PIL image processors requiring torchvision are separate issues." }, { "left": "issue:43994", "right": "issue:44671", "accept": false, - "reason": "Different model-level prediction regressions: SigLIP2 pipeline/AutoModel nonsense vs CamemBERT masked LM outputs." + "reason": "Different models and different regressions: SigLIP2 output quality vs CamemBERT masked-LM predictions." }, { - "left": "issue:43421", - "right": "issue:44186", + "left": "issue:39692", + "right": "issue:44610", "accept": false, - "reason": "Tokenizer special-token post-processor syncing is unrelated to LayoutLMv2Tokenizer NER/padding crashes." + "reason": "Both are vision-model processor mismatches, but one is a docs/quantization example and the other is OmDet-Turbo size mismatch." }, { - "left": "issue:39692", - "right": "issue:44610", + "left": "issue:43421", + "right": "issue:44186", "accept": false, - "reason": "Both mention processor/example issues, but they affect different models and different mismatches." + "reason": "TokenizerBackend post-processor updates and LayoutLMv2 NER/padding crashes are separate tokenizer bugs." }, { "left": "issue:43596", "right": "issue:44220", "accept": false, - "reason": "Deepspeed ZeRO-3/BertModel index error is unrelated to audio fbank feature extraction." + "reason": "Deepspeed/BERT initialization and feature-extraction helper failure are unrelated." }, { "left": "issue:43531", "right": "issue:43645", "accept": false, - "reason": "Sliding-window Qwen3-MoE issue and Jupyter custom-model initialization are unrelated." + "reason": "Qwen3-MoE sliding_window bug does not match the Jupyter custom-model initialization regression." }, { "left": "issue:45081", "right": "issue:45356", "accept": false, - "reason": "Both touch _patch_mistral_regex, but one is a backend_tokenizer AttributeError and the other is a tokenizer regression with codec handling/warning behavior; not clearly the same fix." + "reason": "Both involve Mistral regex/tokenizer handling, but one crashes on backend_tokenizer absence while the other is a Kimi-K2.5 codec regression." }, { - "left": "issue:45183", - "right": "issue:45479", + "left": "issue:43746", + "right": "issue:45381", "accept": false, - "reason": "Validation-error messaging for input_ids/inputs_embeds is unrelated to zero-loss classification with num_labels=1." + "reason": "GraniteSpeech PEFT checkpoint loading and Qwen2.5-VL video position IDs are unrelated." }, { - "left": "issue:43540", - "right": "issue:44008", + "left": "issue:45183", + "right": "issue:45479", "accept": false, - "reason": "Different multimodal model bugs: Qwen3OmniMoe video processing vs Gemma3n variable-name collision." + "reason": "Both are validation/labeling bugs, but one is XOR input validation messaging and the other is degenerate loss for num_labels=1." }, { - "left": "issue:43746", - "right": "issue:45381", + "left": "issue:39692", + "right": "issue:43873", "accept": false, - "reason": "PEFT local checkpoint loading and Qwen2.5-VL video position ids are unrelated." + "reason": "SigLIP2 docs/quantization issues and quantized offloading behavior are not the same bug." }, { "left": "issue:44220", "right": "issue:44661", "accept": false, - "reason": "Audio fbank extraction and add-new-model-like tokenizer mapping failure are unrelated." + "reason": "Audio feature extraction and add-new-model-like tokenizer mapping failure are unrelated." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet, and the lone accepted pair (42886/44843) is well supported by the titles as the same HF_HUB_OFFLINE tokenizer-loading regression. The remaining verdicts are conservative: they reject pairs that share only broad subsystem similarity without clear evidence of the same bug. No overstatement is apparent." - }, - "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 }, @@ -3979,11 +3964,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:30064", "issue:36010", - "issue:39692", "issue:40444", "issue:43335", "issue:43421", @@ -4006,6 +3990,7 @@ "issue:43818", "issue:43819", "issue:43825", + "issue:43827", "issue:43873", "issue:43927", "issue:43994", @@ -4017,251 +4002,258 @@ "issue:44488", "issue:44561", "issue:44671", - "issue:44805", "issue:44811", + "issue:44821", "issue:44843", "issue:44877", "issue:44898", - "issue:44936", "issue:44977", "issue:45020", "issue:45072", "issue:45081", + "issue:45084", "issue:45381" ], "soft_pairs": [ - "issue:39692|issue:43873", + "issue:43540|issue:44008", "issue:43746|issue:43873", - "issue:40444|issue:43746", - "issue:43716|issue:44008", "issue:43618|issue:43697", + "issue:40444|issue:43746", "issue:44843|issue:45081", + "issue:43716|issue:44008", "issue:43720|issue:43819", - "issue:43335|issue:43756", "issue:43653|issue:43720", + "issue:43335|issue:43756", "issue:44062|issue:44488", "issue:30064|issue:36010", - "issue:43421|issue:44811", "issue:44186|issue:44898", - "issue:43540|issue:45381", + "issue:43421|issue:44811", "issue:43550|issue:45072", + "issue:43540|issue:45381", "issue:43577|issue:43818", "issue:43550|issue:43927", "issue:43643|issue:44877", - "issue:43782|issue:44977", "issue:43577|issue:43582", - "issue:44561|issue:45020", - "issue:44448|issue:44671", "issue:43531|issue:43643", + "issue:43782|issue:44977", + "issue:44448|issue:44671", "issue:43644|issue:43749", + "issue:44561|issue:45020", "issue:43653|issue:43819", "issue:43825|issue:43994", - "issue:44805|issue:44936", + "issue:44821|issue:45084", "issue:43550|issue:43819", - "issue:44186|issue:44442" + "issue:44186|issue:44442", + "issue:43825|issue:43827" ], - "prepared_review_unit_hash": "c0b0dd36863765c902c3ddf57dec0d4a437bed337826718e0f1632df65987523", + "prepared_review_unit_hash": "6e5f4e11276b36489bdc69578675c01afec2389cc2263536b3e02e692685f432", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 36071, - "estimated_input_tokens": 9018, - "estimated_eval_tokens": 18292 + "soft_pair_count": 30, + "serialized_chars": 36284, + "estimated_input_tokens": 9071, + "estimated_eval_tokens": 18398 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly heterogeneous: the soft-edge pairs overlap lexically but describe different models, subsystems, or failure modes. I did not find a convincing duplicate pair, and there are no PRs in the cluster.", - "confidence": 0.31, - "canonical_issue_reason": "No single true canonical emerges from these unrelated reports; if one issue must anchor the cluster, issue 45020 is the broadest umbrella ('recent transformers versions break models using remote_code') and is the least-specific representative.", + "summary": "The soft-edge candidates are mostly near-miss similarities rather than true duplicates: they share broad subsystems (tokenizers, multimodal models, loading, v5 regressions) but not the same concrete bug or fix path.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue cleanly covers this cluster; the items are heterogeneous. If forced to pick a representative, issue:45020 is the broadest umbrella, but it only reflects the remote_code regression subset.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the best global representative because it is the most general, open umbrella for a family of recent-version regressions affecting remote_code models.", + "best_issue_reason": "issue:45020 is the best umbrella-style representative because it summarizes the broadest recent-version regression theme, but it still does not subsume the unrelated tokenizer, multimodal, and loading bugs in the set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:39692", - "right": "issue:43873", + "left": "issue:43540", + "right": "issue:44008", "accept": false, - "reason": "SigLIP2 documentation example errors vs quantization/offloading behavior; same library, different bug." + "reason": "Different models and failures: Qwen3OmniMoe video-input processing vs Gemma 3n forward variable collision." }, { "left": "issue:43746", "right": "issue:43873", "accept": false, - "reason": "GraniteSpeech local-checkpoint loading with PEFT adapters is unrelated to offloading with quantization." + "reason": "Both are loading-related, but one is PEFT/local checkpoint loading and the other is quantization offloading behavior; not the same bug." }, { - "left": "issue:40444", - "right": "issue:43746", + "left": "issue:43618", + "right": "issue:43697", "accept": false, - "reason": "Qwen2.5-VL iterable-dataset multi-image finetuning failure is a different code path than GraniteSpeech checkpoint loading." + "reason": "CLIP attentions field regression and RTDetr output drift are unrelated code paths and fixes." }, { - "left": "issue:43716", - "right": "issue:44008", + "left": "issue:40444", + "right": "issue:43746", "accept": false, - "reason": "Mistral image preprocessor dtype mismatch and Gemma3n audio-mask variable collision are unrelated model-specific bugs." + "reason": "IterableDataset multi-image finetuning failure is unrelated to GraniteSpeech PEFT checkpoint loading." }, { - "left": "issue:43618", - "right": "issue:43697", + "left": "issue:44843", + "right": "issue:45081", "accept": false, - "reason": "CLIP attentions field regression and RTDetrV2 output drift are separate model behaviors." + "reason": "Both touch _patch_mistral_regex, but one is an offline model_info call and the other is a backend_tokenizer attribute crash; different concrete bugs." }, { - "left": "issue:44843", - "right": "issue:45081", + "left": "issue:43716", + "right": "issue:44008", "accept": false, - "reason": "Both touch _patch_mistral_regex, but one is offline model_info access and the other is a backend_tokenizer AttributeError; different failures." + "reason": "Blip/Mistral dtype mismatch and Gemma3n variable collision are unrelated." }, { "left": "issue:43720", "right": "issue:43819", "accept": false, - "reason": "BitNet packed-weight loading in accelerate is unrelated to DAC.from_latents vs forward-pass mismatch." + "reason": "BitNet packed-weight loading and DAC from_latents/STE mismatch are different model-specific issues." }, { - "left": "issue:43335", - "right": "issue:43756", + "left": "issue:43653", + "right": "issue:43720", "accept": false, - "reason": "SwitchTransformers sparse-layer creation and Smollm3 RoPE-layer dropping are different configuration/model-construction bugs." + "reason": "BigBirdTokenizer special-token decode issue is unrelated to BitNet accelerate loading." }, { - "left": "issue:43653", - "right": "issue:43720", + "left": "issue:43335", + "right": "issue:43756", "accept": false, - "reason": "BigBirdTokenizer special-token registration and BitNet packed-weight loading do not share the same bug." + "reason": "Different models and mechanisms: SwitchTransformers sparse-layer config vs Smollm3 RoPE-layer dropping." }, { "left": "issue:44062", "right": "issue:44488", "accept": false, - "reason": "AddedToken keyword collision and failure to load cjvt/sleng-bert are unrelated tokenizer issues." + "reason": "Tokenizer AddedToken special-arg duplication is not the same as failing to load cjvt/sleng-bert." }, { "left": "issue:30064", "right": "issue:36010", "accept": false, - "reason": "Void segmentation map processing and GenerationMixin import failure are completely different subsystems." - }, - { - "left": "issue:43421", - "right": "issue:44811", - "accept": false, - "reason": "Runtime post-processor updates for special tokens and Whisper batch_decode skip_special_tokens behavior are distinct tokenizer/processor bugs." + "reason": "ImageProcessor segmentation-map handling and GenerationMixin import failure are unrelated." }, { "left": "issue:44186", "right": "issue:44898", "accept": false, - "reason": "LayoutLMv2 NER/batched padding crashes and Perceiver non-default-resolution failures are unrelated." + "reason": "LayoutLMv2 tokenizer padding/NER crash and Perceiver interpolation failure are different bugs." }, { - "left": "issue:43540", - "right": "issue:45381", + "left": "issue:43421", + "right": "issue:44811", "accept": false, - "reason": "Both are video-related, but Qwen3OmniMoe video-input processing and Qwen2.5-VL vision_position_ids are different model families and bugs." + "reason": "Runtime special-token post-processor update and Whisper batch_decode skip_special_tokens are separate tokenizer/processor behaviors." }, { "left": "issue:43550", "right": "issue:45072", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba and dtype mismatches in SwitchTransformers/TimmWrapperModel do not look like the same underlying bug." + "reason": "Different failures in different models: torch.compile+SDPA vs bfloat16 dtype mismatch." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both involve video inputs, but the models and reported failures differ (Qwen3OmniMoe processing vs Qwen2.5-VL vision_position_ids)." }, { "left": "issue:43577", "right": "issue:43818", "accept": false, - "reason": "BLIP2 dtype loading issue and Video-LLaVA missing temporal attention/shared weights are unrelated." + "reason": "BLIP2 dtype retention and Video-LLaVA temporal-attention/weight-sharing issues are unrelated." }, { "left": "issue:43550", "right": "issue:43927", "accept": false, - "reason": "Bamba compile/SDPA regression and DiaConfig custom-token-ID loss are different failures." + "reason": "Bamba torch.compile/SDPA failure is unrelated to DiaConfig custom token ID loss on save/load." }, { "left": "issue:43643", "right": "issue:44877", "accept": false, - "reason": "trust_remote_code missing fields and strict granite_speech config loading are both config-related, but not the same bug." + "reason": "AutoConfig missing fields with trust_remote_code and strict granite_speech config loading are different config-path issues." }, { - "left": "issue:43782", - "right": "issue:44977", + "left": "issue:43577", + "right": "issue:43582", "accept": false, - "reason": "Qwen3VL weight_only loading error and Qwen3.5 flash-attention generation issue are different code paths." + "reason": "Model dtype retention and Apple Silicon caching_allocator_warmup TypeError are unrelated." }, { - "left": "issue:43577", - "right": "issue:43582", + "left": "issue:43531", + "right": "issue:43643", "accept": false, - "reason": "BLIP2 dtype issue and Apple Silicon caching_allocator_warmup TypeError are unrelated." + "reason": "Qwen3-MoE sliding_window behavior and trust_remote_code AutoConfig fields are unrelated." }, { - "left": "issue:44561", - "right": "issue:45020", + "left": "issue:43782", + "right": "issue:44977", "accept": false, - "reason": "Issue 45020 is a broad umbrella about remote_code regressions, while 44561 pinpoints one specific cause; related, but not the same concrete bug." + "reason": "Qwen3VL weight_only load error and Qwen3.5 flash-attention generation issue are different failures." }, { "left": "issue:44448", "right": "issue:44671", "accept": false, - "reason": "Pegasus v4/v5 output differences and CamemBERT masked-LM prediction errors are separate model regressions." + "reason": "Both are v4/v5 output regressions, but they concern different models and tasks." }, { - "left": "issue:43531", - "right": "issue:43643", + "left": "issue:43644", + "right": "issue:43749", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior and trust_remote_code config loading are unrelated." + "reason": "Non-persistent buffer initialization junk and FSDP CPU RAM efficient loading are separate loading/state issues." }, { - "left": "issue:43644", - "right": "issue:43749", + "left": "issue:44561", + "right": "issue:45020", "accept": false, - "reason": "Non-persistent-buffer junk in Transformers 5.0 and FSDP CPU RAM efficient loading are different loading/runtime problems." + "reason": "45020 is a broad umbrella around remote_code regressions; 44561 is one specific trust_remote_code breakage from a removed helper, not the same issue." }, { "left": "issue:43653", "right": "issue:43819", "accept": false, - "reason": "Tokenizer special-token registration and DAC latent reconstruction mismatch are unrelated." + "reason": "BigBirdTokenizer special-token decode and DAC latent reconstruction mismatch are unrelated." }, { "left": "issue:43825", "right": "issue:43994", "accept": false, - "reason": "Pipeline translation-task messaging and siglip2 AutoModel/pipeline output issues are different pipeline/model bugs." + "reason": "Pipeline translation error messaging and SigLIP2 nonsensical outputs are unrelated." }, { - "left": "issue:44805", - "right": "issue:44936", + "left": "issue:44821", + "right": "issue:45084", "accept": false, - "reason": "Mask-shape IndexError and trainer.evaluate after trainer.train failure are not the same underlying issue." + "reason": "AutoImageProcessor URL loading and template-node compilation errors are unrelated." }, { "left": "issue:43550", "right": "issue:43819", "accept": false, - "reason": "Bamba SDPA/torch.compile regression and DAC.from_latents mismatch are unrelated." + "reason": "Different model families and different code paths: SDPA compile failure vs DAC latent decoding mismatch." }, { "left": "issue:44186", "right": "issue:44442", "accept": false, - "reason": "LayoutLMv2 tokenizer crash and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer-loading paths." + "reason": "LayoutLMv2 tokenizer crash and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer issues." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Both are about pipeline deprecation fallout, but one is a bad runtime error message and the other is outdated docs; separate fixes." } ] }, @@ -4283,7 +4275,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:30064", "issue:34634", @@ -4309,8 +4301,6 @@ "issue:43749", "issue:43782", "issue:43819", - "issue:43825", - "issue:43827", "issue:43906", "issue:44060", "issue:44291", @@ -4322,6 +4312,7 @@ "issue:44589", "issue:44610", "issue:44671", + "issue:44805", "issue:44811", "issue:44857", "issue:44912", @@ -4335,231 +4326,219 @@ "issue:45488" ], "soft_pairs": [ - "issue:41720|issue:44991", - "issue:43825|issue:43827", - "issue:42175|issue:45081", "issue:43650|issue:43906", - "issue:30064|issue:36331", + "issue:41720|issue:44991", + "issue:44805|issue:44936", "issue:45081|issue:45406", + "issue:30064|issue:36331", + "issue:42175|issue:45081", "issue:43606|issue:45072", - "issue:43782|issue:45325", "issue:44610|issue:45072", + "issue:43782|issue:45325", "issue:44485|issue:44509", "issue:34634|issue:36010", - "issue:43720|issue:45072", - "issue:43653|issue:44671", - "issue:43582|issue:44857", "issue:44060|issue:44368", + "issue:43720|issue:45072", "issue:43262|issue:43819", + "issue:43582|issue:44857", "issue:43749|issue:44291", + "issue:43653|issue:44671", "issue:43526|issue:43697", "issue:43526|issue:44554", - "issue:43638|issue:44936", "issue:43653|issue:44912", - "issue:44568|issue:45488", "issue:43454|issue:43577", + "issue:44568|issue:45488", "issue:43452|issue:44991", "issue:34634|issue:36331", + "issue:43638|issue:44936", + "issue:43257|issue:43329", "issue:43746|issue:44811", - "issue:44589|issue:45042", - "issue:43257|issue:43329" + "issue:44589|issue:45042" ], - "prepared_review_unit_hash": "b1803874004fa8afc9130b9347bf688bb3c0710b504f4c71f56bcf6ef4ed0a4f", + "prepared_review_unit_hash": "185078575abf880b9b189dca2165dfef5743d6fad9c4ce2f230db3e63c3d21ca", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 28, - "serialized_chars": 36638, - "estimated_input_tokens": 9160, - "estimated_eval_tokens": 18576 + "serialized_chars": 36088, + "estimated_input_tokens": 9022, + "estimated_eval_tokens": 18300 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No PRs are present. The soft pairs are overwhelmingly false positives: several are in the same broad area (tokenizers/loading/docs), but they describe different bugs, code paths, or model families rather than the same underlying change.", - "confidence": 0.38, - "canonical_issue_reason": "issue:43749 is the clearest canonical issue candidate: it is a well-scoped regression in a named loading feature, with substantial discussion and inbound references, and it is concrete enough to anchor triage.", + "summary": "These are mostly unrelated bug reports spanning tokenizers, model loading, processors, training, and backend-specific failures. None of the soft pairs look like the same underlying bug/change closely enough to merge, so all soft edges should be rejected. The clearest standalone issue is #45081, which has a concrete traceback and a specific tokenizer code path.", + "confidence": 0.91, + "canonical_issue_reason": "#45081 is the most canonical issue in this set: it is specific, has an explicit traceback, names the exact failing function and flag, and is scoped to a concrete tokenizer loading path.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43749 is the strongest overall issue candidate for canonicalization because it is specific, actionable, and appears to be the most central bug report among the set.", + "best_issue_reason": "#45081 is the strongest issue candidate because it is actionable, well-scoped, and tied to a reproducible AttributeError in a single code path.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:41720", - "right": "issue:44991", + "left": "issue:43650", + "right": "issue:43906", "accept": false, - "reason": "Different models and failures: Qwen3 device-mapping cuda assert vs tokenizer loading for EMBEDDIA/est-roberta." + "reason": "Different problems: a vague content issue vs an isolated reproduction of another bug; no shared code path or fix." }, { - "left": "issue:43825", - "right": "issue:43827", + "left": "issue:41720", + "right": "issue:44991", "accept": false, - "reason": "Related v5 pipeline-removal theme, but one is an error-message bug and the other is docs cleanup; not the same fix." + "reason": "One is CUDA assert during auto device mapping for Qwen3, the other is tokenizer loading for EMBEDDIA/est-roberta; unrelated failures." }, { - "left": "issue:42175", - "right": "issue:45081", + "left": "issue:44805", + "right": "issue:44936", "accept": false, - "reason": "Unrelated subsystems: missing TensorFlow backend dependency vs Mistral tokenizer regex patch crash." + "reason": "Mask/tensor shape mismatch in one report versus trainer evaluate-after-train behavior in the other; not the same bug." }, { - "left": "issue:43650", - "right": "issue:43906", + "left": "issue:45081", + "right": "issue:45406", "accept": false, - "reason": "Generic placeholder issue vs isolated reproduction of another upstream bug; no shared underlying change." + "reason": "Both are processor/tokenizer errors, but they fail in different classes and code paths: Mistral regex patching vs Gemma4Processor missing _tokenizer." }, { "left": "issue:30064", "right": "issue:36331", "accept": false, - "reason": "Image segmentation processing failure and CustomTrainer keyword-arg regression are unrelated." + "reason": "Image segmentation map processing and CustomTrainer.compute_loss signature mismatch are unrelated." }, { - "left": "issue:45081", - "right": "issue:45406", + "left": "issue:42175", + "right": "issue:45081", "accept": false, - "reason": "Tokenizer patch crash and Gemma4Processor serve crash are different code paths and model families." + "reason": "Package/backend dependency issue versus a tokenizer patch crash; no common concrete bug." }, { "left": "issue:43606", "right": "issue:45072", "accept": false, - "reason": "Both are inference/loading problems, but the model types and root causes differ (CPU offload device mismatch vs dtype mismatch)." + "reason": "CPU offload device mismatch for Bark-small is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." }, { - "left": "issue:43782", - "right": "issue:45325", + "left": "issue:44610", + "right": "issue:45072", "accept": false, - "reason": "Different VL models and distinct bugs: weight_only loading error vs rope index scaling issue." + "reason": "Different domains and failures: OmDet-Turbo image size mismatch versus dtype mismatch in inference." }, { - "left": "issue:44610", - "right": "issue:45072", + "left": "issue:43782", + "right": "issue:45325", "accept": false, - "reason": "Processor image-size mismatch is unrelated to bfloat16 dtype mismatches." + "reason": "Both involve Qwen/VL models, but one is a from_pretrained weight_only error and the other is a RoPE position-id scaling bug." }, { "left": "issue:44485", "right": "issue:44509", "accept": false, - "reason": "RoPE implementation discussion and v5 pipeline-doc cleanup are not the same underlying change." + "reason": "GLM-5 RoPE implementation bug and docs mentioning removed pipeline tasks are separate issues." }, { "left": "issue:34634", "right": "issue:36010", "accept": false, - "reason": "Bark voice preset behavior is unrelated to a GenerationMixin import error." + "reason": "BarkProcessor voice_preset bug is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings, but they concern different models and different symptoms; not a single concrete bug." }, { "left": "issue:43720", "right": "issue:45072", "accept": false, - "reason": "Packed-weight loading bug and dtype mismatch bug are different loading paths and fixes." + "reason": "BitNet accelerate-loading unpacking bug is unrelated to dtype mismatches in other models." }, { - "left": "issue:43653", - "right": "issue:44671", + "left": "issue:43262", + "right": "issue:43819", "accept": false, - "reason": "Both involve tokenizers/model outputs, but BigBird special-token registration and CamemBERT MLM prediction drift are distinct bugs." + "reason": "Audio chat-template sampling-rate defaulting and DAC.from_latents vs forward mismatch are separate code paths." }, { "left": "issue:43582", "right": "issue:44857", "accept": false, - "reason": "Apple Silicon allocator TypeError is unrelated to CUDA AMP loss crashes." - }, - { - "left": "issue:44060", - "right": "issue:44368", - "accept": false, - "reason": "Both mention tied-weights warnings, but the concrete bug is different: incorrect binding to dt_bias vs a generic config warning." + "reason": "Apple Silicon TypeError in caching allocator warmup is unrelated to CUDA AMP crashes in LwDetrImageLoss." }, { - "left": "issue:43262", - "right": "issue:43819", + "left": "issue:43749", + "right": "issue:44291", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and DAC latent/STE mismatch are unrelated." + "reason": "Both touch loading, but one is FSDP CPU RAM efficient loading and the other is init_empty_weights passing an unexpected argument; not the same bug." }, { - "left": "issue:43749", - "right": "issue:44291", + "left": "issue:43653", + "right": "issue:44671", "accept": false, - "reason": "Both are loading-related, but FSDP RAM-efficient loading and init_empty_weights/_is_hf_initialized are different code paths." + "reason": "Tokenizer special-token registration and masked LM prediction drift are different issues." }, { "left": "issue:43526", "right": "issue:43697", "accept": false, - "reason": "BEiT reduce_labels bug and RTDetrV2 output differences are different models and failure modes." + "reason": "BeitImageProcessorFast reduce_labels bug does not match RTDetrV2 output differences in v5." }, { "left": "issue:43526", "right": "issue:44554", "accept": false, - "reason": "Image-label reduction bug is unrelated to the MPS attention correctness issue." - }, - { - "left": "issue:43638", - "right": "issue:44936", - "accept": false, - "reason": "Zero3/non-pretrained Bert index error and trainer evaluate-after-train failure are distinct trainer/runtime problems." + "reason": "Image processor label reduction and an MPS attention correctness issue are unrelated." }, { "left": "issue:43653", "right": "issue:44912", "accept": false, - "reason": "BigBird tokenizer special-token bug and MXFP4 quantization fallback are unrelated." + "reason": "BigBirdTokenizer decode bug and MXFP4 quantization loading fallback are different code paths." }, { - "left": "issue:44568", - "right": "issue:45488", + "left": "issue:43454", + "right": "issue:43577", "accept": false, - "reason": "Both are tokenizer regressions, but one is about add_special_tokens not inserting BOS/EOS and the other is LlamaTokenizer overriding pre-tokenizer behavior." + "reason": "AyaVision lm_head tying and Blip2 dtype propagation are separate model-loading problems." }, { - "left": "issue:43454", - "right": "issue:43577", + "left": "issue:44568", + "right": "issue:45488", "accept": false, - "reason": "AyaVision weight tying and Blip2 dtype propagation are different model-loading issues." + "reason": "Tokenizer special-token handling for mdeberta-v3-base and LlamaTokenizer pre-tokenizer override for DeepSeek are not the same fix." }, { "left": "issue:43452", "right": "issue:44991", "accept": false, - "reason": "Both concern tokenizer/model loading, but the gguf_file breakage and est-roberta tokenizer failure are not the same bug." + "reason": "gguf_file loading breakage and tokenizer loading failure for a specific model are unrelated." }, { - "left": "issue:34634", - "right": "issue:36331", + "left": "issue:43638", + "right": "issue:44936", "accept": false, - "reason": "Bark voice_preset and CustomTrainer.compute_loss kwarg mismatch are unrelated." + "reason": "DeepSpeed ZeRO-3 + non-pretrained Bert index error is not the same as evaluate-after-train behavior." }, { "left": "issue:43746", "right": "issue:44811", "accept": false, - "reason": "GraniteSpeech local-checkpoint loading with PEFT adapters is unrelated to Whisper batch_decode skip_special_tokens handling." + "reason": "PEFT local checkpoint loading and Whisper batch_decode skip_special_tokens handling are unrelated." }, { "left": "issue:44589", "right": "issue:45042", "accept": false, - "reason": "Float8 storage lookup failure and PIL backend requiring torchvision are unrelated issues." - }, - { - "left": "issue:43257", - "right": "issue:43329", - "accept": false, - "reason": "Qwen3 MOE weight conversion/loading and multimodal video token counting are different bugs with no shared code path." + "reason": "Float8 storage lookup failure is unrelated to PIL backend processors requiring torchvision." } ] }, @@ -4581,8 +4560,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ + "issue:33357", + "issue:34567", "issue:36010", "issue:36331", "issue:38175", @@ -4608,14 +4589,12 @@ "issue:43749", "issue:43756", "issue:43819", - "issue:43824", "issue:43854", "issue:43927", "issue:43931", "issue:44062", "issue:44291", "issue:44479", - "issue:44488", "issue:44560", "issue:44568", "issue:44610", @@ -4635,263 +4614,256 @@ "soft_pairs": [ "issue:43854|issue:44977", "issue:38175|issue:44610", + "issue:39692|issue:44912", "issue:36010|issue:43645", "issue:43257|issue:43696", - "issue:44568|issue:45356", - "issue:39692|issue:44912", "issue:43819|issue:43927", - "issue:42175|issue:44938", + "issue:43262|issue:44811", "issue:43720|issue:44610", "issue:43550|issue:44912", - "issue:43262|issue:44811", + "issue:44568|issue:45356", + "issue:42175|issue:44938", "issue:43749|issue:44756", "issue:43531|issue:45042", "issue:43329|issue:44479", "issue:36010|issue:36331", - "issue:43257|issue:43756", "issue:43257|issue:43931", - "issue:43577|issue:45072", - "issue:41720|issue:43540", - "issue:43452|issue:43582", + "issue:43257|issue:43756", "issue:43329|issue:43450", + "issue:43577|issue:45072", "issue:43697|issue:45357", - "issue:44291|issue:45042", + "issue:43452|issue:43582", + "issue:41720|issue:43540", "issue:44479|issue:45381", - "issue:42175|issue:43824", + "issue:44291|issue:45042", "issue:44062|issue:44991", - "issue:42915|issue:44821", "issue:43540|issue:44560", - "issue:42175|issue:44488", "issue:42915|issue:43742", "issue:43927|issue:44912", - "issue:43540|issue:44479" + "issue:43540|issue:44479", + "issue:42915|issue:44821", + "issue:33357|issue:34567" ], - "prepared_review_unit_hash": "8a690a773844957a5beb2b45c5f662963ad77d2f570ae46443df3dbf7167760d", + "prepared_review_unit_hash": "b2d1458166ec460135b1a93115be94b3cf7ef1da0944bc385202760abd6d2bbc", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37405, - "estimated_input_tokens": 9352, - "estimated_eval_tokens": 18960 + "soft_pair_count": 31, + "serialized_chars": 37099, + "estimated_input_tokens": 9275, + "estimated_eval_tokens": 18806 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly unrelated issues; the only strong duplicate is the Qwen2.5-VL video-input regression pair. 44479 is the broader umbrella report, while 45381 looks like a narrower manifestation of the same bug.", - "confidence": 0.78, - "canonical_issue_reason": "issue:44479 is the broadest, clearest umbrella report for the video-input regression across multiple Qwen variants; it best represents the underlying bug cluster.", + "summary": "Most candidate pairs are unrelated and should stay separate. The only clear overlaps are the Qwen video-input regressions around v5.3.0, which appear to share the same underlying multimodal/video code-path issue.", + "confidence": 0.76, + "canonical_issue_reason": "issue:44479 is the broadest and clearest report of the shared Qwen video-input regression, covering multiple affected Qwen VL variants and matching the other video-related soft edges.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44479 is the best cluster anchor because it describes the shared regression at a higher level and covers the model family overlap seen in 45381.", + "best_issue_reason": "issue:44479 is the best canonical issue because it describes the user-facing regression most generally and is broad enough to subsume the narrower video-position/tokenization reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43854", "right": "issue:44977", "accept": false, - "reason": "Different models and failures: GLM-4.7-Flash loading/test issue vs Qwen3.5 flash-attention generation regression." + "reason": "Different models and symptoms; CLIP load crash vs Qwen3.5 flash-attention generation bug." }, { "left": "issue:38175", "right": "issue:44610", "accept": false, - "reason": "SigLIP2 zero-probability behavior is unrelated to OmDet-Turbo processor input-size mismatch." + "reason": "Unrelated subsystems: SigLIP2 probability issue vs OmDet-Turbo image-size mismatch." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Documentation/example errors vs MXFP4 quantization fallback; not the same bug." }, { "left": "issue:36010", "right": "issue:43645", "accept": false, - "reason": "Both mention import/runtime breakage, but one is a GenerationMixin import error and the other is a custom-model/Jupyter regression in v5.0." + "reason": "Both are version-5 breakages, but one is a GenerationMixin import error and the other is custom-model notebook init behavior." }, { "left": "issue:43257", "right": "issue:43696", "accept": false, - "reason": "Different concrete problems: Qwen3 MoE weight conversion with accelerate/deepspeed vs GPT-oss-20b OOM." - }, - { - "left": "issue:44568", - "right": "issue:45356", - "accept": false, - "reason": "Both are tokenizer bugs, but they affect different models and different failure modes (special-token insertion vs codec/warning regression)." - }, - { - "left": "issue:39692", - "right": "issue:44912", - "accept": false, - "reason": "Docs-example errors for SigLIP2 are not the same bug as git-oss-20b quantization fallback." + "reason": "Qwen3 MOE weight conversion/loading bug vs GPT-oss OOM; different failure modes." }, { "left": "issue:43819", "right": "issue:43927", "accept": false, - "reason": "Different subsystems: DAC forward/from_latents mismatch vs DiaConfig save/load token-id loss." + "reason": "Different components entirely: DAC latent conversion vs DiaConfig token-id persistence." }, { - "left": "issue:42175", - "right": "issue:44938", + "left": "issue:43262", + "right": "issue:44811", "accept": false, - "reason": "Packaging/backend dependency issue is unrelated to Python 3.14 import/load failure." + "reason": "Audio chat-template sample-rate default vs Whisper batch_decode skip_special_tokens handling." }, { "left": "issue:43720", "right": "issue:44610", "accept": false, - "reason": "Packed-weight unpacking during accelerate loading is unrelated to OmDet-Turbo image-size mismatch." + "reason": "Accelerate packed-weight loading bug vs OmDet processor input-size mismatch." }, { "left": "issue:43550", "right": "issue:44912", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba is unrelated to quantization fallback in git-oss-20b." + "reason": "torch.compile/SDPA runtime bug vs quantization fallback; different code paths." }, { - "left": "issue:43262", - "right": "issue:44811", + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer-related, but the titles indicate different models and different tokenizer regressions." + }, + { + "left": "issue:42175", + "right": "issue:44938", "accept": false, - "reason": "Both are audio-related, but one is apply_chat_template sampling-rate defaulting and the other is batch_decode skip_special_tokens handling." + "reason": "Packaging/backend dependency issue vs Python 3.14 import failure." }, { "left": "issue:43749", "right": "issue:44756", "accept": false, - "reason": "FSDP CPU RAM efficient loading and Strix Halo mmap OOM are different loading/memory issues with no clear shared code path." + "reason": "Both involve loading/memory, but one is FSDP CPU RAM efficient loading and the other is mmap/OOM on Strix Halo." }, { "left": "issue:43531", "right": "issue:45042", "accept": false, - "reason": "Qwen3-MoE sliding_window bug is unrelated to PIL backend requiring torchvision." + "reason": "Qwen3-MoE sliding window bug vs PIL backend torchvision dependency regression." }, { "left": "issue:43329", "right": "issue:44479", - "accept": false, - "reason": "Both involve video/multimodal code, but one is an undefined-function bug in token counting and the other is a video-input regression/positioning issue." + "accept": true, + "reason": "Both point to the same Qwen video-input/multimodal-token regression in the video code path; one is the specific undefined-symbol bug, the other the broader downstream regression report." }, { "left": "issue:36010", "right": "issue:36331", "accept": false, - "reason": "Different API breakages: GenerationMixin import path vs CustomTrainer.compute_loss signature change." + "reason": "Both are API breakages, but one is GenerationMixin import and the other is compute_loss signature mismatch." }, { "left": "issue:43257", - "right": "issue:43756", + "right": "issue:43931", "accept": false, - "reason": "Qwen3 MoE weight conversion and Smollm3 RoPE-layer dropping are unrelated model-specific bugs." + "reason": "Different Qwen3 model-loading failures with different symptoms and causes." }, { "left": "issue:43257", - "right": "issue:43931", + "right": "issue:43756", + "accept": false, + "reason": "Qwen3 MOE loading vs Smollm3 RoPE layer dropping; unrelated." + }, + { + "left": "issue:43329", + "right": "issue:43450", "accept": false, - "reason": "Accelerate/deepspeed MoE loading bug is not the same as Qwen3-VL weight-shape mismatch." + "reason": "Both are video-related, but one is token counting in multimodal utilities and the other is a batched video processor shape bug." }, { "left": "issue:43577", "right": "issue:45072", "accept": false, - "reason": "Both are dtype-related, but they affect different models and likely different loading/inference paths." + "reason": "Different dtype mismatch regressions in different model families." }, { - "left": "issue:41720", - "right": "issue:43540", + "left": "issue:43697", + "right": "issue:45357", "accept": false, - "reason": "Auto device mapping cuda assert on A800 is unrelated to Qwen3OmniMoe video-input processing." + "reason": "RTDetrV2 output regression vs Qwen3.5 save_pretrained visual-key regression; unrelated." }, { "left": "issue:43452", "right": "issue:43582", "accept": false, - "reason": "GGUF tokenizer/model loading failure is unrelated to Apple Silicon caching allocator warmup TypeError." + "reason": "Tokenizer/model gguf loading issue vs Apple Silicon allocator warmup TypeError." }, { - "left": "issue:43329", - "right": "issue:43450", + "left": "issue:41720", + "right": "issue:43540", "accept": false, - "reason": "Multimodal token counting bug is different from batched video processor output shape bug." + "reason": "Qwen3 auto device-mapping CUDA assert vs Qwen3OmniMoe video ValueError; different bugs." }, { - "left": "issue:43697", - "right": "issue:45357", - "accept": false, - "reason": "RTDetrV2 output regression is unrelated to Qwen3.5 save_pretrained visual-encoder key corruption." + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both are Qwen2.5-VL/Qwen video-input regressions; 45381 looks like a narrower manifestation of the same bad video position handling." }, { "left": "issue:44291", "right": "issue:45042", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized loading bug is unrelated to PIL backend torchvision dependency handling." - }, - { - "left": "issue:44479", - "right": "issue:45381", - "accept": true, - "reason": "Both describe the same Qwen2.5-VL video-input regression; 45381 is a narrower statement of the broader v5.3.0 issue in 44479." - }, - { - "left": "issue:42175", - "right": "issue:43824", - "accept": false, - "reason": "Package install backend selection is unrelated to the missing Qwen2_5_VLForConditionalGeneration import." + "reason": "init_empty_weights argument handling vs PIL/torchvision dependency issue." }, { "left": "issue:44062", "right": "issue:44991", "accept": false, - "reason": "AddedToken keyword conflict is unrelated to tokenizer loading failure for EMBEDDIA/est-roberta." - }, - { - "left": "issue:42915", - "right": "issue:44821", - "accept": false, - "reason": "Qwen3Moe FP8Config failure is unrelated to AutoImageProcessor loading from URL." + "reason": "Both are tokenizer loading problems, but the titles do not show the same concrete failure or model/config." }, { "left": "issue:43540", "right": "issue:44560", "accept": false, - "reason": "Different video bugs in different model families: Qwen3OmniMoe ValueError vs Qwen3-vl-embedding StopIteration." - }, - { - "left": "issue:42175", - "right": "issue:44488", - "accept": false, - "reason": "Backend installation issue is unrelated to sleng-bert tokenizer/model loading failure." + "reason": "Different Qwen model variants and different video failures; not clearly the same bug." }, { "left": "issue:42915", "right": "issue:43742", "accept": false, - "reason": "FineGrainedFP8Config failure for Qwen3Moe is unrelated to MobileLLM-125M key loading." + "reason": "FineGrainedFP8Config issue vs MobileLLM key error; unrelated." }, { "left": "issue:43927", "right": "issue:44912", "accept": false, - "reason": "DiaConfig custom token IDs/save-load bug is unrelated to git-oss-20b quantization fallback." + "reason": "Custom token ID persistence vs quantization fallback; unrelated." }, { "left": "issue:43540", "right": "issue:44479", "accept": false, - "reason": "Both involve video inputs, but the titles point to different models and different failure modes, so they are not the same concrete bug." + "reason": "Related by model family and video inputs, but the reported failures and model targets are different enough to keep separate." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Unrelated: FP8 Qwen3Moe loading vs AutoImageProcessor-from-URL loading." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error on CLIP vs TrainerState token counter not updating; unrelated." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: 44479 is the broader Qwen video-input regression umbrella, and 45381 is a narrower manifestation that matches it well. The other soft-edge verdicts stay conservative by distinguishing different models, subsystems, or failure modes rather than forcing duplicates." + "feedback": "The summary is grounded in the packet and stays conservative. The canonical issue choice is supported by the titles, and the only accepted soft-edge links are the clearly video-regression-related Qwen pairs, which look plausibly mergeable into a single bugfix. No obvious overstatement stands out." }, "evaluator_used": true, "retried": false, @@ -4910,11 +4882,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:28282", "issue:29127", - "issue:33357", "issue:34567", "issue:34689", "issue:36010", @@ -4935,6 +4906,7 @@ "issue:43749", "issue:43756", "issue:43782", + "issue:43824", "issue:43906", "issue:43994", "issue:44062", @@ -4947,6 +4919,7 @@ "issue:44704", "issue:44792", "issue:44821", + "issue:44843", "issue:44898", "issue:44964", "issue:45042", @@ -4961,80 +4934,262 @@ "issue:45479" ], "soft_pairs": [ - "issue:43526|issue:45479", - "issue:33357|issue:34567", - "issue:43906|issue:45381", + "issue:42175|issue:43824", "issue:45137|issue:45405", - "issue:44484|issue:44704", - "issue:36010|issue:42175", + "issue:42175|issue:44488", + "issue:43906|issue:45381", + "issue:43526|issue:45479", "issue:38175|issue:44898", "issue:43582|issue:43782", - "issue:43742|issue:44291", - "issue:43646|issue:43749", + "issue:44484|issue:44704", "issue:28282|issue:34689", + "issue:43646|issue:43749", + "issue:36010|issue:42175", + "issue:43742|issue:44291", + "issue:43493|issue:43994", "issue:43643|issue:45375", + "issue:28282|issue:29127", "issue:43756|issue:45325", - "issue:44291|issue:44821", - "issue:43493|issue:43994", "issue:44062|issue:45081", - "issue:28282|issue:29127", + "issue:44291|issue:44821", "issue:43452|issue:44488", "issue:34567|issue:36331", - "issue:43688|issue:43749", "issue:44792|issue:44964", - "issue:43742|issue:44589", - "issue:43531|issue:44821", + "issue:43688|issue:43749", "issue:44514|issue:45290", - "issue:45042|issue:45081", + "issue:43742|issue:44589", "issue:43452|issue:43577", "issue:43643|issue:44561", + "issue:45042|issue:45081", + "issue:43531|issue:44821", "issue:44514|issue:45381", - "issue:43697|issue:45216" + "issue:43697|issue:45216", + "issue:43452|issue:44843" ], - "prepared_review_unit_hash": "2edad70a2a723a7e8b14bf7753a611b105db6f977de7a861b48a3cbb11693df7", + "prepared_review_unit_hash": "6f74cd32eb2dd254ef3829a0a33f4c068071ea5c66d70804a2b23130d23aa0bb", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 29, - "serialized_chars": 36062, - "estimated_input_tokens": 9016, - "estimated_eval_tokens": 18288 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37287, + "estimated_input_tokens": 9322, + "estimated_eval_tokens": 18900 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The pair is related in the same chat-template area, but the packet does not support treating 44514 and 45290 as an exact duplicate. The safer triage is to keep them as closely related, not merged.", - "confidence": 0.68, - "canonical_issue_reason": "44514 is the cleaner canonical issue for the batched `apply_chat_template` crash because its title and context are more specific about the failure surface. 45290 looks adjacent but describes a different trigger (`tokenize=True` with assistant tool-call messages and no content).", + "summary": "This cluster remains heterogeneous overall: most items are unrelated loading, dependency, training, or model-specific bugs. I only accept one soft-edge duplicate pair, the two SigLIP2 reports, because they point to the same model-family output regression. The other soft edges are rejected as separate failure modes, with the Qwen2.5-VL and tool-call pairs kept conservative and grounded in the packet text.", + "confidence": 0.95, + "canonical_issue_reason": "No single issue cleanly represents the cluster; the items span many distinct code paths and bug classes, so any one issue would be a poor canonical for the whole set.", "canonical_pr_reason": null, - "best_issue_reason": "44514 is the best issue representative for this cluster because it is the more concrete and better-scoped chat-template regression, while 45290 appears to be a related but not identical crash.", + "best_issue_reason": "No global best issue: the only clear local duplicate is the SigLIP2 pair, but the cluster as a whole is too mixed for one issue to serve as a reliable representative.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Backend packaging/dependency problem vs missing Qwen2.5-VL symbol import; not the same bug." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque failure is unrelated to a PEFT version pin regression." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Torch backend packaging issue vs a specific model-loading failure; different failures." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "The packet only shows an isolated repro and a Qwen2.5-VL video position_ids bug; no grounded evidence they are the same defect." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BEiT fast processor label-reduction bug vs sequence-classification zero-loss bug; different code paths." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision-model related, but the model and symptom are different enough that they do not look like one bug." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError is unrelated to Qwen3VL weight_only loading." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard sizing vs AutoProcessor kwargs forwarding are separate issues." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Missing PyTorch ImportError is not the same as a Llama 3.2 Vision loading regression." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization breakage and FSDP CPU RAM-efficient loading are different code paths." + }, + { + "left": "issue:36010", + "right": "issue:42175", + "accept": false, + "reason": "GenerationMixin import failure and TensorFlow backend packaging are unrelated." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "MobileLLM key error and init_empty_weights/_is_hf_initialized mismatch are separate problems." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both reports describe the same SigLIP2 model producing incorrect or nonsensical outputs, indicating the same underlying bug." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both involve config fields, but they concern different models and different failure mechanisms." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "PyTorch-missing ImportError and LayoutLMv3 error-message clarity are not duplicates." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both touch RoPE/position handling, but they are different models and distinct bugs." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "AddedToken constructor conflict is unrelated to the Mistral regex crash." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights argument handling and AutoImageProcessor URL loading are separate failures." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf loading breakage and a specific model load failure for cjvt/sleng-bert are different issues." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token-counting update and CustomTrainer signature mismatch are different trainer bugs." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "A janus test failure is not the same as loading Phi-4 multimodal instruct." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Auxiliary-loss normalization and FSDP CPU RAM-efficient loading are unrelated." + }, { "left": "issue:44514", "right": "issue:45290", "accept": false, - "reason": "Same broad chat-template area, but the titles describe different failure modes and triggers: batched `padding=False` crash vs assistant tool-call/no-content crash. Not enough evidence for an exact duplicate." + "reason": "The packet describes batched-input/padding=False crashes on one side and tool-call assistant-message crashes on the other; not the same bug." + }, + { + "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:43577", + "accept": false, + "reason": "gguf loading breakage and Blip2 dtype propagation are separate issues." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both affect trust_remote_code models, but the concrete failure modes differ." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "PIL backend torchvision dependency issue is unrelated to the Mistral regex crash." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Batched-input/padding=False apply_chat_template crash and Qwen2.5-VL video position_ids bug are different problems." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output regression and Qwen3.5 checkpoint corruption are not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "gguf loading breakage and offline model_info calls in _patch_mistral_regex are different code paths." } ] }, "evaluator_result": { "accept": false, - "feedback": "The overall framing is mostly conservative, but the accepted pair is overstated. In the packet, issue 44514 is titled as a batched `apply_chat_template`/`padding=False` crash for Qwen2.5-VL, while 45290 is a `tokenize=True` crash on assistant messages with tool calls and no content. Those titles suggest a related chat-template area, but they do not establish an exact one-to-one duplicate or the same failure mode. Likewise, the canonicalization rationale leans too hard on evidence not present in the packet. I would keep this as a likely-related/soft match rather than an exact duplicate." + "feedback": "The overall heterogeneity summary is fine, but at least a couple of soft-edge reasons overreach the packet. In particular, issue:44514 is described in the packet as a batched-input/padding=False crash, not a tool-call crash, and issue:43906 relies on unknown external context (issue 38071) to justify the comparison to issue:45381. That makes the soft verdicts insufficiently grounded/conservative, so I can\u2019t accept this clustering." }, "evaluator_used": true, "retried": true, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -5049,7 +5204,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:24643", "issue:28282", @@ -5066,7 +5221,6 @@ "issue:43122", "issue:43295", "issue:43329", - "issue:43452", "issue:43525", "issue:43531", "issue:43577", @@ -5101,259 +5255,255 @@ "issue:45440" ], "soft_pairs": [ - "issue:41762|issue:44936", - "issue:43452|issue:44843", "issue:43688|issue:45440", - "issue:43653|issue:44568", - "issue:43720|issue:44898", - "issue:43819|issue:44811", "issue:44625|issue:44877", - "issue:43295|issue:45356", + "issue:43819|issue:44811", + "issue:43720|issue:44898", + "issue:41762|issue:44936", + "issue:43653|issue:44568", "issue:44464|issue:44743", - "issue:40444|issue:44811", "issue:44466|issue:45005", - "issue:41720|issue:43906", + "issue:43295|issue:45356", "issue:24643|issue:29127", + "issue:41720|issue:43906", "issue:43525|issue:43643", + "issue:40444|issue:44811", "issue:44843|issue:44991", "issue:44361|issue:44442", "issue:43577|issue:43697", "issue:38175|issue:43295", "issue:43653|issue:44361", "issue:44479|issue:44936", - "issue:30064|issue:33666", - "issue:44977|issue:45356", - "issue:43122|issue:43906", - "issue:28282|issue:33357", "issue:28282|issue:32090", + "issue:28282|issue:33357", + "issue:30064|issue:33666", "issue:28282|issue:35141", + "issue:43122|issue:43906", + "issue:44977|issue:45356", "issue:44464|issue:44977", "issue:43531|issue:43742", "issue:43577|issue:44554", "issue:43653|issue:43927", "issue:43329|issue:45325" ], - "prepared_review_unit_hash": "fa4cbd75ab017a92747ea1823097dd57ea22fd5929fb7c0ebe82915b2b8f7cee", + "prepared_review_unit_hash": "20f9973b334efb4625fd81110350eb911477a185c65a853d07fc522688fd360f", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 31, - "serialized_chars": 37377, - "estimated_input_tokens": 9345, - "estimated_eval_tokens": 18946 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36486, + "estimated_input_tokens": 9122, + "estimated_eval_tokens": 18500 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a grab-bag of unrelated Hugging Face/Transformers issues. The closest soft links are around tokenizer/config regressions, but none are strong enough to treat as true duplicates, so I would not merge any of the proposed pairs.", - "confidence": 0.26, - "canonical_issue_reason": "Issue 43653 is the most central and broadly reusable report among the soft-linked set: it describes a concrete tokenizer special-token handling bug on a generic decode path, which is a better canonical anchor than the more model-specific failures.", + "summary": "Mostly unrelated issues spanning tokenizer/config problems, multimodal regressions, trainer/DeepSpeed failures, and model-specific bugs. Only one plausible duplicate edge stands out: the tied-weights v5 regressions around 44466 and 45005.", + "confidence": 0.91, + "canonical_issue_reason": "If a canonical issue must be chosen, issue 45005 is the broadest representative of the tied-weights v5 regression family and best covers the one plausible duplicate pair (44466/45005).", "canonical_pr_reason": null, - "best_issue_reason": "Issue 43653 is the best representative issue in this cluster because it is concrete, actionable, and closest to the recurring tokenizer-related themes seen in the weaker soft matches.", + "best_issue_reason": "Issue 45005 is the best representative issue in this set because it describes the broader tied-weights regression, while 44466 looks like a narrower symptom of the same underlying bug.", "best_pr_reason": null, "soft_edge_verdicts": [ - { - "left": "issue:41762", - "right": "issue:44936", - "accept": false, - "reason": "Different failures and code paths: DeepSpeed ZeRO-3 loading index error vs trainer.evaluate() failure after training." - }, - { - "left": "issue:43452", - "right": "issue:44843", - "accept": false, - "reason": "One is gguf_file / model loading, the other is an unconditional model_info() call in tokenizer regex patching; not the same bug." - }, { "left": "issue:43688", "right": "issue:45440", "accept": false, - "reason": "Both concern MoE models, but they are different model families and different correctness bugs." - }, - { - "left": "issue:43653", - "right": "issue:44568", - "accept": false, - "reason": "Both are tokenizer special-token bugs, but they affect different tokenizers and have different symptoms and fixes." + "reason": "Both involve MoE-related model behavior, but one is auxiliary-loss normalization and the other is DeepSeekV3 implementation divergence; different bug surfaces and code paths." }, { - "left": "issue:43720", - "right": "issue:44898", + "left": "issue:44625", + "right": "issue:44877", "accept": false, - "reason": "Packed-weight loading in BitNet is unrelated to Perceiver positional interpolation failure." + "reason": "Both are config-related, but one is num_labels propagation in Qwen3.5 and the other is strict loading for granite_speech; not the same underlying issue." }, { "left": "issue:43819", "right": "issue:44811", "accept": false, - "reason": "DAC latent decoding mismatch and Whisper batch_decode skip_special_tokens are unrelated behaviors." + "reason": "DAC latent decoding mismatch and Whisper batch_decode skip_special_tokens are unrelated tokenizer/model behaviors." }, { - "left": "issue:44625", - "right": "issue:44877", + "left": "issue:43720", + "right": "issue:44898", "accept": false, - "reason": "Both involve config loading, but one is num_labels propagation and the other is strict config rejection; not the same underlying issue." + "reason": "Packed-weight loading for BitNet and Perceiver interpolation failures are different model paths with no shared bug." }, { - "left": "issue:43295", - "right": "issue:45356", + "left": "issue:41762", + "right": "issue:44936", "accept": false, - "reason": "Both are tokenizer regressions, but they target different releases, models, and failure modes." + "reason": "Gemma3 ZeRO-3 loading index error and trainer.evaluate-after-train failure are unrelated trainer/loading issues." }, { - "left": "issue:44464", - "right": "issue:44743", + "left": "issue:43653", + "right": "issue:44568", "accept": false, - "reason": "Both are generation-related, but one is compiled-forward inconsistency and the other is recurrent-state reset in modular Qwen3.5." + "reason": "Both touch special-token handling, but they affect different tokenizers and different behaviors (mask token registration vs BOS/EOS insertion)." }, { - "left": "issue:40444", - "right": "issue:44811", + "left": "issue:44464", + "right": "issue:44743", "accept": false, - "reason": "Multi-image IterableDataset finetuning and Whisper decoding are unrelated." + "reason": "Chunked generation with compiled forward and Qwen3.5 recurrent-state reset are different generation bugs." }, { "left": "issue:44466", "right": "issue:45005", - "accept": false, - "reason": "Both mention tied weights, but one is serialization/device-dependent lm_head.weight behavior while the other is a broader translation-model tied-weights report." + "accept": true, + "reason": "Both describe tied-weights regressions in v5; 45005 is the broader formulation and 44466 reads like a concrete manifestation of the same serialization/tied-weight bug." }, { - "left": "issue:41720", - "right": "issue:43906", + "left": "issue:43295", + "right": "issue:45356", "accept": false, - "reason": "Auto device mapping cudaErrorAssert and an isolated reproduction of another issue are not enough to establish the same bug." + "reason": "Processor.tokenizer regression and Kimi-K2.5 codec/regex issues are unrelated tokenizer regressions." }, { "left": "issue:24643", "right": "issue:29127", "accept": false, - "reason": "DeepSpeed weight dimensionality error and LayoutLMv3 box-count error-message clarity are unrelated." + "reason": "DeepSpeed weight shape error and LayoutLMv3 box-validation messaging are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "Qwen3 auto device-mapping cuda assert and an isolated reproduction of a different issue are not the same bug." }, { "left": "issue:43525", "right": "issue:43643", "accept": false, - "reason": "Both involve missing config fields, but one is a specific Llama4 pad_token_id attribute issue and the other is generic missing fields with trust_remote_code." + "reason": "Missing pad_token_id in Llama4Config and missing fields from trust_remote_code AutoConfig are distinct config-loading problems." + }, + { + "left": "issue:40444", + "right": "issue:44811", + "accept": false, + "reason": "IterableDataset multi-image fine-tuning and Whisper batch_decode token skipping are unrelated." }, { "left": "issue:44843", "right": "issue:44991", "accept": false, - "reason": "Both are tokenizer-loading regressions, but the titles point to different failure mechanisms and no shared concrete code path." + "reason": "Offline model_info call in AutoTokenizer patching and tokenizer load failure for EMBEDDIA/est-roberta are different failures." }, { "left": "issue:44361", "right": "issue:44442", "accept": false, - "reason": "Different tokenizer failures: MLukeTokenizer task AttributeError vs FastSpeech2ConformerTokenizer load failure." + "reason": "MLukeTokenizer AttributeError and FastSpeech2ConformerTokenizer load failure are separate tokenizer-specific bugs." }, { "left": "issue:43577", "right": "issue:43697", "accept": false, - "reason": "Blip2 dtype loading and RTDetrV2 output drift are unrelated." + "reason": "BLIP2 dtype propagation and RTDetrV2 output differences are unrelated model-behavior bugs." }, { "left": "issue:38175", "right": "issue:43295", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and a tokenizer regression are unrelated." + "reason": "Zero probabilities in SigLIP2 and processor.tokenizer regression are unrelated." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Both are token-related, but one is special-token registration/decode output and the other is custom token IDs being lost after save/load." + "reason": "BigBirdTokenizer special-token registration and DiaConfig custom token ID loss are different tokenizer/config issues." }, { "left": "issue:44479", "right": "issue:44936", "accept": false, - "reason": "Video-input regression for Qwen VL models and trainer.evaluate() failure are unrelated." + "reason": "Video-input regression across Qwen models and trainer.evaluate-after-train are unrelated." }, { - "left": "issue:30064", - "right": "issue:33666", + "left": "issue:28282", + "right": "issue:32090", "accept": false, - "reason": "Void segmentation-map processing and Qwen2-VL multi-GPU training are unrelated." + "reason": "Missing PyTorch import and a NoneType GPU broadcast error are unrelated." }, { - "left": "issue:44977", - "right": "issue:45356", + "left": "issue:28282", + "right": "issue:33357", "accept": false, - "reason": "Flash-attention generation issues and tokenizer codec/regression issues are different problems." + "reason": "ImportError for missing torch and MacOS bus error are unrelated." }, { - "left": "issue:43122", - "right": "issue:43906", + "left": "issue:30064", + "right": "issue:33666", "accept": false, - "reason": "Different tokenization regression reports with no evidence they are the same bug." + "reason": "Void segmentation map processing and Qwen2-VL multi-GPU training are not the same problem." }, { "left": "issue:28282", - "right": "issue:33357", + "right": "issue:35141", "accept": false, - "reason": "PyTorch import error for AutoModel and a MacOS bus error on a CLIP model are unrelated." + "reason": "Torch import failure and tied-embedding reinitialization are unrelated." }, { - "left": "issue:28282", - "right": "issue:32090", + "left": "issue:43122", + "right": "issue:43906", "accept": false, - "reason": "Missing PyTorch and a NoneType broadcast error are unrelated." + "reason": "Tokenization differences across versions and an isolated reproduction of another issue are not enough to establish the same bug." }, { - "left": "issue:28282", - "right": "issue:35141", + "left": "issue:44977", + "right": "issue:45356", "accept": false, - "reason": "ImportError for PyTorch and tied-output-embedding reinitialization are unrelated." + "reason": "Qwen3.5 flash-attention generation failure and Kimi-K2.5 tokenizer regression are unrelated." }, { "left": "issue:44464", "right": "issue:44977", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is unrelated to Qwen3.5 flash-attention generation failure." + "reason": "Both are generation-related, but one is compiled-forward chunked generation and the other is flash-attention model behavior; different code paths." }, { "left": "issue:43531", "right": "issue:43742", "accept": false, - "reason": "Sliding-window behavior in Qwen3-MoE and a MobileLLM key error are unrelated." + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM-125M loading key errors are unrelated." }, { "left": "issue:43577", "right": "issue:44554", "accept": false, - "reason": "Blip2 dtype loading and MPS attention correctness are unrelated." + "reason": "BLIP2 dtype issue and MPS attention correctness issue are different model/runtime problems." }, { "left": "issue:43653", "right": "issue:43927", "accept": false, - "reason": "Special-token registration/decode output and loss of custom token IDs after save/load are different bugs." + "reason": "Tokenizer special-token registration and DiaConfig save/load token-ID loss are unrelated." }, { "left": "issue:43329", "right": "issue:45325", "accept": false, - "reason": "Undefined helpers in a video-token counting branch and still-image RoPE scaling are different regressions." + "reason": "Video-branch undefined variables in multimodal token counting and still-image temporal position scaling are both multimodal bugs, but not the same underlying issue." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet and is mostly conservative. The only accepted soft edge, 44466/45005, is plausibly the same tied-weights v5 regression family based on the titles, and the other pairs are appropriately rejected as different bug surfaces." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -5368,14 +5518,12 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:32090", "issue:33666", "issue:34567", "issue:36010", - "issue:36331", - "issue:39692", "issue:41720", "issue:43012", "issue:43257", @@ -5407,6 +5555,7 @@ "issue:44464", "issue:44479", "issue:44488", + "issue:44589", "issue:44610", "issue:44625", "issue:44821", @@ -5417,256 +5566,256 @@ "issue:45084", "issue:45200", "issue:45229", - "issue:45357" + "issue:45357", + "issue:45520" ], "soft_pairs": [ "issue:43452|issue:44451", + "issue:44479|issue:44625", "issue:43012|issue:43577", "issue:41720|issue:44464", - "issue:44479|issue:44625", "issue:43504|issue:43720", "issue:43550|issue:44361", - "issue:44849|issue:45357", "issue:43504|issue:43550", + "issue:32090|issue:33666", + "issue:32090|issue:34567", + "issue:43504|issue:44610", + "issue:43262|issue:45084", + "issue:43278|issue:43688", "issue:45200|issue:45229", + "issue:44849|issue:45357", "issue:43334|issue:43572", "issue:43742|issue:44821", "issue:43540|issue:43716", - "issue:32090|issue:34567", - "issue:32090|issue:33666", - "issue:43504|issue:44610", - "issue:43525|issue:43572", - "issue:43278|issue:43688", - "issue:43262|issue:45084", - "issue:43756|issue:44913", "issue:43493|issue:43825", + "issue:43756|issue:44913", + "issue:43525|issue:43572", "issue:33666|issue:36010", "issue:43688|issue:43873", "issue:43720|issue:43927", "issue:43606|issue:43927", - "issue:44488|issue:44843", "issue:43278|issue:43844", + "issue:44488|issue:44843", + "issue:43257|issue:43299", + "issue:44589|issue:45520", "issue:43540|issue:44977", "issue:43720|issue:44361", - "issue:43257|issue:43299", - "issue:43329|issue:43606", - "issue:39692|issue:43493", - "issue:33666|issue:36331" + "issue:43329|issue:43606" ], - "prepared_review_unit_hash": "c7dc6f2f1c6f1c09a15d2ed0cfee8d615dc47d9a5e2b36ab2d34bf1292090636", + "prepared_review_unit_hash": "f4f03343c43b047ee764ce8209e2e1900ff7dfb5d977960f0fe85cbe85e9d552", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37328, - "estimated_input_tokens": 9332, - "estimated_eval_tokens": 18920 + "soft_pair_count": 31, + "serialized_chars": 36961, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items are mostly unrelated issue reports across different models and code paths. The soft pairs only share broad subsystem overlap (loading, dtype, multimodal, training, config) and do not look like the same underlying bug/change.", - "confidence": 0.86, - "canonical_issue_reason": "issue:43329 is the most concrete, code-path-specific report here: it names the exact function, the failing video branch, and the undefined symbols, making it the clearest representative issue.", + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues; the soft pairs share broad vocabulary but not the same concrete bug or fix path. No single issue stands out as a safe canonical duplicate target.", + "confidence": 0.95, + "canonical_issue_reason": "No single issue is a clear canonical representative: these items span unrelated regressions in loading, training, quantization, multimodal processing, config serialization, and warning/error-message behavior.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43329 is the best issue choice because it is narrowly scoped, actionable, and describes a precise bug surface rather than a broad symptom cluster.", + "best_issue_reason": "No issue is strong enough to serve as a global best canonical for the cluster, since the pairwise similarities are mostly superficial and the underlying failures differ.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43452", "right": "issue:44451", "accept": false, - "reason": "Both are model-loading failures, but they affect different models and failure modes." + "reason": "Both are model-loading failures, but they involve different APIs/models and likely different root causes; too broad to merge." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "One is a v5 video-input regression across Qwen VL variants; the other is a Qwen3.5 num_labels propagation bug. Different code paths." }, { "left": "issue:43012", "right": "issue:43577", "accept": false, - "reason": "Both involve dtype/precision, but one is a compile warning and the other is BLIP2 dtype propagation on load." + "reason": "Both mention dtype/precision, but one is a compile-time warning and the other is incorrect load-time dtypes for Blip2; not the same bug." }, { "left": "issue:41720", "right": "issue:44464", "accept": false, - "reason": "Different problems: device-mapping cuda assert vs inconsistent outputs from compiled forward." - }, - { - "left": "issue:44479", - "right": "issue:44625", - "accept": false, - "reason": "A video-input regression and a num_labels propagation bug are separate issues." + "reason": "CUDA assert with auto device mapping vs inconsistent chunked generation under compiled forward are unrelated failures." }, { "left": "issue:43504", "right": "issue:43720", "accept": false, - "reason": "Different loading bugs in different models: legacy field handling vs packed-weight unpacking." + "reason": "Legacy-field load failure for BEiT segmentation and unpacked-weights accelerate loading for BitNet are distinct issues." }, { "left": "issue:43550", "right": "issue:44361", "accept": false, - "reason": "Unrelated failures: torch.compile+SDPA on Bamba vs tokenizer AttributeError on tasks." - }, - { - "left": "issue:44849", - "right": "issue:45357", - "accept": false, - "reason": "Different Qwen3.5 bugs: hidden-state output handling vs save_pretrained key regression." + "reason": "A torch.compile/SDPA model bug and a tokenizer AttributeError are unrelated." }, { "left": "issue:43504", "right": "issue:43550", "accept": false, - "reason": "No shared underlying bug; one is a pretrained-load compatibility issue, the other a compile/runtime issue." + "reason": "Different models, different subsystems, and different failure modes; no common concrete bug." }, { - "left": "issue:45200", - "right": "issue:45229", + "left": "issue:32090", + "right": "issue:33666", "accept": false, - "reason": "Text-only fine-tuning defaults and multi-GPU inference OOM are distinct problems." + "reason": "Generic Trainer broadcast NoneType error vs Qwen2-VL multi-GPU training problem; too little evidence of the same root cause." }, { - "left": "issue:43334", - "right": "issue:43572", + "left": "issue:32090", + "right": "issue:34567", "accept": false, - "reason": "Both are missing-pad-token style config errors, but for different models and different config fields." + "reason": "Trainer broadcast TypeError and TrainerState token counter not updating are separate training bugs." }, { - "left": "issue:43742", - "right": "issue:44821", + "left": "issue:43504", + "right": "issue:44610", "accept": false, - "reason": "A key error loading MobileLLM is unrelated to AutoImageProcessor URL loading." + "reason": "Pretrained preset loading regression and OmDet-Turbo processor/model size mismatch are unrelated." }, { - "left": "issue:43540", - "right": "issue:43716", + "left": "issue:43262", + "right": "issue:45084", "accept": false, - "reason": "Video-processing validation and image-preprocessor dtype mismatch are separate bugs." + "reason": "Audio chat-template sampling-rate default and non-template-node compilation error are different parts of the stack." }, { - "left": "issue:32090", - "right": "issue:34567", + "left": "issue:43278", + "right": "issue:43688", "accept": false, - "reason": "Trainer broadcast NoneType and num_input_tokens_seen not updating are different trainer issues." + "reason": "Embedding dtype drift in eval and auxiliary-loss normalization are unrelated training behaviors." }, { - "left": "issue:32090", - "right": "issue:33666", + "left": "issue:45200", + "right": "issue:45229", "accept": false, - "reason": "Multi-GPU training context overlaps, but the concrete failures are different." + "reason": "Text-only mm_token_type_ids defaulting and Gemma4 multi-GPU inference OOM are separate issues." }, { - "left": "issue:43504", - "right": "issue:44610", + "left": "issue:44849", + "right": "issue:45357", "accept": false, - "reason": "Legacy-field model loading and OmDet input-size mismatch are unrelated." + "reason": "Both are Qwen3.5 regressions, but one is output_hidden_states behavior and the other is save_pretrained key serialization; different bugs." }, { - "left": "issue:43525", + "left": "issue:43334", "right": "issue:43572", "accept": false, - "reason": "Both mention missing token-id fields, but they are model-specific config regressions with different attributes." + "reason": "Missing pad_token_id on Qwen3-VL text config vs missing pad_token_idx on StableLmConfig are similar symptom classes but not the same bug." }, { - "left": "issue:43278", - "right": "issue:43688", + "left": "issue:43742", + "right": "issue:44821", "accept": false, - "reason": "Embedding dtype drift and auxiliary-loss normalization are unrelated." + "reason": "KeyError while loading a model vs AutoImageProcessor URL loading are unrelated loader paths." }, { - "left": "issue:43262", - "right": "issue:45084", + "left": "issue:43540", + "right": "issue:43716", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting is unrelated to template-node compilation errors." + "reason": "Video-processing ValueError in Qwen3OmniMoe and image-preprocessor dtype mismatch in Mistral-3 are different multimodal bugs." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and pipeline task-support error message are not the same underlying issue." }, { "left": "issue:43756", "right": "issue:44913", "accept": false, - "reason": "Both involve rotary/RoPE settings, but they describe different bugs in different model/config paths." + "reason": "Both touch RoPE-related configuration, but they affect different models and different mechanisms; not a mergeable duplicate." }, { - "left": "issue:43493", - "right": "issue:43825", + "left": "issue:43525", + "right": "issue:43572", "accept": false, - "reason": "SigLIP2 implementation mismatch and pipeline error-message wording are unrelated." + "reason": "Both are missing token-id fields after v5 changes, but they concern different configs and different field names." }, { "left": "issue:33666", "right": "issue:36010", "accept": false, - "reason": "Qwen2-VL multi-GPU training is unrelated to the GenerationMixin import regression." + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are unrelated." }, { "left": "issue:43688", "right": "issue:43873", "accept": false, - "reason": "Aux-loss normalization and quantization offloading are different mechanisms and failures." + "reason": "Auxiliary-loss normalization and quantization/offloading behavior are distinct issues." }, { "left": "issue:43720", "right": "issue:43927", "accept": false, - "reason": "Packed-weight loading and lost custom token IDs are unrelated bugs." + "reason": "Packed-weight handling in accelerate loading and DiaConfig save/load token IDs are unrelated regression classes." }, { "left": "issue:43606", "right": "issue:43927", "accept": false, - "reason": "CPU offload device mismatch is unrelated to DiaConfig token-ID persistence." + "reason": "CPU offload device mismatch and config token-ID persistence are different bugs." }, { "left": "issue:43278", "right": "issue:43844", "accept": false, - "reason": "Different training-time anomalies: dtype conversion vs exploding gradients under ZeRO-3." + "reason": "Eval dtype drift vs exploding gradients under ZeRO-3 have no concrete overlap." }, { - "left": "issue:43540", - "right": "issue:44977", + "left": "issue:44488", + "right": "issue:44843", "accept": false, - "reason": "Different Qwen3-family issues: video input processing vs flash-attention generation." + "reason": "Model loading failure and offline-mode hub lookup bug are different problems." }, { - "left": "issue:43720", - "right": "issue:44361", + "left": "issue:43257", + "right": "issue:43299", "accept": false, - "reason": "Packed-weight accelerate loading and tokenizer task errors are unrelated." + "reason": "Both concern Qwen3 MOE loading, but one is accelerate+deepspeed conversion and the other is a transformer-dev version breakage; not clearly the same fix." }, { - "left": "issue:43257", - "right": "issue:43299", + "left": "issue:44589", + "right": "issue:45520", "accept": false, - "reason": "Both are Qwen MoE loading issues, but they have different causes and model variants." + "reason": "Missing Float8 storage object and flash_attn import key error are unrelated environment/compatibility issues." }, { - "left": "issue:43329", - "right": "issue:43606", + "left": "issue:43540", + "right": "issue:44977", "accept": false, - "reason": "Video multimodal token counting and Bark CPU offload device mismatch are unrelated." + "reason": "Video-input processing error and flash-attention generation issue in Qwen3.5 are different code paths." }, { - "left": "issue:39692", - "right": "issue:43493", + "left": "issue:43720", + "right": "issue:44361", "accept": false, - "reason": "SigLIP2 docs-example failures and SigLIP2 implementation discrepancies are different bug reports." + "reason": "Accelerate loading of packed BitNet weights and MLukeTokenizer task AttributeError are unrelated." }, { - "left": "issue:33666", - "right": "issue:36331", + "left": "issue:43329", + "right": "issue:43606", "accept": false, - "reason": "Multi-GPU Qwen2-VL training and the compute_loss signature change are separate trainer issues." + "reason": "Undefined multimodal token helper in video branch and Bark CPU-offload device mismatch are different bugs." } ] }, @@ -5688,15 +5837,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:15354", "issue:33357", "issue:33666", + "issue:36331", "issue:38175", - "issue:39401", "issue:39692", - "issue:43232", "issue:43257", "issue:43278", "issue:43295", @@ -5726,230 +5874,246 @@ "issue:44493", "issue:44521", "issue:44554", - "issue:44568", "issue:44589", "issue:44734", "issue:44792", - "issue:44849", "issue:44964", "issue:44977", + "issue:45072", + "issue:45084", + "issue:45198", "issue:45200", "issue:45290", "issue:45405", - "issue:45446" + "issue:45446", + "issue:45520" ], "soft_pairs": [ - "issue:43756|issue:44060", + "issue:39692|issue:43493", "issue:44964|issue:45405", + "issue:33666|issue:36331", "issue:43493|issue:43844", - "issue:43824|issue:44589", + "issue:43756|issue:44060", + "issue:43646|issue:44466", "issue:44423|issue:44734", "issue:43688|issue:43746", - "issue:43572|issue:44291", - "issue:43646|issue:44466", + "issue:43824|issue:44589", "issue:44521|issue:45290", - "issue:43994|issue:44493", - "issue:43504|issue:43606", "issue:33357|issue:33666", + "issue:43572|issue:44291", "issue:39692|issue:43295", + "issue:43504|issue:43606", + "issue:43994|issue:44493", + "issue:43575|issue:45405", "issue:43504|issue:43653", "issue:43927|issue:44361", - "issue:43575|issue:45405", - "issue:43278|issue:43701", "issue:43257|issue:45405", - "issue:43606|issue:44112", + "issue:43278|issue:43701", "issue:44792|issue:45405", + "issue:43606|issue:44112", + "issue:45084|issue:45520", "issue:15354|issue:38175", - "issue:43540|issue:45200", "issue:44554|issue:45446", - "issue:43232|issue:44849", + "issue:43540|issue:45200", "issue:43582|issue:44977", - "issue:39401|issue:44568" + "issue:45072|issue:45198" ], - "prepared_review_unit_hash": "8b230a8c7dd5eb56530e7912b6b865be0d079d014c6e0417fd15c6cb8ff68840", + "prepared_review_unit_hash": "cd996b56b5fbe79d03a432894d5fb5561c7b2b6ec420c4a2c595b4b774bc99ac", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 35124, - "estimated_input_tokens": 8781, - "estimated_eval_tokens": 17818 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36304, + "estimated_input_tokens": 9076, + "estimated_eval_tokens": 18408 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The candidates are a broad mix of unrelated model, tokenizer, training, serving, and dependency-version bugs. I found no soft pair that clearly represents the same underlying change or defect, so all soft edges should be rejected.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue stands out because the set is heterogeneous and does not form one underlying bug cluster; if a representative must be chosen, issue:44521 is the strongest standalone report due to its activity and reproducibility.", + "summary": "The cluster is heterogeneous and does not present a single duplicate theme. The soft-pair candidates share only broad subsystem keywords at most, not the same concrete bug or fix.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the best standalone issue candidate: it is open, has strong discussion/inbound signal, and describes a clear reproducible multimodal bug.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43756", - "right": "issue:44060", + "left": "issue:39692", + "right": "issue:43493", "accept": false, - "reason": "Different model families and different defects: RoPE-layer dropping vs a tied-weights warning on Qwen3-Next." + "reason": "Both involve SigLIP2, but one is a documentation/example error and the other is a model implementation discrepancy; different bugs and fixes." }, { "left": "issue:44964", "right": "issue:45405", "accept": false, - "reason": "One is a Phi-4 multimodal load failure; the other is a PEFT minimum-version packaging issue." + "reason": "Phi-4 multimodal loading failure vs. a PEFT version bump problem; unrelated failure modes." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Multi-GPU training question for Qwen2-VL is unrelated to a custom trainer signature regression." }, { "left": "issue:43493", "right": "issue:43844", "accept": false, - "reason": "SigLIP2 HF-vs-JAX discrepancy is unrelated to the DeepSpeed/ZeRO-3 gradient growth bug." + "reason": "SigLIP2 implementation mismatch and ZeRO-3 gradient inflation are different code paths and symptoms." }, { - "left": "issue:43824", - "right": "issue:44589", + "left": "issue:43756", + "right": "issue:44060", "accept": false, - "reason": "Importing a Qwen2-VL class and a Float8 storage lookup failure are different code paths and failure modes." + "reason": "Smollm3 RoPE-layer bug and Qwen3-Next tied-weights warning are unrelated model-specific issues." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5-related, but one breaks model initialization while the other is about state dict serialization/tied weights." }, { "left": "issue:44423", "right": "issue:44734", "accept": false, - "reason": "Both involve serve, but one is a multimodal continuous-batching crash and the other is KV-cache continuation indexing." + "reason": "Both are serve crashes, but they hit different pathways: multimodal continuous batching vs KV-cache continuation indexing." }, { "left": "issue:43688", "right": "issue:43746", "accept": false, - "reason": "Auxiliary-loss normalization in OLMoE/GPT OSS is unrelated to PEFT adapter loading from local checkpoints." + "reason": "Aux-loss normalization bug and PEFT-local-checkpoint loading bug are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "ImportError for a model class vs missing float8 storage type; no shared underlying defect." }, { - "left": "issue:43572", - "right": "issue:44291", + "left": "issue:44521", + "right": "issue:45290", "accept": false, - "reason": "Both touch 5.0 migration, but one is a missing config field and the other is an unexpected init argument." + "reason": "Both touch apply_chat_template, but one returns bad masks for multimodal inputs while the other crashes on tool-call assistant messages." }, { - "left": "issue:43646", - "right": "issue:44466", + "left": "issue:33357", + "right": "issue:33666", "accept": false, - "reason": "Custom model initialization regressions are different from lm_head.weight serialization inconsistencies." + "reason": "MacOS bus error on a CLIP model is unrelated to Qwen2-VL multi-GPU training." }, { - "left": "issue:44521", - "right": "issue:45290", + "left": "issue:43572", + "right": "issue:44291", "accept": false, - "reason": "Both use apply_chat_template, but one is zero assistant masks for multimodal input and the other is a tool-call crash." + "reason": "StableLmConfig pad_token_idx regression and init_empty_weights unexpected argument are separate v5 breakages." }, { - "left": "issue:43994", - "right": "issue:44493", + "left": "issue:39692", + "right": "issue:43295", "accept": false, - "reason": "SigLIP2 bad outputs and unexpected position-id keys are different bugs." + "reason": "Doc example errors for SigLIP2 do not match the processor.tokenizer regression in multimodal custom model code." }, { "left": "issue:43504", "right": "issue:43606", "accept": false, - "reason": "A legacy field load failure in Beit is unrelated to a CPU-offload device mismatch in Bark." + "reason": "Legacy field load failure in Beit differs from CPU-offload device mismatch in bark-small." }, { - "left": "issue:33357", - "right": "issue:33666", + "left": "issue:43994", + "right": "issue:44493", "accept": false, - "reason": "MacOS bus error on a CLIP model and Qwen2-VL multi-GPU training are unrelated." + "reason": "SigLIP2 bad outputs and unexpected position-id key warnings are distinct issues." }, { - "left": "issue:39692", - "right": "issue:43295", + "left": "issue:43575", + "right": "issue:45405", "accept": false, - "reason": "Both mention processor/tokenizer multimodal usage, but one is a broken docs example and the other is a regression in custom code handling." + "reason": "Tensor-parallel OOM when loading Qwen2-57B is unrelated to the PEFT version constant bump." }, { "left": "issue:43504", "right": "issue:43653", "accept": false, - "reason": "Legacy preset-field loading and BigBirdTokenizer special-token registration are unrelated tokenizer/model issues." + "reason": "Model preset loading with a legacy field is unrelated to BigBirdTokenizer special-token registration." }, { "left": "issue:43927", "right": "issue:44361", "accept": false, - "reason": "A save/load token-ID mismatch in DiaConfig is not the same as MLukeTokenizer's AttributeError." + "reason": "Custom token IDs lost on save/load and MLukeTokenizer task AttributeError are different tokenizer bugs." }, { - "left": "issue:43575", + "left": "issue:43257", "right": "issue:45405", "accept": false, - "reason": "Tensor-parallel OOM on Qwen2-57B is unrelated to the PEFT version bump issue." + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to the PEFT release-version mismatch." }, { "left": "issue:43278", "right": "issue:43701", "accept": false, - "reason": "Embedding dtype drift during eval and resume-from-checkpoint key mismatch are different problems." + "reason": "Embedding dtype drift during eval and resume_from_checkpoint key mismatch are unrelated regressions." }, { - "left": "issue:43257", + "left": "issue:44792", "right": "issue:45405", "accept": false, - "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed is unrelated to the PEFT release-version issue." + "reason": "Janus image-generation test failure has nothing to do with the PEFT version bump." }, { "left": "issue:43606", "right": "issue:44112", "accept": false, - "reason": "Same project area, but one is a runtime CPU-offload mismatch and the other is a stale GraniteSpeech CI test." + "reason": "Both mention GraniteSpeech only indirectly, but one is a real CPU-offload device mismatch and the other is a stale CI test." }, { - "left": "issue:44792", - "right": "issue:45405", + "left": "issue:45084", + "right": "issue:45520", "accept": false, - "reason": "A Janus image-generation test failure is unrelated to the PEFT minimum-version bump." + "reason": "Template-node compilation error and flash_attn import KeyError on Python 3.13 are unrelated." }, { "left": "issue:15354", "right": "issue:38175", "accept": false, - "reason": "TorchScript GeneratorExp export failure and SigLIP2 zero-probabilities are different model/export bugs." - }, - { - "left": "issue:43540", - "right": "issue:45200", - "accept": false, - "reason": "Qwen3OmniMoe video processing ValueError is unrelated to Gemma 4 token-type defaults for text-only fine-tuning." + "reason": "TorchScript GeneratorExp export failure and unexpected zero probabilities in SigLIP2 are unrelated." }, { "left": "issue:44554", "right": "issue:45446", "accept": false, - "reason": "An MPS attention correctness issue is not the same as a PyTorch version-check bug in flex_attention." + "reason": "Attention correctness with mismatched head dims is a different bug from a PyTorch version check in flex_attention imports." }, { - "left": "issue:43232", - "right": "issue:44849", + "left": "issue:43540", + "right": "issue:45200", "accept": false, - "reason": "Generation state update after sync_gpus and Qwen3.5 output_hidden_states behavior are distinct generation bugs." + "reason": "Video-input processing error in Qwen3OmniMoe is unrelated to Gemma 4 text-only fine-tuning defaults." }, { "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 and Qwen3.5 flash-attention generation problems are different issues." }, { - "left": "issue:39401", - "right": "issue:44568", + "left": "issue:45072", + "right": "issue:45198", "accept": false, - "reason": "Both are tokenizer issues, but they affect different models and different behaviors: offset mapping vs special-token insertion." + "reason": "bfloat16 dtype mismatches in inference and Wav2Vec2 save/tokenization failures are unrelated." } ] }, @@ -5971,12 +6135,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:37428", "issue:38175", + "issue:39401", "issue:40990", "issue:42915", + "issue:43232", "issue:43278", "issue:43295", "issue:43334", @@ -6004,6 +6170,7 @@ "issue:44521", "issue:44554", "issue:44561", + "issue:44568", "issue:44589", "issue:44610", "issue:44625", @@ -6012,8 +6179,6 @@ "issue:44936", "issue:44964", "issue:44987", - "issue:45003", - "issue:45072", "issue:45127", "issue:45198", "issue:45200", @@ -6022,255 +6187,252 @@ "issue:45412" ], "soft_pairs": [ - "issue:45072|issue:45198", - "issue:37428|issue:43957", + "issue:43232|issue:44849", "issue:43844|issue:44964", + "issue:39401|issue:44568", + "issue:37428|issue:43957", "issue:44315|issue:44849", - "issue:43334|issue:43525", "issue:44554|issue:45356", - "issue:44898|issue:45412", + "issue:43334|issue:43525", "issue:43295|issue:44610", "issue:45200|issue:45325", - "issue:37428|issue:43824", - "issue:44466|issue:45127", "issue:43493|issue:43701", + "issue:44898|issue:45412", + "issue:44466|issue:45127", "issue:43873|issue:44464", "issue:40990|issue:44625", + "issue:37428|issue:43824", "issue:43388|issue:43526", - "issue:43572|issue:44589", "issue:43572|issue:44987", "issue:44514|issue:44521", "issue:43844|issue:44936", - "issue:43824|issue:43975", + "issue:43572|issue:44589", "issue:43575|issue:44464", + "issue:43824|issue:43975", "issue:44442|issue:45198", - "issue:38175|issue:45412", - "issue:42915|issue:43824", - "issue:43295|issue:43873", "issue:43295|issue:43450", + "issue:43295|issue:43873", + "issue:38175|issue:45412", "issue:43278|issue:43856", - "issue:44849|issue:45003", "issue:44360|issue:44561", - "issue:43295|issue:43994" + "issue:43295|issue:43994", + "issue:42915|issue:43824" ], - "prepared_review_unit_hash": "929396bcbf91ec348f7df66d57b560014db6a19c193e3371298f8f99e72dfd27", + "prepared_review_unit_hash": "8d01796bb013ead3f5dd809762577d0b7db38e0655c69843e165c717a4eb1c01", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36253, - "estimated_input_tokens": 9064, - "estimated_eval_tokens": 18384 + "serialized_chars": 36838, + "estimated_input_tokens": 9210, + "estimated_eval_tokens": 18676 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Most items are unrelated issue reports spanning different regressions and subsystems. I found one plausible duplicate pair around the missing `pad_token_id` regression across config classes; the rest are too symptom- or model-specific to merge confidently. No PRs were present.", - "confidence": 0.63, - "canonical_issue_reason": "Issue 44561 is the best broad canonical issue: it describes a clear v5 regression with a systemic API removal affecting remote-code models, making it a stronger representative than narrower model-specific reports.", + "summary": "I don\u2019t see any true duplicate pairs here; the list spans many unrelated issues across loading, tokenization, generation, training, and multimodal processing. No PRs are present.", + "confidence": 0.92, + "canonical_issue_reason": "No clear canonical issue emerges because the items are heterogeneous and do not share one underlying bug or code path.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 44561 has the clearest combination of regression framing, broad applicability, and actionable root-cause direction, so it is the strongest issue candidate overall.", + "best_issue_reason": "If forced to pick one representative, issue:44521 is the most active/referenced recent bug, but it is only a narrow multimodal chat-template regression and not representative of the whole set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:45072", - "right": "issue:45198", + "left": "issue:43232", + "right": "issue:44849", "accept": false, - "reason": "Different failures: bfloat16 inference dtype mismatch vs save/tokenization failure. Same broad area, but not the same bug." + "reason": "Different problems: generation kwargs after sync_gpus vs Qwen3.5 hidden-states behavior." }, { - "left": "issue:37428", - "right": "issue:43957", + "left": "issue:43844", + "right": "issue:44964", "accept": false, - "reason": "Unrelated breakages: flash-attention import error vs meta-device loading failure." + "reason": "ZeRO-3 gradient anomaly and Phi-4 multimodal loading failure are unrelated." }, { - "left": "issue:43844", - "right": "issue:44964", + "left": "issue:39401", + "right": "issue:44568", "accept": false, - "reason": "Different code paths and symptoms: DeepSpeed ZeRO-3 gradient growth vs model-loading failure for a multimodal checkpoint." + "reason": "Both are tokenizer bugs, but one is offset mapping and the other is special-token insertion; different code paths." }, { - "left": "issue:44315", - "right": "issue:44849", + "left": "issue:37428", + "right": "issue:43957", "accept": false, - "reason": "Different models and bugs: Liger Kernel application with `model_init` vs Qwen3.5 hidden-state behavior." + "reason": "ImportError for a missing flash-attention helper is not the same as meta-device model-loading breakage." }, { - "left": "issue:43334", - "right": "issue:43525", - "accept": true, - "reason": "Both report the same underlying regression: `pad_token_id` is missing on model config objects and causes an `AttributeError` during loading/usage." + "left": "issue:44315", + "right": "issue:44849", + "accept": false, + "reason": "Liger Kernel application in model_init is unrelated to Qwen3.5 output_hidden_states." }, { "left": "issue:44554", "right": "issue:45356", "accept": false, - "reason": "Different problems: an MPS attention correctness issue vs a tokenizer/codec regression in Kimi-K2.5." + "reason": "MPS attention correctness issue and Kimi tokenizer regression do not share a concrete bug." }, { - "left": "issue:44898", - "right": "issue:45412", + "left": "issue:43334", + "right": "issue:43525", "accept": false, - "reason": "Different subsystems entirely: image-classification resolution handling vs RT-DETR memory release behavior." + "reason": "Similar symptom (missing pad_token_id), but different models/configs and likely separate fixes." }, { "left": "issue:43295", "right": "issue:44610", "accept": false, - "reason": "Different processor regressions: custom processor/tokenizer access vs OmDet-Turbo input-size mismatch." + "reason": "Processor.tokenizer regression is unrelated to OmDet-Turbo image size mismatch." }, { "left": "issue:45200", "right": "issue:45325", "accept": false, - "reason": "Both involve multimodal/VL models, but one is about `mm_token_type_ids` defaults and the other about rope position scaling." + "reason": "Different multimodal bugs: token_type_ids defaulting vs rope/position_ids scaling." }, { - "left": "issue:37428", - "right": "issue:43824", + "left": "issue:43493", + "right": "issue:43701", "accept": false, - "reason": "Both are import failures, but for different symbols and different model families; no shared underlying bug is evident." + "reason": "SigLIP2 implementation discrepancy and checkpoint key mismatch are unrelated." }, { - "left": "issue:44466", - "right": "issue:45127", + "left": "issue:44898", + "right": "issue:45412", "accept": false, - "reason": "Serialization/tied-weights bug vs LoRA merge collapse with extended vocabulary; related theme, but not the same concrete issue." + "reason": "Perceiver resolution handling and RT-DETR memory reclamation are different issues." }, { - "left": "issue:43493", - "right": "issue:43701", + "left": "issue:44466", + "right": "issue:45127", "accept": false, - "reason": "SigLIP2 correctness discrepancy vs checkpoint resume key mismatch are distinct failures." + "reason": "Serialization of tied weights is separate from LoRA merge collapse with extended vocab." }, { "left": "issue:43873", "right": "issue:44464", "accept": false, - "reason": "Quantization offloading bug vs compiled-forward chunked generation mismatch; different code paths." + "reason": "Quantization offloading and compiled-forward chunked generation are different failure modes." }, { "left": "issue:40990", "right": "issue:44625", "accept": false, - "reason": "Different model behaviors: perplexity anomaly vs `num_labels` propagation bug." + "reason": "High perplexity on gpt-oss-20b is unrelated to Qwen3.5 num_labels propagation." }, { - "left": "issue:43388", - "right": "issue:43526", + "left": "issue:37428", + "right": "issue:43824", "accept": false, - "reason": "Different label-handling bugs in different components: metrics gathering vs image processor reduction." + "reason": "Two different missing-import symbols; not the same underlying import regression." }, { - "left": "issue:43572", - "right": "issue:44589", + "left": "issue:43388", + "right": "issue:43526", "accept": false, - "reason": "Config attribute regression vs float8 storage lookup error; unrelated issues." + "reason": "Both involve labels, but gather_for_metrics batching and BeitImageProcessorFast reduce_labels are distinct bugs." }, { "left": "issue:43572", "right": "issue:44987", "accept": false, - "reason": "A missing `pad_token_idx` config attribute is not the same as a model-load failure for a specific checkpoint." + "reason": "StableLm pad_token_idx regression and loading physical-intelligence/fast on 5.1 are not the same bug." }, { "left": "issue:44514", "right": "issue:44521", "accept": false, - "reason": "Same processor family, but one is a batched-input crash and the other is incorrect assistant masks; too different to treat as the same bug." + "reason": "Both hit apply_chat_template, but one crashes on batched padding=False input while the other produces wrong assistant masks; too different to merge." }, { "left": "issue:43844", "right": "issue:44936", "accept": false, - "reason": "Training instability under ZeRO-3 is unrelated to `trainer.evaluate()` failing after training." + "reason": "Gradient growth under ZeRO-3 and trainer.evaluate after train are unrelated training issues." }, { - "left": "issue:43824", - "right": "issue:43975", + "left": "issue:43572", + "right": "issue:44589", "accept": false, - "reason": "Different import/tokenization issues for different models; no concrete shared fix is apparent." + "reason": "Missing pad_token_idx and Float8 storage lookup failure are different compatibility problems." }, { "left": "issue:43575", "right": "issue:44464", "accept": false, - "reason": "TP/OOM during model load is unrelated to chunked generation with compiled forward." + "reason": "TP OOM on Qwen2-57B and chunked-generation inconsistency are unrelated." }, { - "left": "issue:44442", - "right": "issue:45198", + "left": "issue:43824", + "right": "issue:43975", "accept": false, - "reason": "Different surfaces: tokenizer loading failure vs save_pretrained/tokenization failure for Wav2Vec2." + "reason": "Import failure for Qwen2.5-VL is unrelated to DeepSeek-Coder detokenization." }, { - "left": "issue:38175", - "right": "issue:45412", + "left": "issue:44442", + "right": "issue:45198", "accept": false, - "reason": "SigLIP2 zero-probability output is unrelated to RT-DETR memory retention after deletion." + "reason": "AutoTokenizer loading failure for one model and save_pretrained/tokenization failure for another are different bugs." }, { - "left": "issue:42915", - "right": "issue:43824", + "left": "issue:43295", + "right": "issue:43450", "accept": false, - "reason": "FP8 config failure vs missing imported VL model class are unrelated." + "reason": "Processor.tokenizer regression is not the same as batched video-shape handling." }, { "left": "issue:43295", "right": "issue:43873", "accept": false, - "reason": "Processor API regression vs quantization/offloading behavior; different bugs." + "reason": "Custom processor regression and quantization offloading are unrelated." }, { - "left": "issue:43295", - "right": "issue:43450", + "left": "issue:38175", + "right": "issue:45412", "accept": false, - "reason": "Different processor issues: tokenizer/images API regression vs video processor batch-shape bug." + "reason": "SigLIP2 zero probabilities and RT-DETR memory retention are entirely different." }, { "left": "issue:43278", "right": "issue:43856", "accept": false, - "reason": "Embedding dtype drift in evaluate is unrelated to Qwen3 MoE memory usage." - }, - { - "left": "issue:44849", - "right": "issue:45003", - "accept": false, - "reason": "Qwen3.5 hidden-states bug and unsafe `sys.modules` access are not the same issue." + "reason": "Embedding dtype drift in eval is not the same as Qwen3 MoE memory inefficiency." }, { "left": "issue:44360", "right": "issue:44561", "accept": false, - "reason": "DSA indexer content issue vs a v5 API-removal regression; different code paths and impact." + "reason": "DSA indexer/ReLU discussion is unrelated to removal of is_torch_fx_available." }, { "left": "issue:43295", "right": "issue:43994", "accept": false, - "reason": "Custom processor regression vs SigLIP2 bad outputs; unrelated failures." + "reason": "Processor.tokenizer regression and SigLIP2 nonsensical outputs do not share the same root cause." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Qwen3Moe FP8 failure is unrelated to Qwen2.5-VL import errors." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The summary is grounded in the packet and stays conservative overall. The one accepted duplicate pair is plausible from the titles (both missing pad_token_id on config objects), and the other verdicts appropriately distinguish different symptoms/subsystems. No PRs were present, so that part is also consistent." - }, - "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 }, @@ -6285,9 +6447,8 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ - "issue:29127", "issue:29942", "issue:32090", "issue:37428", @@ -6300,7 +6461,6 @@ "issue:43425", "issue:43450", "issue:43493", - "issue:43504", "issue:43526", "issue:43550", "issue:43575", @@ -6313,6 +6473,7 @@ "issue:43901", "issue:43931", "issue:44112", + "issue:44292", "issue:44351", "issue:44448", "issue:44462", @@ -6322,322 +6483,288 @@ "issue:44610", "issue:44743", "issue:44779", + "issue:44849", "issue:44855", "issue:44857", - "issue:44977", "issue:44991", + "issue:45003", "issue:45083", "issue:45127", "issue:45200", - "issue:45229", "issue:45245", "issue:45290", "issue:45310", "issue:45381", - "issue:45478" + "issue:45478", + "issue:45520" ], "soft_pairs": [ "issue:43827|issue:43901", + "issue:44849|issue:45003", "issue:38175|issue:43493", "issue:43653|issue:44112", "issue:44568|issue:44991", - "issue:45127|issue:45245", "issue:43450|issue:43493", - "issue:37428|issue:43425", - "issue:44462|issue:44561", - "issue:43819|issue:45083", + "issue:45127|issue:45245", "issue:45290|issue:45381", - "issue:44561|issue:44779", + "issue:43819|issue:45083", + "issue:44462|issue:44561", "issue:43450|issue:43575", + "issue:44561|issue:44779", "issue:44610|issue:44857", - "issue:37428|issue:44351", + "issue:37428|issue:43425", "issue:43819|issue:44112", "issue:44448|issue:44743", "issue:43262|issue:43450", - "issue:43606|issue:45083", - "issue:45310|issue:45478", - "issue:43819|issue:44493", "issue:43278|issue:43825", + "issue:43606|issue:45083", "issue:29942|issue:32090", + "issue:45310|issue:45478", "issue:43611|issue:43931", + "issue:37428|issue:44351", + "issue:43819|issue:44493", "issue:39692|issue:43526", - "issue:45200|issue:45381", "issue:43335|issue:43550", + "issue:45200|issue:45381", + "issue:43450|issue:43526", "issue:44855|issue:45083", "issue:43404|issue:45200", + "issue:44292|issue:45520", "issue:39692|issue:44493", - "issue:43450|issue:43526", - "issue:38175|issue:44493", - "issue:44977|issue:45229", - "issue:44448|issue:44779", - "issue:43335|issue:43653", - "issue:29127|issue:32090", - "issue:43335|issue:45083", - "issue:43335|issue:43504", - "issue:43550|issue:44112" + "issue:44448|issue:44779" ], - "prepared_review_unit_hash": "77a7c187de2fdb40d674006870a6adfa2b1a511bd6e95c840bd5a76228f62143", + "prepared_review_unit_hash": "b21fb352f604897138bef0b6b5e734c3ef62600a300126d2c12909e6cc2dbd5c", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 38, - "serialized_chars": 39192, - "estimated_input_tokens": 9798, - "estimated_eval_tokens": 19852 + "soft_pair_count": 33, + "serialized_chars": 37621, + "estimated_input_tokens": 9406, + "estimated_eval_tokens": 19068 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: the issues span unrelated models, processors, tokenizers, and version-regression reports. Only one soft edge looks like a true duplicate pair: the repeated Qwen3.5 MoE from_pretrained error report.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 45310 is the best canonical issue: it is the earlier, more complete report for the Qwen3.5 MoE from_pretrained failure, and 45478 reads like a duplicate follow-up.", + "summary": "Most pairs are broad topical similarities, not true duplicates. The only clear duplicate-looking pair is the repeated Qwen3.5 MoE from_pretrained error (45310/45478), with 45310 the better canonical target.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 45310 is the clearer, more detailed report of the same Qwen3.5 MoE from_pretrained failure; 45478 reads like a later duplicate/restatement.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45310 is the strongest representative of the only clear duplicate subcluster and has the fuller bug description.", + "best_issue_reason": "45310 is the most triage-worthy issue here: it names the exact failing model, version area, and concrete loading error, and it cleanly anchors the only strong duplicate pair.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43827", "right": "issue:43901", "accept": false, - "reason": "Both are docs/API drift around pipelines, but they concern different behaviors and different pipeline classes." + "reason": "Both are pipeline-v5 docs/problems, but one is about translation pipeline references and the other is about TextClassificationPipeline return_all_scores; different concrete issues." + }, + { + "left": "issue:44849", + "right": "issue:45003", + "accept": false, + "reason": "Qwen3.5 hidden_states behavior vs unsafe sys.modules access are unrelated code paths." }, { "left": "issue:38175", "right": "issue:43493", "accept": false, - "reason": "Both mention SigLIP2, but one is zero probabilities while the other is HF-vs-JAX implementation mismatch; not the same concrete bug." + "reason": "Same model family, but one reports zero probabilities and the other an HF-vs-JAX implementation discrepancy; not the same bug." }, { "left": "issue:43653", "right": "issue:44112", "accept": false, - "reason": "BigBirdTokenizer special-token decode issue vs a GraniteSpeech CI device-override failure; unrelated." + "reason": "BigBirdTokenizer special-token registration and GraniteSpeech CI device override are unrelated." }, { "left": "issue:44568", "right": "issue:44991", "accept": false, - "reason": "Tokenizer add_special_tokens behavior and EMBEDDIA/est-roberta loading are different tokenizer bugs." - }, - { - "left": "issue:45127", - "right": "issue:45245", - "accept": false, - "reason": "Extended-vocab LoRA collapse and category-count runtime limit are unrelated failures." + "reason": "Both involve tokenizers after v5, but they affect different models and different failure modes." }, { "left": "issue:43450", "right": "issue:43493", "accept": false, - "reason": "Video processor shape bug vs SigLIP2 discrepancy; different components and symptoms." + "reason": "Video processor batching shape bug vs SigLIP2 model discrepancy are different subsystems and symptoms." }, { - "left": "issue:37428", - "right": "issue:43425", + "left": "issue:45127", + "right": "issue:45245", "accept": false, - "reason": "Flash-support import error and Torch 2.10 compatibility are version issues in different code paths." + "reason": "LoRA/vocab merge collapse is unrelated to the categories > 2^24 runtime error." }, { - "left": "issue:44462", - "right": "issue:44561", + "left": "issue:45290", + "right": "issue:45381", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json is unrelated to removal of is_torch_fx_available." + "reason": "Chat-template tool-call crash and Qwen2.5-VL video vision_position_ids bug are different code paths." }, { "left": "issue:43819", "right": "issue:45083", "accept": false, - "reason": "DAC latent/forward mismatch and qwen3_omni_moe feature-length helper behavior are different bugs." + "reason": "DAC latent/STE mismatch is unrelated to qwen3_omni_moe feature-extract length behavior." }, { - "left": "issue:45290", - "right": "issue:45381", + "left": "issue:44462", + "right": "issue:44561", "accept": false, - "reason": "Chat-template tool-call crash and Qwen2.5-VL video vision_position_ids are unrelated preprocessing issues." + "reason": "AutoTokenizer ignoring tokenizer.json and v5 removal of is_torch_fx_available are unrelated." }, { - "left": "issue:44561", - "right": "issue:44779", + "left": "issue:43450", + "right": "issue:43575", "accept": false, - "reason": "Removed helper API breakage vs Deepseek tokenizer regression; different subsystems." + "reason": "Batched video shape output and tensor-parallel OOM are not the same underlying bug." }, { - "left": "issue:43450", - "right": "issue:43575", + "left": "issue:44561", + "right": "issue:44779", "accept": false, - "reason": "Video processor batching shape bug vs tensor-parallel OOM while loading Qwen2-57B; not the same issue." + "reason": "Removed helper breaking trust_remote_code models vs Deepseek tokenizer regression are different issues." }, { "left": "issue:44610", "right": "issue:44857", "accept": false, - "reason": "Processor input-size mismatch and LwDetrImageLoss AMP crash are unrelated vision bugs." + "reason": "OmDet-Turbo input-size mismatch and LwDetrImageLoss AMP/CUDA crash are unrelated." }, { "left": "issue:37428", - "right": "issue:44351", - "accept": false, - "reason": "Flash-attention import problem vs HybridCache import failure; separate missing-symbol regressions." - }, - { - "left": "issue:43819", - "right": "issue:44112", - "accept": false, - "reason": "DAC from_latents mismatch and stale GraniteSpeech device override test are unrelated." - }, - { - "left": "issue:44448", - "right": "issue:44743", - "accept": false, - "reason": "Pegasus v4/v5 output difference and qwen3_5 recurrent-state reset are different model behaviors." - }, - { - "left": "issue:43262", - "right": "issue:43450", - "accept": false, - "reason": "Audio chat-template sampling-rate default and video processor batched-shape bug are unrelated processor issues." - }, - { - "left": "issue:43606", - "right": "issue:45083", - "accept": false, - "reason": "CPU offload device mismatch in bark-small is unrelated to qwen3_omni_moe feature-length helper behavior." - }, - { - "left": "issue:45310", - "right": "issue:45478", - "accept": true, - "reason": "Same Qwen3.5 MoE from_pretrained failure, with 45478 appearing to be a duplicate follow-up of 45310." + "right": "issue:43425", + "accept": false, + "reason": "ImportError for a flash-attention helper and Torch 2.10 incompatibility are only loosely version-related, not the same bug." }, { "left": "issue:43819", - "right": "issue:44493", + "right": "issue:44112", "accept": false, - "reason": "DAC latent mismatch and unexpected position-id keys are different bugs." + "reason": "DAC latent mismatch and GraniteSpeech CI stale device override are unrelated." }, { - "left": "issue:43278", - "right": "issue:43825", + "left": "issue:44448", + "right": "issue:44743", "accept": false, - "reason": "Embedding dtype drift during evaluate and pipeline() translation messaging are unrelated." + "reason": "Pegasus v4/v5 output difference and qwen3_5 recurrent-state reset bug do not describe the same failure." }, { - "left": "issue:29942", - "right": "issue:32090", + "left": "issue:43262", + "right": "issue:43450", "accept": false, - "reason": "Flash Attention 2 test failure and Trainer _gpu_broadcast_one NoneType error are different runtime issues." + "reason": "Audio chat-template sampling-rate default and video processor batch-shape bug are different processor issues." }, { - "left": "issue:43611", - "right": "issue:43931", + "left": "issue:43278", + "right": "issue:43825", "accept": false, - "reason": "base_model_prefix loading breakage and Qwen3-VL weight-shape mismatch are distinct model-loading failures." + "reason": "Embedding dtype drift in evaluate vs pipeline translation messaging/docs are unrelated." }, { - "left": "issue:39692", - "right": "issue:43526", + "left": "issue:43606", + "right": "issue:45083", "accept": false, - "reason": "SigLIP2 docs errors and BeitImageProcessorFast reduce_labels bug are unrelated." + "reason": "CPU-offload device mismatch and qwen3_omni_moe feat-length helper behavior are different problems." }, { - "left": "issue:45200", - "right": "issue:45381", + "left": "issue:29942", + "right": "issue:32090", "accept": false, - "reason": "Gemma 4 mm_token_type_ids default issue is unrelated to Qwen2.5-VL vision_position_ids." + "reason": "Flash Attention 2 test failures and Trainer NoneType GPU broadcast errors are unrelated." }, { - "left": "issue:43335", - "right": "issue:43550", - "accept": false, - "reason": "Sparse-layer config bug and Bamba torch.compile+SDPA failure are different model-specific issues." + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "These are the same reported Qwen3.5 MoE from_pretrained error; 45478 appears to be a duplicate of 45310." }, { - "left": "issue:44855", - "right": "issue:45083", + "left": "issue:43611", + "right": "issue:43931", "accept": false, - "reason": "Python 3.13 IndentationError from torch.jit parsing and qwen3_omni_moe helper behavior are unrelated." + "reason": "base_model_prefix loading breakage and Qwen3-VL weight-shape mismatch are different load failures." }, { - "left": "issue:43404", - "right": "issue:45200", + "left": "issue:37428", + "right": "issue:44351", "accept": false, - "reason": "Mistral3 lm_head tying bug and Gemma 4 mm_token_type_ids default bug are unrelated generation/training issues." + "reason": "Different import errors affecting different symbols; not the same underlying change." }, { - "left": "issue:39692", + "left": "issue:43819", "right": "issue:44493", "accept": false, - "reason": "SigLIP2 documentation problems and unexpected position-id key reports are not the same bug." + "reason": "DAC STE mismatch and unexpected position-id keys are unrelated." }, { - "left": "issue:43450", + "left": "issue:39692", "right": "issue:43526", "accept": false, - "reason": "Video batched-shape bug and BeitImageProcessorFast label reduction bug are unrelated processor issues." + "reason": "SigLIP2 doc-example errors and BeitImageProcessorFast label reduction bug are unrelated." }, { - "left": "issue:38175", - "right": "issue:44493", + "left": "issue:43335", + "right": "issue:43550", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 and unexpected position-id keys are different symptoms with different causes." + "reason": "SwitchTransformers sparse-layer config bug and Bamba torch.compile SDPA failure are different models and code paths." }, { - "left": "issue:44977", - "right": "issue:45229", + "left": "issue:45200", + "right": "issue:45381", "accept": false, - "reason": "Qwen3.5 flash-attention generation bug and Gemma4 multi-GPU OOM are unrelated model/runtime problems." + "reason": "mm_token_type_ids defaulting for text-only fine-tuning and Qwen2.5-VL video position-id bug are not the same issue." }, { - "left": "issue:44448", - "right": "issue:44779", + "left": "issue:43450", + "right": "issue:43526", "accept": false, - "reason": "Pegasus output regression and Deepseek tokenizer regression are separate model-specific changes." + "reason": "Video batching shape bug and BeitImageProcessorFast label reduction bug are unrelated processor issues." }, { - "left": "issue:43335", - "right": "issue:43653", + "left": "issue:44855", + "right": "issue:45083", "accept": false, - "reason": "SwitchTransformers sparse-layer config and BigBirdTokenizer special-token decode are unrelated." + "reason": "Python 3.13 IndentationError in DebertaV2Model import and qwen3_omni_moe helper behavior are unrelated." }, { - "left": "issue:29127", - "right": "issue:32090", + "left": "issue:43404", + "right": "issue:45200", "accept": false, - "reason": "LayoutLMv3 error-message clarity and Trainer NoneType broadcast error are not the same bug." + "reason": "Mistral3 lm_head tie issue and Gemma 4 mm_token_type_ids defaulting are different generation/fine-tuning bugs." }, { - "left": "issue:43335", - "right": "issue:45083", + "left": "issue:44292", + "right": "issue:45520", "accept": false, - "reason": "SwitchTransformers sparse-layer config and qwen3_omni_moe helper behavior are unrelated." + "reason": "Qwen-3-8B-NVFP4 runtime error and Python 3.13 flash_attn import KeyError are different failures." }, { - "left": "issue:43335", - "right": "issue:43504", + "left": "issue:39692", + "right": "issue:44493", "accept": false, - "reason": "Sparse-layer creation bug and BEiT semantic-segmentation legacy-field load failure are different loading/config issues." + "reason": "SigLIP2 doc-example problems and unexpected position-id key behavior are unrelated." }, { - "left": "issue:43550", - "right": "issue:44112", + "left": "issue:44448", + "right": "issue:44779", "accept": false, - "reason": "torch.compile+SDPA failure in Bamba and GraniteSpeech CI device-override failure are unrelated." + "reason": "Pegasus output drift and Deepseek tokenizer regression are different model-specific bugs." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: the cluster is largely heterogeneous, and the 45310/45478 pair is a well-supported duplicate candidate based on the matching Qwen3.5 MoE from_pretrained failure and strong textual similarity. The other soft-edge rejections are conservative and appropriately distinguish different bug classes. Minor wording note: 'more complete report' is a bit stronger than the packet directly proves, but it doesn\u2019t undermine the overall assessment." + "feedback": "Grounded and conservative overall. The only accepted duplicate pair, issue:45310/issue:45478, is well supported by the identical Qwen3.5 MoE from_pretrained failure title/version. The remaining soft-edge rejects are appropriately cautious and do not overstate duplicate evidence." }, "evaluator_used": true, "retried": false, @@ -6656,15 +6783,17 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:25251", + "issue:29127", "issue:29942", "issue:32090", "issue:33357", "issue:34689", "issue:35141", "issue:37428", + "issue:38175", "issue:40444", "issue:41720", "issue:43065", @@ -6672,6 +6801,7 @@ "issue:43278", "issue:43329", "issue:43335", + "issue:43504", "issue:43526", "issue:43550", "issue:43606", @@ -6683,284 +6813,302 @@ "issue:43976", "issue:43994", "issue:44112", - "issue:44186", - "issue:44265", - "issue:44361", "issue:44448", + "issue:44451", "issue:44462", "issue:44466", "issue:44493", "issue:44554", "issue:44568", - "issue:44743", "issue:44857", "issue:44977", "issue:44987", "issue:45072", "issue:45083", - "issue:45198", "issue:45229", "issue:45372", "issue:45406", "issue:45412", "issue:45440", "issue:45459", - "issue:45491" + "issue:45491", + "issue:45520" ], "soft_pairs": [ + "issue:29127|issue:32090", + "issue:44977|issue:45229", + "issue:43335|issue:43653", + "issue:38175|issue:44493", + "issue:43335|issue:45083", + "issue:43335|issue:43504", + "issue:43550|issue:44112", + "issue:43335|issue:45072", "issue:44568|issue:45491", + "issue:43720|issue:44112", "issue:44448|issue:44977", "issue:40444|issue:43526", - "issue:43335|issue:45072", "issue:43723|issue:44554", - "issue:43720|issue:44112", - "issue:45372|issue:45406", - "issue:43066|issue:44462", "issue:43335|issue:44112", - "issue:37428|issue:44987", + "issue:34689|issue:35141", + "issue:45372|issue:45406", "issue:43606|issue:44857", - "issue:43873|issue:45229", "issue:44857|issue:45072", - "issue:34689|issue:35141", + "issue:43066|issue:44462", "issue:43873|issue:44493", + "issue:43873|issue:45229", + "issue:37428|issue:44987", "issue:43278|issue:43329", "issue:43701|issue:45083", - "issue:41720|issue:45229", - "issue:45372|issue:45459", "issue:29942|issue:35141", + "issue:41720|issue:45229", "issue:43550|issue:45083", + "issue:44451|issue:45520", "issue:43976|issue:45440", "issue:32090|issue:33357", + "issue:45372|issue:45459", "issue:44448|issue:44466", "issue:43065|issue:45412", "issue:43065|issue:43994", - "issue:44186|issue:45198", - "issue:43653|issue:44743", - "issue:43994|issue:45412", - "issue:25251|issue:32090", - "issue:44361|issue:45198", - "issue:44265|issue:45198" + "issue:25251|issue:32090" ], - "prepared_review_unit_hash": "78321a4f098046dfbe485fb96f79e154b396b5518cfd104b8f0d14a00e92bb82", + "prepared_review_unit_hash": "5ec1634ce9c4b0c07e981ab9c9c37ba317cbbeb35547b0e3aa7bd5954da83e8e", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37544, - "estimated_input_tokens": 9386, - "estimated_eval_tokens": 19028 + "soft_pair_count": 35, + "serialized_chars": 38376, + "estimated_input_tokens": 9594, + "estimated_eval_tokens": 19444 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is largely heterogeneous: most items are unrelated single bugs across different models, tokenizers, and runtime paths. I found no duplicate pair with enough evidence to merge, so the cluster does not have a clear canonical issue/PR anchor.", - "confidence": 0.88, - "canonical_issue_reason": "No single issue stands out as the canonical duplicate target; the items span unrelated subsystems and failure modes, so there is no meaningful shared root cause to anchor the cluster.", + "summary": "The set is mostly a grab bag of unrelated Transformer bugs across different models and subsystems. None of the soft pairs look like the same underlying issue, so all soft edges are rejected. Issue 43335 is the strongest representative because it is the most central item in the soft-link graph and describes a concrete, reproducible bug.", + "confidence": 0.72, + "canonical_issue_reason": "Issue 43335 is the most cluster-central item, appearing in multiple soft-similarity links, and it has a precise, actionable bug description about SwitchTransformers config behavior.", "canonical_pr_reason": null, - "best_issue_reason": "None of the issues is a strong global representative for this set. The closest to a broad anchor is one of the higher-level regression reports, but the collection is too disparate for a reliable canonical issue choice.", + "best_issue_reason": "Issue 43335 is the best representative of this set because it is the most connected and most clearly scoped bug among otherwise unrelated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44568", - "right": "issue:45491", + "left": "issue:29127", + "right": "issue:32090", "accept": false, - "reason": "Different tokenizer behavior vs. Gemma3 NaN embeddings; distinct models and code paths." + "reason": "Different failures: LayoutLMv3 error messaging vs Trainer GPU broadcast of NoneType." }, { - "left": "issue:44448", - "right": "issue:44977", + "left": "issue:44977", + "right": "issue:45229", "accept": false, - "reason": "Both mention generation/output problems, but they affect different models and likely different attention/decoding paths." + "reason": "Different problems: flash-attention generation bug vs multi-GPU OOM." }, { - "left": "issue:40444", - "right": "issue:43526", + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers config bug and BigBirdTokenizer special-token registration are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Siglip2 zero probabilities and position-id key warnings are different symptoms in different code paths." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Sparse-layer construction bug vs multimodal feature-length helper behavior; no shared underlying bug." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Different models and failure modes: SwitchTransformers config vs Beit legacy-field loading." + }, + { + "left": "issue:43550", + "right": "issue:44112", "accept": false, - "reason": "Unrelated bugs: multimodal dataset finetuning failure vs. BeitImageProcessorFast label reduction bug." + "reason": "Bamba torch.compile/SDPA issue is unrelated to a stale GraniteSpeech CI test." }, { "left": "issue:43335", "right": "issue:45072", "accept": false, - "reason": "Same broad model family name, but one is a sparse-layer config bug and the other is a dtype mismatch in inference." + "reason": "SwitchTransformers sparse-layer bug and bfloat16 dtype mismatch in CI are unrelated." }, { - "left": "issue:43723", - "right": "issue:44554", + "left": "issue:44568", + "right": "issue:45491", "accept": false, - "reason": "Tokenizer loading in v5 is unrelated to an MPS attention correctness issue." + "reason": "Tokenizer special-token addition bug is unrelated to Gemma3 NaN embeddings from sliding-window attention." }, { "left": "issue:43720", "right": "issue:44112", "accept": false, - "reason": "BitNet accelerate-loading bug and GraniteSpeech CI device-override test failure are unrelated." + "reason": "BitNet packed-weight loading issue is unrelated to GraniteSpeech device-override CI failure." }, { - "left": "issue:45372", - "right": "issue:45406", + "left": "issue:44448", + "right": "issue:44977", "accept": false, - "reason": "Both involve Gemma4, but one is a mistral_common import dependency issue and the other is a missing _tokenizer attribute in serve." + "reason": "Pegasus output differences across versions and Qwen3.5 flash-attention failure are distinct bugs." }, { - "left": "issue:43066", - "right": "issue:44462", + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Qwen2.5-VL multi-image dataset failure and BeitImageProcessorFast label reduction bug are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44554", "accept": false, - "reason": "Both concern tokenizer behavior, but one is wrong decoder type and the other is AutoTokenizer ignoring tokenizer.json." + "reason": "Tokenizer loading regression and MPS attention correctness issue do not share the same code path." }, { "left": "issue:43335", "right": "issue:44112", "accept": false, - "reason": "Different failures: SwitchTransformers config creation vs. GraniteSpeech CI test staleness." + "reason": "SwitchTransformers config bug and GraniteSpeech stale test failure are unrelated." }, { - "left": "issue:37428", - "right": "issue:44987", + "left": "issue:34689", + "right": "issue:35141", "accept": false, - "reason": "Flash-attention import error and a model-loading failure are not the same underlying bug." + "reason": "Model loading regression for Llama vision and token-embedding resizing/reinit bug are different issues." }, { - "left": "issue:43606", - "right": "issue:44857", + "left": "issue:45372", + "right": "issue:45406", "accept": false, - "reason": "CPU offload device mismatch and AMP/CUDA loss crash are unrelated runtime errors." + "reason": "Gemma4 processor import dependency failure and Gemma4Processor missing _tokenizer are separate loading bugs." }, { - "left": "issue:43873", - "right": "issue:45229", + "left": "issue:43606", + "right": "issue:44857", "accept": false, - "reason": "Both touch memory/offloading concerns, but one is quantization/offload behavior and the other is multi-GPU OOM on Gemma4." + "reason": "CPU-offload device mismatch and float16 AMP crash are different failure modes." }, { "left": "issue:44857", "right": "issue:45072", "accept": false, - "reason": "Different models and different dtype-related failures; no shared code-path evidence." + "reason": "Different bugs: LwDetr loss AMP crash vs dtype mismatch in SwitchTransformers/TimmWrapperModel inference." }, { - "left": "issue:34689", - "right": "issue:35141", + "left": "issue:43066", + "right": "issue:44462", "accept": false, - "reason": "Model-loading regression for a specific model vs. a general embedding-resize/post_init bug." + "reason": "Wrong tokenizer decoder type in v5 is unrelated to AutoTokenizer ignoring tokenizer.json." }, { "left": "issue:43873", "right": "issue:44493", "accept": false, - "reason": "Quantization/offloading behavior is unrelated to position-id key mismatches." + "reason": "Quantization/offloading behavior and unexpected position-id keys are unrelated." }, { - "left": "issue:43278", - "right": "issue:43329", + "left": "issue:43873", + "right": "issue:45229", "accept": false, - "reason": "BF16/FP32 dtype drift in evaluation is unrelated to the multimodal token-counting video-branch bug." + "reason": "Quantization offloading issue is not the same as Gemma4 multi-GPU OOM." }, { - "left": "issue:43701", - "right": "issue:45083", + "left": "issue:37428", + "right": "issue:44987", "accept": false, - "reason": "Checkpoint key mismatch and qwen3_omni_moe feature-length helper behavior are different problems." + "reason": "Flash-attention import error and loading a specific model in v5 are not the same bug." }, { - "left": "issue:41720", - "right": "issue:45229", + "left": "issue:43278", + "right": "issue:43329", "accept": false, - "reason": "Both are inference/runtime resource issues, but on different models with different failure modes (cudaErrorAssert vs. OOM)." + "reason": "Embedding dtype drift and multimodal token-count helper bug are unrelated." }, { - "left": "issue:45372", - "right": "issue:45459", + "left": "issue:43701", + "right": "issue:45083", "accept": false, - "reason": "Gemma4 processor import failure is unrelated to tokenizer error masking when protobuf is absent." + "reason": "Checkpoint resume key mismatch and qwen3_omni_moe length helper issue are different code paths." }, { "left": "issue:29942", "right": "issue:35141", "accept": false, - "reason": "FlashAttention test failures and embedding reinitialization on resize are separate regressions." + "reason": "Flash Attention 2 test failures are unrelated to output embedding reinitialization after resizing embeddings." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "cudaErrorAssert in Qwen3 auto device mapping is unrelated to Gemma4 OOM." }, { "left": "issue:43550", "right": "issue:45083", "accept": false, - "reason": "Torch.compile/SDPA failure in Bamba is unrelated to qwen3_omni_moe feature-length helper behavior." + "reason": "Bamba SDPA compile failure and qwen3_omni_moe helper behavior are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:45520", + "accept": false, + "reason": "ScandiBERT loading failure and flash_attn KeyError on Python 3.13 are different import/load issues." }, { "left": "issue:43976", "right": "issue:45440", "accept": false, - "reason": "Python version compatibility issue and DeepseekV3MoE divergence are unrelated." + "reason": "Python version compatibility bug and DeepseekV3MoE divergence are unrelated." }, { "left": "issue:32090", "right": "issue:33357", "accept": false, - "reason": "Trainer broadcast type error and MacOS bus error on CLIP are different runtime failures." + "reason": "Trainer NoneType broadcast error and MacOS bus error on CLIP are different runtime failures." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma4 processor import failure and tokenizer error masking when protobuf is absent are separate issues." }, { "left": "issue:44448", "right": "issue:44466", "accept": false, - "reason": "Both are v5 regressions, but one is generation output drift and the other is lm_head.weight serialization inconsistency." + "reason": "Pegasus output differences and lm_head.weight serialization inconsistency are not the same bug." }, { "left": "issue:43065", "right": "issue:45412", "accept": false, - "reason": "Dummy Conv2d in Sam3PixelDecoder and RT-DETR memory release are unrelated." + "reason": "Dummy Conv2d in Sam3PixelDecoder and RT-DETR memory leak are unrelated." }, { "left": "issue:43065", "right": "issue:43994", "accept": false, - "reason": "Sam3PixelDecoder dummy layer bug and SigLIP2 nonsensical outputs are not the same issue." - }, - { - "left": "issue:44186", - "right": "issue:45198", - "accept": false, - "reason": "Tokenizer crashes on NER/padding and Wav2Vec2 save/tokenization failures are unrelated tokenizer bugs." - }, - { - "left": "issue:43653", - "right": "issue:44743", - "accept": false, - "reason": "BigBirdTokenizer special-token registration and Qwen3 recurrent-state reset are different code paths." - }, - { - "left": "issue:43994", - "right": "issue:45412", - "accept": false, - "reason": "SigLIP2 output corruption and RT-DETR GC/memory-release behavior are unrelated." + "reason": "Sam3PixelDecoder dummy layer bug and Siglip2 nonsensical outputs are unrelated." }, { "left": "issue:25251", "right": "issue:32090", "accept": false, - "reason": "Pipeline top_k nesting issue and Trainer broadcast NoneType error are unrelated." - }, - { - "left": "issue:44361", - "right": "issue:45198", - "accept": false, - "reason": "MLukeTokenizer task AttributeError and Wav2Vec2 save/tokenization failure are distinct tokenizer bugs." - }, - { - "left": "issue:44265", - "right": "issue:45198", - "accept": false, - "reason": "torch.export export failure via torch_compilable_check is unrelated to Wav2Vec2 tokenization/save_pretrained problems." + "reason": "Pipeline top_k nesting bug and Trainer GPU broadcast NoneType are unrelated." } ] }, @@ -6982,7 +7130,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:29127", "issue:29942", @@ -6996,13 +7144,11 @@ "issue:39692", "issue:42915", "issue:42994", - "issue:43232", "issue:43257", "issue:43262", "issue:43278", "issue:43334", "issue:43381", - "issue:43388", "issue:43421", "issue:43504", "issue:43526", @@ -7010,22 +7156,24 @@ "issue:43653", "issue:43701", "issue:43756", - "issue:43761", "issue:43824", - "issue:43931", - "issue:44079", + "issue:43994", "issue:44186", "issue:44206", + "issue:44265", "issue:44279", + "issue:44361", "issue:44442", "issue:44464", "issue:44479", "issue:44610", + "issue:44743", "issue:44857", "issue:44871", "issue:44977", "issue:45072", "issue:45103", + "issue:45198", "issue:45229", "issue:45245", "issue:45372", @@ -7034,252 +7182,252 @@ "issue:45491" ], "soft_pairs": [ + "issue:44186|issue:45198", "issue:43278|issue:43381", - "issue:42915|issue:42994", + "issue:43653|issue:44743", + "issue:44265|issue:45198", + "issue:44361|issue:45198", "issue:43257|issue:43278", + "issue:42915|issue:42994", "issue:43421|issue:45491", + "issue:43994|issue:45412", "issue:44871|issue:44977", "issue:44206|issue:44479", "issue:43756|issue:45440", + "issue:43504|issue:44857", "issue:44610|issue:45412", - "issue:45103|issue:45372", "issue:29942|issue:30064", - "issue:43504|issue:44857", - "issue:43824|issue:44279", - "issue:45072|issue:45229", + "issue:45103|issue:45372", "issue:33357|issue:39692", "issue:29127|issue:34567", "issue:32090|issue:33290", - "issue:43334|issue:43531", + "issue:45072|issue:45229", + "issue:43824|issue:44279", + "issue:43381|issue:43701", "issue:43653|issue:44442", - "issue:44977|issue:45245", + "issue:43334|issue:43531", "issue:43526|issue:44186", "issue:44464|issue:45491", "issue:43262|issue:44206", - "issue:43381|issue:43701", - "issue:43761|issue:44079", + "issue:44977|issue:45245", "issue:29942|issue:36010", "issue:29127|issue:33357", - "issue:43232|issue:43388", - "issue:29127|issue:35141", - "issue:43526|issue:44857", - "issue:44442|issue:45072", - "issue:42915|issue:43931" + "issue:29127|issue:35141" ], - "prepared_review_unit_hash": "cf99a4226dd32fe4f67636b2ab3cdb83b06fe489eb5b26d70cc70f20e1e57944", + "prepared_review_unit_hash": "b83316dc1edb8d9e1d660aece1e253b350628850b2906f8d7734d2ca59b65348", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 31, - "serialized_chars": 37073, - "estimated_input_tokens": 9269, - "estimated_eval_tokens": 18794 + "serialized_chars": 37060, + "estimated_input_tokens": 9265, + "estimated_eval_tokens": 18786 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No duplicates accepted: the soft pairs span unrelated bugs, regressions, and feature requests across different models/subsystems, so this is not a coherent duplicate cluster.", + "summary": "This cluster is a loose semantic grab bag across unrelated bug reports (tokenizers, vision processors, training/runtime, generation, and model-loading issues). None of the soft pairs look like the same underlying defect, so all soft edges are rejected.", "confidence": 0.97, - "canonical_issue_reason": "No single canonical issue fits this set; the items are heterogeneous and do not describe the same underlying bug/change.", + "canonical_issue_reason": "No single canonical issue: the items span unrelated bugs in different subsystems, and the soft pairs do not read as true duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "No issue here is a good global representative because the cluster is broad and the paired items are not the same problem.", + "best_issue_reason": "No clear best issue representative for the same reason; choosing any one would be arbitrary across these unrelated reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different bugs: LayoutLMv2 NER/padding crash vs Wav2Vec2 save_pretrained/tokenization failure." + }, { "left": "issue:43278", "right": "issue:43381", "accept": false, - "reason": "Embedding dtype drift in evaluation is unrelated to gradient checkpointing being disallowed in eval mode." + "reason": "Embedding dtype regression and eval-mode gradient-checkpointing are unrelated failures." }, { - "left": "issue:42915", - "right": "issue:42994", + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBird tokenizer special-token decode issue vs Qwen3-MoE recurrent-state reset bug." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failure with torch_compilable_check is unrelated to Wav2Vec2 save/tokenization errors." + }, + { + "left": "issue:44361", + "right": "issue:45198", "accept": false, - "reason": "FP8 config failure for Qwen3Moe is a different problem from quantized model saving." + "reason": "MLuke tokenizer AttributeError is a different code path from Wav2Vec2 save_pretrained/tokenization." }, { "left": "issue:43257", "right": "issue:43278", "accept": false, - "reason": "Weight conversion under accelerate+deepspeed is unrelated to BF16/FP32 embedding dtype changes." + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is not the same as BF16-to-FP32 embedding dtype drift." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "FineGrainedFP8 failure on Qwen3Moe is distinct from quantized model saving." }, { "left": "issue:43421", "right": "issue:45491", "accept": false, - "reason": "Runtime special-token updates in tokenizers backend are unrelated to Gemma3 sliding-window NaN embeddings." + "reason": "Runtime special-token post-processor updates and Gemma3 NaN embeddings are different problems." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 inference mismatch/quantization issue is unrelated to RT-DETR memory release behavior." }, { "left": "issue:44871", "right": "issue:44977", "accept": false, - "reason": "Gemma eos_token_id inconsistency and Qwen3.5 flash-attention generation failures are different model/config issues." + "reason": "Gemma-3 eos_token_id inconsistency is separate from Qwen3.5 flash-attention generation issues." }, { "left": "issue:44206", "right": "issue:44479", "accept": false, - "reason": "Unsupported center arg in LasrFeatureExtractor is unrelated to the Qwen VL video regression." + "reason": "Unsupported center arg in LasrFeatureExtractor is unrelated to the Qwen video-input regression." }, { "left": "issue:43756", "right": "issue:45440", "accept": false, - "reason": "Smollm3 RoPE-layer omission and DeepSeekV3MoE divergence affect different models and code paths." + "reason": "Smollm3 RoPE layer omission and DeepseekV3 divergence are different model-specific bugs." }, { - "left": "issue:44610", - "right": "issue:45412", + "left": "issue:43504", + "right": "issue:44857", "accept": false, - "reason": "Processor image-size mismatch is unrelated to RT-DETR model memory not being released." + "reason": "BEiT legacy-field loading failure is unrelated to LwDetrImageLoss AMP/CUDA crashes." }, { - "left": "issue:45103", - "right": "issue:45372", + "left": "issue:44610", + "right": "issue:45412", "accept": false, - "reason": "auto_docstring annotation parsing crash is unrelated to the mistral_common import error in Gemma 4 loading." + "reason": "Processor size mismatch for OmDet-Turbo is not the same as RT-DETR memory leakage." }, { "left": "issue:29942", "right": "issue:30064", "accept": false, - "reason": "Flash Attention test failures do not match the void segmentation-map processing bug." - }, - { - "left": "issue:43504", - "right": "issue:44857", - "accept": false, - "reason": "Legacy-field model loading failure for BEiT is unrelated to LwDetrImageLoss AMP/CUDA crashes." - }, - { - "left": "issue:43824", - "right": "issue:44279", - "accept": false, - "reason": "Missing Qwen2_5_VL export is not the same as a generic dependency issue." + "reason": "Flash Attention 2 test failures and void segmentation-map processing are unrelated." }, { - "left": "issue:45072", - "right": "issue:45229", + "left": "issue:45103", + "right": "issue:45372", "accept": false, - "reason": "bfloat16 dtype mismatches in two model wrappers are unrelated to Gemma4 multi-GPU OOM." + "reason": "auto_docstring annotation handling and mistral_common import failures have different root causes." }, { "left": "issue:33357", "right": "issue:39692", "accept": false, - "reason": "MacOS bus error in CLIP loading and SigLIP2 doc-example errors are different problems." + "reason": "MacOS bus error for CLIP is unrelated to SigLIP2 documentation/model-processor mismatch." }, { "left": "issue:29127", "right": "issue:34567", "accept": false, - "reason": "LayoutLMv3 error-message clarity is unrelated to TrainerState token accounting." + "reason": "LayoutLMv3 error messaging and TrainerState token counter updating are different issues." }, { "left": "issue:32090", "right": "issue:33290", "accept": false, - "reason": "NoneType broadcast errors in Trainer are not the same as Adafactor+deepspeed OOM." + "reason": "NoneType broadcast in Trainer is unrelated to Adafactor/deepspeed OOM." }, { - "left": "issue:43334", - "right": "issue:43531", + "left": "issue:45072", + "right": "issue:45229", + "accept": false, + "reason": "bfloat16 inference dtype mismatches do not match a Gemma4 multi-GPU OOM." + }, + { + "left": "issue:43824", + "right": "issue:44279", "accept": false, - "reason": "Qwen3-VL pad_token_id loading failure and Qwen3-MoE sliding-window behavior are different bugs." + "reason": "Qwen2.5-VL import error is broader than the dependency issue and not the same concrete bug." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing in eval mode is unrelated to resume_from_checkpoint key mismatch." }, { "left": "issue:43653", "right": "issue:44442", "accept": false, - "reason": "BigBirdTokenizer special-token registration is unrelated to AutoTokenizer loading FastSpeech2ConformerTokenizer." + "reason": "BigBirdTokenizer decode output bug and AutoTokenizer loading FastSpeech2ConformerTokenizer are distinct." }, { - "left": "issue:44977", - "right": "issue:45245", + "left": "issue:43334", + "right": "issue:43531", "accept": false, - "reason": "Qwen3.5 flash-attention generation and the 2^24 category limit are unrelated." + "reason": "Qwen3-VL pad_token_id import failure and Qwen3-MoE sliding-window bug are different code paths." }, { "left": "issue:43526", "right": "issue:44186", "accept": false, - "reason": "BEiT fast processor reduce_labels bug is unrelated to LayoutLMv2 tokenizer padding/truncation crashes." + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to LayoutLMv2 tokenizer crashes." }, { "left": "issue:44464", "right": "issue:45491", "accept": false, - "reason": "Compiled-forward chunked generation inconsistency is not the same as Gemma3 NaN embeddings from mixed-length sliding windows." + "reason": "Chunked generation with compiled forward and Gemma3 sliding-window NaNs are different generation bugs." }, { "left": "issue:43262", "right": "issue:44206", "accept": false, - "reason": "Audio chat-template sampling-rate defaulting and LasrFeatureExtractor center-arg crashes are unrelated." - }, - { - "left": "issue:43381", - "right": "issue:43701", - "accept": false, - "reason": "Gradient checkpointing in eval mode and resume_from_checkpoint key mismatches are different trainer issues." + "reason": "Audio chat-template sampling-rate defaulting is unrelated to the LasrFeatureExtractor center-arg crash." }, { - "left": "issue:43761", - "right": "issue:44079", + "left": "issue:44977", + "right": "issue:45245", "accept": false, - "reason": "CLIPVisionModel hidden_states regression is not the same as ModelOutput key assignment when previous values were None." + "reason": "Qwen3.5 flash-attention generation and category-count overflow are unrelated." }, { "left": "issue:29942", "right": "issue:36010", "accept": false, - "reason": "Flash Attention 2 test failures do not match the GenerationMixin import error." + "reason": "Flash Attention 2 tests do not match the GenerationMixin import regression." }, { "left": "issue:29127", "right": "issue:33357", "accept": false, - "reason": "LayoutLMv3 error messaging and CLIP bus errors on MacOS are unrelated." - }, - { - "left": "issue:43232", - "right": "issue:43388", - "accept": false, - "reason": "Generation kwargs synchronization after sync_gpus is unrelated to gather_for_metrics dropping tuple labels." + "reason": "LayoutLMv3 error clarity and CLIP MacOS bus error are different artifacts." }, { "left": "issue:29127", "right": "issue:35141", "accept": false, - "reason": "LayoutLMv3 error messaging and embedding reinitialization in post_init are different bugs." - }, - { - "left": "issue:43526", - "right": "issue:44857", - "accept": false, - "reason": "BEiT reduce_labels output is unrelated to LwDetrImageLoss float16 CUDA crashes." - }, - { - "left": "issue:44442", - "right": "issue:45072", - "accept": false, - "reason": "Tokenizer loading failure for FastSpeech2ConformerTokenizer is unrelated to bfloat16 dtype mismatches in model wrappers." - }, - { - "left": "issue:42915", - "right": "issue:43931", - "accept": false, - "reason": "Qwen3Moe FP8 config failure and Qwen3-VL weight-shape mismatch are distinct Qwen-related issues, not the same bug." + "reason": "LayoutLMv3 error messaging and token-embedding reinitialization in post_init are unrelated." } ] }, @@ -7301,7 +7449,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:29942", "issue:33290", @@ -7311,39 +7459,40 @@ "issue:35141", "issue:39692", "issue:41628", + "issue:42915", + "issue:43232", "issue:43316", "issue:43381", + "issue:43388", "issue:43493", "issue:43525", "issue:43526", "issue:43575", "issue:43611", - "issue:43643", "issue:43646", "issue:43673", "issue:43701", "issue:43723", - "issue:43749", "issue:43761", "issue:43827", "issue:43873", "issue:43931", "issue:43994", "issue:44077", + "issue:44079", "issue:44188", - "issue:44190", "issue:44206", "issue:44387", "issue:44442", "issue:44479", - "issue:44496", + "issue:44610", "issue:44743", "issue:44811", - "issue:44871", + "issue:44857", "issue:44898", - "issue:44913", "issue:44964", "issue:45003", + "issue:45072", "issue:45216", "issue:45290", "issue:45372", @@ -7352,245 +7501,245 @@ "issue:45491" ], "soft_pairs": [ - "issue:43723|issue:45479", - "issue:43526|issue:43994", + "issue:43761|issue:44079", + "issue:44442|issue:45072", + "issue:42915|issue:43931", + "issue:43232|issue:43388", + "issue:43526|issue:44857", + "issue:43526|issue:44610", + "issue:44479|issue:45290", "issue:45003|issue:45440", + "issue:43723|issue:45479", "issue:44743|issue:44811", - "issue:44479|issue:45290", + "issue:43575|issue:43701", "issue:43316|issue:43525", - "issue:43761|issue:44188", - "issue:43994|issue:44206", "issue:43931|issue:45216", - "issue:43575|issue:43701", "issue:43646|issue:44077", + "issue:43761|issue:44188", "issue:44964|issue:45440", - "issue:39692|issue:45491", "issue:43575|issue:44387", - "issue:43673|issue:44743", + "issue:43994|issue:44206", "issue:29942|issue:33666", + "issue:39692|issue:45491", + "issue:43673|issue:44743", "issue:44442|issue:44898", "issue:33290|issue:34567", - "issue:41628|issue:45372", "issue:34689|issue:43611", "issue:43493|issue:43827", "issue:43381|issue:43873", "issue:44077|issue:44479", - "issue:39692|issue:44206", - "issue:33290|issue:35141", "issue:33290|issue:34689", - "issue:43873|issue:44871", - "issue:43749|issue:43931", - "issue:44190|issue:44496", - "issue:43643|issue:44913" + "issue:33290|issue:35141", + "issue:41628|issue:45372" ], - "prepared_review_unit_hash": "e376b2baee4b90e8a9615466587e182c2fab59c6677e9ac21ab99db65657072a", + "prepared_review_unit_hash": "14a1f2fbdb367af83375774584380cdb5083a00a41c9049d164d92c6add0ce5c", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 47, - "item_count": 47, + "node_count": 48, + "item_count": 48, "soft_pair_count": 30, - "serialized_chars": 36195, - "estimated_input_tokens": 9049, - "estimated_eval_tokens": 18354 + "serialized_chars": 36720, + "estimated_input_tokens": 9180, + "estimated_eval_tokens": 18616 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is highly heterogeneous: it mixes unrelated regressions in model loading, quantization, tokenizer/config APIs, docs, and attention behavior. I do not see any pair that is clearly the same underlying bug or change, so all soft edges are rejected.", - "confidence": 0.96, - "canonical_issue_reason": "No true canonical duplicate stands out because the items span unrelated issues; if forced to use an anchor, issue:43873 is the broadest and most discussion-heavy item, but it does not represent the rest of the set well.", + "summary": "This is a loose grab-bag of unrelated Transformers issues spanning model loading, generation, quantization, processors, docs, and training behavior. None of the soft pairs look like true duplicates or the same underlying fix, so all soft edges are rejected. The most umbrella-like issue is the generic Transformers 5.0.0 custom model initialization regression.", + "confidence": 0.27, + "canonical_issue_reason": "issue:43646 is the broadest, most umbrella-style report in the set and best fits as a weak cluster representative, even though the cluster items are not true duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "Issue:43873 is the strongest weak anchor by activity and inbound references, but it is still not a good representative of the cluster\u2019s mixed topics.", + "best_issue_reason": "issue:43646 is the least specific and most generally applicable issue here, so it is the best single issue to anchor the cluster despite weak semantic cohesion.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43723", - "right": "issue:45479", + "left": "issue:43761", + "right": "issue:44079", "accept": false, - "reason": "Tokenizer loading failure in v5 vs a classification loss bug; different APIs and failure modes." + "reason": "Both touch output/model-result handling, but one is a CLIPVisionModel regression and the other is a generic ModelOutput key assignment bug; different code paths." }, { - "left": "issue:43526", - "right": "issue:43994", + "left": "issue:44442", + "right": "issue:45072", "accept": false, - "reason": "BEiT label reduction bug vs SigLIP2 nonsensical outputs; unrelated models and code paths." + "reason": "Tokenizer loading for FastSpeech2Conformer is unrelated to dtype mismatch failures in SwitchTransformers/TimmWrapperModel." }, { - "left": "issue:45003", - "right": "issue:45440", + "left": "issue:42915", + "right": "issue:43931", "accept": false, - "reason": "Unsafe sys.modules access in modeling utils vs DeepSeekV3 implementation divergence; not the same bug." + "reason": "Both are Qwen-related loading problems, but one is a FineGrainedFP8 config failure and the other is a shape mismatch for a different model family; not the same bug." }, { - "left": "issue:44743", - "right": "issue:44811", + "left": "issue:43232", + "right": "issue:43388", "accept": false, - "reason": "Qwen3_5 cache/state reset bug vs Whisper batch_decode skip_special_tokens handling; unrelated behavior." + "reason": "Generation kwargs synchronization and metric gathering are unrelated subsystems with different failure modes." }, { - "left": "issue:44479", - "right": "issue:45290", + "left": "issue:43526", + "right": "issue:44857", "accept": false, - "reason": "Video-input regression in Qwen models vs chat-template tool-call crash; different features." + "reason": "Beit label reduction and LwDetr AMP/CUDA crashes are separate image-processing vs loss-computation issues." }, { - "left": "issue:43316", - "right": "issue:43525", + "left": "issue:43526", + "right": "issue:44610", "accept": false, - "reason": "Gemma3TextConfig API mismatch vs Llama4Config missing pad_token_id; different config issues." + "reason": "Both involve vision processors, but one is a label-reduction bug and the other is a processor/model size mismatch; different fixes." }, { - "left": "issue:43761", - "right": "issue:44188", + "left": "issue:44479", + "right": "issue:45290", "accept": false, - "reason": "CLIPVisionModel hidden_states regression vs attention-kernel divergence under torch.compile; different code paths." + "reason": "Video-input regression in Qwen VL models is unrelated to chat template crashes on tool-call assistant messages." }, { - "left": "issue:43994", - "right": "issue:44206", + "left": "issue:45003", + "right": "issue:45440", "accept": false, - "reason": "SigLIP2 incorrect outputs vs LasrFeatureExtractor center-arg crash; unrelated failures." + "reason": "Unsafe sys.modules access in modeling_utils is unrelated to DeepseekV3MoE implementation divergence." }, { - "left": "issue:43931", - "right": "issue:45216", + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading in v5 and zero-loss sequence classification with num_labels=1 are different bugs in different APIs." + }, + { + "left": "issue:44743", + "right": "issue:44811", "accept": false, - "reason": "Qwen3-VL weight-shape load mismatch vs Qwen3.5 save_pretrained checkpoint corruption; separate load/save bugs." + "reason": "Recurrent-state cache reset and Whisper batch_decode skip_special_tokens handling are unrelated." }, { "left": "issue:43575", "right": "issue:43701", "accept": false, - "reason": "TP loading OOM for Qwen2-57B vs resume_from_checkpoint key mismatch; unrelated issues." + "reason": "Tensor-parallel OOM when loading Qwen2-57B is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config API discrepancies, but they concern different models and different missing attributes." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Model weight-shape mismatch during load is not the same as an incorrect save_pretrained checkpoint regression." }, { "left": "issue:43646", "right": "issue:44077", "accept": false, - "reason": "Custom model init broken in v5 vs patchtsmixer post_init policy change; not the same concrete bug." + "reason": "Generic custom-model initialization breakage and patchtsmixer post_init policy are not the same concrete bug." }, { - "left": "issue:44964", - "right": "issue:45440", + "left": "issue:43761", + "right": "issue:44188", "accept": false, - "reason": "Phi-4 multimodal loading failure vs DeepSeekV3 implementation divergence; different models and causes." + "reason": "CLIP hidden_states regression and attention-kernel divergence under torch.compile are separate issues." }, { - "left": "issue:39692", - "right": "issue:45491", + "left": "issue:44964", + "right": "issue:45440", "accept": false, - "reason": "SigLIP2 docs/example errors vs Gemma3 NaN embeddings; documentation vs runtime kernel bug." + "reason": "Phi-4 multimodal loading failure and DeepSeekV3MoE divergence are unrelated model-specific problems." }, { "left": "issue:43575", "right": "issue:44387", "accept": false, - "reason": "Both are OOMs, but one is TP model load OOM and the other is int4 reserved-memory growth; distinct causes." + "reason": "OOM from tensor-parallel loading and increased reserved memory under int4 quantization are both memory-related but not the same bug." }, { - "left": "issue:43673", - "right": "issue:44743", + "left": "issue:43994", + "right": "issue:44206", "accept": false, - "reason": "Missing GenerationMixin cache in v5 chunked_prefill vs Qwen recurrent-state reset; both cache-related but different mechanisms." + "reason": "SigLIP2 nonsensical outputs and LasrFeatureExtractor center-arg crash are different failure modes." }, { "left": "issue:29942", "right": "issue:33666", "accept": false, - "reason": "Flash Attention 2 test failures vs Qwen2-VL multi-GPU training; no shared underlying bug." + "reason": "Flash Attention 2 test failures and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:45491", + "accept": false, + "reason": "SigLIP2 docs/example problems and Gemma3 mixed-length NaN embeddings are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache, but one is a GenerationMixin cache regression and the other is Qwen3_5 recurrent-state behavior; not the same fix." }, { "left": "issue:44442", "right": "issue:44898", "accept": false, - "reason": "AutoTokenizer loading failure vs Perceiver non-default-resolution image classification failure; unrelated subsystems." + "reason": "FastSpeech2Conformer tokenizer loading and Perceiver resolution handling are unrelated vision/audio issues." }, { "left": "issue:33290", "right": "issue:34567", "accept": false, - "reason": "Adafactor+DeepSpeed OOM vs TrainerState token counter not updating; unrelated bugs." - }, - { - "left": "issue:41628", - "right": "issue:45372", - "accept": false, - "reason": "Missing AutoImageProcessor export vs Gemma 4 processor import breakage from ReasoningEffort; both import-related but different symbols." + "reason": "Adafactor/deepspeed OOM and TrainerState token-count tracking are completely different bugs." }, { "left": "issue:34689", "right": "issue:43611", "accept": false, - "reason": "Llama 3.2 Vision model-loading regression vs base_model_prefix loading issue in v5; separate loading failures." + "reason": "Model loading breakage for a specific Llama vision model and a base_model_prefix regression are different loading paths and root causes." }, { "left": "issue:43493", "right": "issue:43827", "accept": false, - "reason": "SigLIP2 HF vs JAX discrepancy vs docs still mentioning pipeline(); one is model behavior, the other documentation." + "reason": "SigLIP2 implementation discrepancy and docs still referencing removed pipeline() APIs do not share a concrete code-path problem." }, { "left": "issue:43381", "right": "issue:43873", "accept": false, - "reason": "Gradient checkpointing in eval mode vs offloading/quantization bug; unrelated training/runtime behavior." + "reason": "Gradient checkpointing in eval mode and quantization offloading issues are unrelated training/runtime bugs." }, { "left": "issue:44077", "right": "issue:44479", "accept": false, - "reason": "PatchTSMixer post_init policy vs Qwen video-input regression; different models and failure modes." - }, - { - "left": "issue:39692", - "right": "issue:44206", - "accept": false, - "reason": "SigLIP2 docs/example problems vs feature-extractor crash; docs issue is not the same as runtime bug." - }, - { - "left": "issue:33290", - "right": "issue:35141", - "accept": false, - "reason": "OOM with Adafactor/DeepSpeed vs embedding reinitialization after resize; unrelated defects." + "reason": "patchtsmixer post_init policy and Qwen VL video-input regression are unrelated." }, { "left": "issue:33290", "right": "issue:34689", "accept": false, - "reason": "Optimizer OOM vs model-loading regression; no shared code-path problem." + "reason": "OOM in optimizer setup and model-loading failure for Llama vision are different failure classes." }, { - "left": "issue:43873", - "right": "issue:44871", - "accept": false, - "reason": "Quantization/offloading issue vs Gemma3 eos_token_id mismatch; different configuration areas." - }, - { - "left": "issue:43749", - "right": "issue:43931", - "accept": false, - "reason": "FSDP CPU RAM efficient loading broken vs Qwen3-VL shape mismatch; both loading-related, but not the same bug." - }, - { - "left": "issue:44190", - "right": "issue:44496", + "left": "issue:33290", + "right": "issue:35141", "accept": false, - "reason": "Local dataset loading failure in an example script vs unrecognized-model config error; unrelated." + "reason": "Optimizer OOM and embedding reinitialization in post_init are unrelated." }, { - "left": "issue:43643", - "right": "issue:44913", + "left": "issue:41628", + "right": "issue:45372", "accept": false, - "reason": "AutoConfig missing fields with trust_remote_code vs GPTNeoX rotary_pct not persisting; both config-related but different fields and paths." + "reason": "Import errors for AutoImageProcessor and ReasoningEffort are distinct symbols and contexts." } ] }, @@ -7612,7 +7761,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:16998", "issue:22355", @@ -7622,6 +7771,7 @@ "issue:34689", "issue:36010", "issue:36331", + "issue:39692", "issue:41950", "issue:42175", "issue:43065", @@ -7631,25 +7781,25 @@ "issue:43388", "issue:43404", "issue:43441", - "issue:43452", "issue:43475", "issue:43540", - "issue:43550", - "issue:43606", - "issue:43646", + "issue:43643", "issue:43688", - "issue:43701", "issue:43716", + "issue:43720", + "issue:43749", "issue:43761", "issue:43825", "issue:43856", - "issue:44077", + "issue:43873", + "issue:43931", "issue:44162", - "issue:44242", + "issue:44190", + "issue:44206", + "issue:44442", "issue:44479", "issue:44496", "issue:44514", - "issue:44534", "issue:44617", "issue:44734", "issue:44792", @@ -7663,217 +7813,210 @@ "issue:45406" ], "soft_pairs": [ - "issue:43540|issue:44514", + "issue:43720|issue:44442", + "issue:43749|issue:43931", + "issue:43643|issue:44913", + "issue:43873|issue:44871", + "issue:39692|issue:44206", + "issue:44190|issue:44496", + "issue:33290|issue:36331", + "issue:43381|issue:43856", "issue:43316|issue:44877", + "issue:44855|issue:45137", + "issue:43540|issue:44514", + "issue:34689|issue:36010", "issue:42175|issue:43316", "issue:44162|issue:44792", - "issue:43381|issue:43856", - "issue:44855|issue:45137", - "issue:33290|issue:36331", "issue:43388|issue:43688", - "issue:34689|issue:36010", - "issue:43316|issue:43716", "issue:43404|issue:43441", - "issue:43065|issue:44617", - "issue:41950|issue:43475", - "issue:45127|issue:45200", + "issue:43316|issue:43716", "issue:16998|issue:22355", "issue:25251|issue:28282", - "issue:43761|issue:44479", + "issue:41950|issue:43475", + "issue:43065|issue:44617", + "issue:45127|issue:45200", "issue:43295|issue:43825", - "issue:44734|issue:45406", - "issue:43295|issue:44871", - "issue:44077|issue:44534", - "issue:43550|issue:43701", - "issue:43452|issue:44496", - "issue:43606|issue:43701", - "issue:43646|issue:44913", - "issue:43761|issue:44242" + "issue:43761|issue:44479", + "issue:44734|issue:45406" ], - "prepared_review_unit_hash": "15f15873fe8843609aaa40fef60fddcc7b4d32086046f78faf4589709ca1a6ee", + "prepared_review_unit_hash": "330bd7d447d69758e69e984510d677329bdcaef5bf4574a575121d1ffe5242e7", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 26, - "serialized_chars": 34849, - "estimated_input_tokens": 8713, - "estimated_eval_tokens": 17682 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 25, + "serialized_chars": 35228, + "estimated_input_tokens": 8807, + "estimated_eval_tokens": 17870 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "All suggested soft edges look like false positives: the items share surface terms (processors, configs, video, loading, MoE, etc.) but describe different concrete bugs or regressions. No PRs are present, and the issue set is too heterogeneous to justify a single canonical duplicate target.", - "confidence": 0.96, - "canonical_issue_reason": "No clear canonical issue: the cluster spans unrelated subsystems and model families, so none of these issues is a credible duplicate target for the whole group.", + "summary": "This cluster is heterogeneous: the items cover unrelated bugs across model loading, training, pipelines, configs, and docs, so there is no credible duplicate canonical here. All soft-edge pairs should be rejected.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "No issue stands out as a global best representative because the cluster is heterogeneous and the proposed pairings do not share the same underlying bug.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43540", - "right": "issue:44514", + "left": "issue:43720", + "right": "issue:44442", "accept": false, - "reason": "Both mention video/processor handling, but one is a Qwen3OmniMoe video-input ValueError and the other is a Qwen2.5-VL batched chat-template padding crash; different code paths." + "reason": "Different failures: one is BitNet packed-weight unpacking during accelerate loading, the other is AutoTokenizer loading a FastSpeech2Conformer tokenizer." }, { - "left": "issue:43316", - "right": "issue:44877", + "left": "issue:43749", + "right": "issue:43931", "accept": false, - "reason": "Both are config/API issues, but Gemma3TextConfig discrepancy and granite_speech strict-loading failure are separate model/config problems." + "reason": "Different code paths and symptoms: FSDP CPU RAM-efficient loading vs a Qwen3-VL model weight-shape mismatch." }, { - "left": "issue:42175", - "right": "issue:43316", + "left": "issue:43643", + "right": "issue:44913", "accept": false, - "reason": "Extra/backend dependency packaging for torch is unrelated to a Gemma3TextConfig API mismatch." + "reason": "Both are config-related, but one is trust_remote_code returning incomplete AutoConfig fields and the other is GPTNeoX rotary_pct not persisting on reload." }, { - "left": "issue:44162", - "right": "issue:44792", + "left": "issue:43873", + "right": "issue:44871", "accept": false, - "reason": "ESM2 model breakage and a janus image-generation test failure are different bugs with different failure modes." + "reason": "Quantization/offloading behavior is unrelated to Gemma-3 eos_token_id configuration inconsistency." }, { - "left": "issue:43381", - "right": "issue:43856", + "left": "issue:39692", + "right": "issue:44206", "accept": false, - "reason": "Gradient checkpointing in eval mode is unrelated to MoE training memory usage." + "reason": "A documentation example with model/processor mismatch is not the same bug as LaserFeatureExtractor passing an unsupported center argument." }, { - "left": "issue:44855", - "right": "issue:45137", + "left": "issue:44190", + "right": "issue:44496", "accept": false, - "reason": "A Python 3.13 import/parser issue in DebertaV2Model is unrelated to a DeepSpeed ZeRO3 deque IndexError." + "reason": "Different subsystems: local dataset loading in an image-classification training script vs config loading failure for an unrecognized model." }, { "left": "issue:33290", "right": "issue:36331", "accept": false, - "reason": "Deepspeed+Adafactor OOM and a CustomTrainer signature mismatch are different training/runtime issues." - }, - { - "left": "issue:43388", - "right": "issue:43688", - "accept": false, - "reason": "Metric gathering dropping tuple labels is unrelated to incorrect auxiliary-loss normalization in MoE models." + "reason": "One is OOM with Adafactor under DeepSpeed; the other is a CustomTrainer method signature change." }, { - "left": "issue:34689", - "right": "issue:36010", + "left": "issue:43381", + "right": "issue:43856", "accept": false, - "reason": "Model-loading regression for Llama 3.2 Vision and an import-path error for GenerationMixin are different failures." + "reason": "Gradient checkpointing in eval mode and Qwen3 MoE memory usage are different training behaviors, not one underlying bug." }, { "left": "issue:43316", - "right": "issue:43716", + "right": "issue:44877", "accept": false, - "reason": "Gemma3TextConfig API mismatch and Mistral-3 image-preprocessor dtype mismatch are separate issues." + "reason": "Both mention config APIs, but the former is a Gemma3TextConfig discrepancy and the latter is strict config blocking granite_speech loading." }, { - "left": "issue:43404", - "right": "issue:43441", + "left": "issue:44855", + "right": "issue:45137", "accept": false, - "reason": "Missing lm_head tying in Mistral3ForConditionalGeneration is unrelated to FlashAttention failures in Ministral-3." + "reason": "Python 3.13 torch.jit parsing/IndentationError is unrelated to a DeepSpeed ZeRO3 deque pop error." }, { - "left": "issue:43065", - "right": "issue:44617", + "left": "issue:43540", + "right": "issue:44514", "accept": false, - "reason": "A dummy Conv2d in Sam3PixelDecoder and Sam3Video CUDA OOM are not the same bug." + "reason": "Different processor/input bugs: Qwen3OmniMoe video processing ValueError vs batched chat-template padding crash in Qwen2_5_VLProcessor." }, { - "left": "issue:41950", - "right": "issue:43475", + "left": "issue:34689", + "right": "issue:36010", "accept": false, - "reason": "Video-classification pipeline picking image processors is unrelated to Sam3VisionEncoderOutput missing fpn_position_embeddings." + "reason": "Model-loading regression for Llama 3.2 Vision is unrelated to the GenerationMixin import error." }, { - "left": "issue:45127", - "right": "issue:45200", + "left": "issue:42175", + "right": "issue:43316", "accept": false, - "reason": "LoRA merge collapse with tied embeddings is a different bug from missing default mm_token_type_ids in Gemma 4 fine-tuning." + "reason": "Package/backend installation issue with pip install '.[torch]' is unrelated to a Gemma3TextConfig API discrepancy." }, { - "left": "issue:16998", - "right": "issue:22355", + "left": "issue:44162", + "right": "issue:44792", "accept": false, - "reason": "DeBERTa-V3 model_max_length question and missing transformers.onnx import are unrelated." + "reason": "ESM2 being broken and a janus image-generation test failure are different models and failure modes." }, { - "left": "issue:25251", - "right": "issue:28282", + "left": "issue:43388", + "right": "issue:43688", "accept": false, - "reason": "Pipeline top_k output-shape change has nothing to do with missing PyTorch for AutoModel import." + "reason": "Gathering metrics drops tuple labels in the last batch, while auxiliary-loss normalization is a separate training-loss bug." }, { - "left": "issue:43761", - "right": "issue:44479", + "left": "issue:43404", + "right": "issue:43441", "accept": false, - "reason": "CLIPVisionModel hidden_states regression and Qwen video-input regression in v5.3.0 affect different models and paths." + "reason": "lm_head weight tying in Mistral3ForConditionalGeneration is a different concrete issue than FlashAttention failures in Ministral-3." }, { - "left": "issue:43295", - "right": "issue:43825", + "left": "issue:43316", + "right": "issue:43716", "accept": false, - "reason": "Processor.tokenizer/image-handling regression is not the same as a pipeline error-message regression about translation support." + "reason": "Gemma3TextConfig API mismatch is unrelated to the Mistral-3 image-preprocessor dtype mismatch." }, { - "left": "issue:44734", - "right": "issue:45406", + "left": "issue:16998", + "right": "issue:22355", "accept": false, - "reason": "KV-cache continuation indexing in transformers serve is unrelated to Gemma4Processor missing _tokenizer." + "reason": "A model_max_length question for DeBERTa-V3 is unrelated to a missing transformers.onnx import." }, { - "left": "issue:43295", - "right": "issue:44871", + "left": "issue:25251", + "right": "issue:28282", "accept": false, - "reason": "Custom processor/tokenizer regression and Gemma-3 eos_token_id inconsistency are distinct configuration/API problems." + "reason": "Pipeline top_k output shape change is unrelated to AutoModel requiring PyTorch to be installed." }, { - "left": "issue:44077", - "right": "issue:44534", + "left": "issue:41950", + "right": "issue:43475", "accept": false, - "reason": "Optional post_init validation for patchtsmixer is unrelated to non-persistent buffer initialization junk in v5." + "reason": "Video-classification pipeline using image processors is a different bug from Sam3VisionEncoderOutput missing fpn_position_embeddings." }, { - "left": "issue:43550", - "right": "issue:43701", + "left": "issue:43065", + "right": "issue:44617", "accept": false, - "reason": "torch.compile/SDPA failure in Bamba-9B-v2 is unrelated to resume_from_checkpoint key mismatch." + "reason": "Sam3PixelDecoder dummy Conv2d and Sam3Video CUDA OOM are both SAM3-related but not the same underlying change or code path." }, { - "left": "issue:43452", - "right": "issue:44496", + "left": "issue:45127", + "right": "issue:45200", "accept": false, - "reason": "gguf_file handling for from_pretrained is a different loading-path bug than an unrecognized-model config missing model_type." + "reason": "LoRA merge collapse with tied embeddings is unrelated to Gemma 4 mm_token_type_ids defaulting behavior." }, { - "left": "issue:43606", - "right": "issue:43701", + "left": "issue:43295", + "right": "issue:43825", "accept": false, - "reason": "CPU offload device mismatch and checkpoint resume key mismatch are unrelated." + "reason": "processor.tokenizer regression in v4.57.5 is unrelated to pipeline translation-task error messaging in v5." }, { - "left": "issue:43646", - "right": "issue:44913", + "left": "issue:43761", + "right": "issue:44479", "accept": false, - "reason": "Custom model initialization breakage in Transformers 5.0.0 is not the same as GPTNeoXConfig rotary_pct not round-tripping." + "reason": "CLIPVisionModel hidden_states regression and Qwen video-input regression are different models and different breakages." }, { - "left": "issue:43761", - "right": "issue:44242", + "left": "issue:44734", + "right": "issue:45406", "accept": false, - "reason": "Missing hidden_states in CLIPVisionModel and missing load-balancing loss when output_router_logits=False are different regressions." + "reason": "Transformers serve KV-cache continuation indexing is unrelated to Gemma4Processor missing _tokenizer." } ] }, @@ -7895,34 +8038,37 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:32090", "issue:35141", "issue:41720", "issue:42617", "issue:43065", - "issue:43335", + "issue:43295", "issue:43388", + "issue:43452", "issue:43525", "issue:43531", - "issue:43540", + "issue:43550", "issue:43575", - "issue:43582", - "issue:43644", + "issue:43606", + "issue:43646", "issue:43650", + "issue:43701", "issue:43742", "issue:43761", "issue:43827", "issue:43873", "issue:43906", "issue:43931", - "issue:43976", "issue:43994", "issue:44060", + "issue:44077", + "issue:44079", "issue:44190", "issue:44220", - "issue:44355", + "issue:44242", "issue:44464", "issue:44479", "issue:44496", @@ -7930,9 +8076,7 @@ "issue:44534", "issue:44610", "issue:44734", - "issue:44743", "issue:44792", - "issue:44855", "issue:44871", "issue:44913", "issue:44933", @@ -7943,228 +8087,227 @@ "issue:45290", "issue:45357", "issue:45362", - "issue:45372", "issue:45381" ], "soft_pairs": [ + "issue:44077|issue:44534", + "issue:43550|issue:43701", + "issue:43295|issue:44871", + "issue:43452|issue:44496", + "issue:43606|issue:43701", + "issue:43646|issue:44913", "issue:43906|issue:45290", - "issue:43931|issue:44060", + "issue:43761|issue:44242", "issue:42617|issue:43575", "issue:43575|issue:43742", + "issue:43931|issue:44060", "issue:44521|issue:45381", "issue:43525|issue:43531", - "issue:44479|issue:44734", - "issue:44792|issue:44977", - "issue:43065|issue:44610", "issue:43575|issue:43873", - "issue:44933|issue:45362", + "issue:44792|issue:44977", + "issue:44479|issue:44734", "issue:44464|issue:44871", "issue:32090|issue:35141", - "issue:41720|issue:45245", + "issue:44933|issue:45362", + "issue:43065|issue:44610", "issue:43827|issue:43994", "issue:44534|issue:44913", - "issue:43388|issue:43582", + "issue:41720|issue:45245", "issue:43650|issue:44220", "issue:44190|issue:44964", - "issue:43761|issue:44743", - "issue:45127|issue:45357", - "issue:44355|issue:44855", - "issue:43540|issue:44734", - "issue:44496|issue:45372", - "issue:43976|issue:44855", - "issue:43644|issue:44913", - "issue:43335|issue:44855" + "issue:43388|issue:44079", + "issue:45127|issue:45357" ], - "prepared_review_unit_hash": "e0a13e0abcdf21b36cdac7a10d3eda22c186227887863b2ec0db5f50fa3d7f52", + "prepared_review_unit_hash": "e2cc9d58ad9b1ebbc62d4188589a55c7f308e44ad4dfd28e28c10ed06ecba4c7", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 27, - "serialized_chars": 35807, - "estimated_input_tokens": 8952, - "estimated_eval_tokens": 18160 + "serialized_chars": 35662, + "estimated_input_tokens": 8916, + "estimated_eval_tokens": 18088 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The set is very heterogeneous: most items are unrelated model/config/runtime bugs, so none of the soft edges look like true duplicates. No PRs are present.", - "confidence": 0.89, - "canonical_issue_reason": "issue:44521 is the strongest standalone bug report in the set: open, well-discussed, with inbound references, and a concrete reproducible multimodal regression.", + "summary": "These items are mostly unrelated issue reports across different regressions and subsystems. None of the soft pairs look like the same concrete bug, so I would not merge any of them as duplicates. If I had to anchor the cluster, I\u2019d use issue 44521 because it is current, well-documented, and has the strongest external signal.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44521 is the strongest representative: it is open, detailed, and has multiple inbound references, making it the most useful anchor among these unrelated issues.", "canonical_pr_reason": null, - "best_issue_reason": "issue:44521 is the best representative issue to anchor the cluster because it is high-signal, actionable, and more substantial than the many isolated model-specific reports.", + "best_issue_reason": "issue:44521 is the best global issue candidate for triage because it has the clearest repro and the most activity/references.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43906", - "right": "issue:45290", + "left": "issue:44077", + "right": "issue:44534", "accept": false, - "reason": "Different bugs: an isolated reproduction of issue #38071 versus a chat-template crash on assistant tool-call messages." + "reason": "Both involve model init/buffer behavior in v5, but one is about post_init being disallowed for patchtsmixer and the other about junk non-persistent buffers; different bugs." }, { - "left": "issue:43931", - "right": "issue:44060", + "left": "issue:43550", + "right": "issue:43701", "accept": false, - "reason": "Different model families and failure modes: load-time weight shape mismatch vs a tied-weights warning mapping the wrong parameters." + "reason": "One is a torch.compile/SDPA failure for Bamba, the other is a resume_from_checkpoint key mismatch; no shared code-path bug." }, { - "left": "issue:42617", - "right": "issue:43575", + "left": "issue:43295", + "right": "issue:44871", "accept": false, - "reason": "Different surfaces: a 3d_parallel.py runtime failure versus Qwen2-57B tensor-parallel OOM on model load." + "reason": "Custom processor/tokenizer regression and Gemma eos_token_id inconsistency are distinct model/config issues." }, { - "left": "issue:43575", - "right": "issue:43742", + "left": "issue:43452", + "right": "issue:44496", "accept": false, - "reason": "Unrelated models and errors: tensor-parallel OOM for Qwen2-57B vs a key error loading MobileLLM-125M." + "reason": "Both are loading-related, but gguf_file handling and missing model_type validation are different root causes." }, { - "left": "issue:44521", - "right": "issue:45381", + "left": "issue:43606", + "right": "issue:43701", "accept": false, - "reason": "Both are multimodal, but the reported bugs differ: assistant_masks become all-zero vs video vision_position_ids are wrong." + "reason": "CPU offload device mismatch is unrelated to resume checkpoint key mismatch." }, { - "left": "issue:43525", - "right": "issue:43531", + "left": "issue:43646", + "right": "issue:44913", "accept": false, - "reason": "Different configs/models: missing pad_token_id on Llama4Config vs a sliding_window bug in Qwen3-MoE." + "reason": "Generic custom initialization breakage and GPTNeoX rotary_pct reload regression are separate problems." }, { - "left": "issue:44479", - "right": "issue:44734", + "left": "issue:43906", + "right": "issue:45290", "accept": false, - "reason": "Different code paths: a video-input regression in Qwen models vs KV-cache continuation tensor indexing in transformers serve." + "reason": "An isolated reproduction of one past issue does not match a chat-template tool-call crash; different functionality and failure mode." }, { - "left": "issue:44792", - "right": "issue:44977", + "left": "issue:43761", + "right": "issue:44242", "accept": false, - "reason": "Different model-specific generation problems: janus image-generation test failure vs Qwen3.5 flash-attention generation issue." + "reason": "CLIP hidden_states regression and MoE router loss omission are unrelated model-output vs training-loss bugs." }, { - "left": "issue:43065", - "right": "issue:44610", + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel startup failure and tensor-parallel OOM are both parallelism-related but not the same concrete bug." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "OOM on a large Qwen2 model with tp is not the same as a key error loading MobileLLM-125M." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Shape mismatch loading Qwen3-VL differs from a tied-weights warning in Qwen3-Next." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both touch multimodal chat/vision input handling, but assistant mask generation and video vision_position_ids are different code paths and symptoms." + }, + { + "left": "issue:43525", + "right": "issue:43531", "accept": false, - "reason": "Different components: a dummy Conv2d in Sam3PixelDecoder vs an OmDet-Turbo processor/model input-size mismatch." + "reason": "Missing pad_token_id in Llama4Config and Qwen3-MoE sliding_window behavior are unrelated config bugs." }, { "left": "issue:43575", "right": "issue:43873", "accept": false, - "reason": "Different resource bugs: OOM when loading a Qwen model with TP vs offloading failing under quantization." + "reason": "Tensor-parallel OOM and offloading/quantization behavior are different memory-management issues." }, { - "left": "issue:44933", - "right": "issue:45362", + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation issues are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44734", "accept": false, - "reason": "Different failures: missing import from image_utils vs a Qwen3.5 chat crash." + "reason": "Video-input regression in Qwen VL models and serve KV-cache continuation indexing are different failures." }, { "left": "issue:44464", "right": "issue:44871", "accept": false, - "reason": "Different bugs: inconsistent outputs with compiled forward vs an eos_token_id config mismatch." + "reason": "Compiled chunked generation inconsistency is not the same as a Gemma eos_token_id config mismatch." }, { "left": "issue:32090", "right": "issue:35141", "accept": false, - "reason": "Different issues: Trainer broadcast failure on NoneType vs embedding reinitialization after resizing with untied embeddings." + "reason": "Trainer GPU broadcast NoneType error and embedding reinitialization after resize are different bugs." }, { - "left": "issue:41720", - "right": "issue:45245", + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A missing import in image_utils and a Qwen3.5 chat crash are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:44610", "accept": false, - "reason": "Different root causes: device-mapping cudaErrorAssert on A800 vs an upper bound on category count." + "reason": "Sam3PixelDecoder dummy Conv2d and OmDet-Turbo processor image-size mismatch are different vision-model issues." }, { "left": "issue:43827", "right": "issue:43994", "accept": false, - "reason": "Docs referencing removed pipeline() is unrelated to SIGLIP producing nonsensical outputs." + "reason": "Documentation still referencing pipeline() is unrelated to SigLIP2 producing bad outputs with AutoModel/pipeline." }, { "left": "issue:44534", "right": "issue:44913", "accept": false, - "reason": "Both involve persistence, but one is junk-filled non-persistent buffers and the other is config reload losing rotary_pct." + "reason": "Non-persistent buffer junk and GPTNeoX rotary_pct not persisting are both v5 regressions, but not the same underlying serialization bug." }, { - "left": "issue:43388", - "right": "issue:43582", + "left": "issue:41720", + "right": "issue:45245", "accept": false, - "reason": "Different areas: gather_for_metrics drops tuple labels vs caching_allocator_warmup raises a TypeError on Apple Silicon." + "reason": "Qwen3 auto device mapping cuda assert and category-count overflow are unrelated runtime errors." }, { "left": "issue:43650", "right": "issue:44220", "accept": false, - "reason": "Clearly unrelated: a placeholder issue title vs an fbank feature extraction bug." + "reason": "A vague data request is not the same as a fbank feature extraction bug." }, { "left": "issue:44190", "right": "issue:44964", "accept": false, - "reason": "Different loading paths: local dataset loading in a no-trainer script vs model config loading for Phi-4 multimodal." + "reason": "Local dataset loading in a training script and loading a multimodal model are different loading paths and failures." }, { - "left": "issue:43761", - "right": "issue:44743", + "left": "issue:43388", + "right": "issue:44079", "accept": false, - "reason": "Different model families and symptoms: CLIPVisionModel hidden_states regression vs Qwen recurrent states resetting with cache." + "reason": "Dropping label elements in gather_for_metrics and ModelOutput key assignment with None are distinct tuple/ModelOutput handling issues." }, { "left": "issue:45127", "right": "issue:45357", "accept": false, - "reason": "Both touch weights/save behavior, but one is LoRA merge collapse with extended vocab and the other is incorrect visual-encoder key serialization." - }, - { - "left": "issue:44355", - "right": "issue:44855", - "accept": false, - "reason": "Different failure layers: compiled Python file errors vs a Python 3.13 torch.jit parsing/IndentationError in DebertaV2Model." - }, - { - "left": "issue:43540", - "right": "issue:44734", - "accept": false, - "reason": "Different subsystems: Qwen3OmniMoe video-input ValueError vs serving-time KV-cache tensor indexing crash." - }, - { - "left": "issue:44496", - "right": "issue:45372", - "accept": false, - "reason": "Different models and causes: unrecognized OLMo config vs Gemma 4 processor import failure from missing ReasoningEffort." - }, - { - "left": "issue:43976", - "right": "issue:44855", - "accept": false, - "reason": "Not the same bug: a broader Python-version compatibility complaint vs a specific DebertaV2 torch.jit parsing error." - }, - { - "left": "issue:43644", - "right": "issue:44913", - "accept": false, - "reason": "Different persistence bugs: junk-filled non-persistent buffers vs rotary_pct not surviving config reload." - }, - { - "left": "issue:43335", - "right": "issue:44855", - "accept": false, - "reason": "Unrelated: sparse-layer construction in SwitchTransformers vs a Python 3.13 torch.jit indentation/parsing failure." + "reason": "LoRA merge with extended vocabulary causing collapse and incorrect visual encoder keys in save_pretrained are separate model-saving bugs." } ] }, @@ -8186,7 +8329,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:36246", "issue:38175", @@ -8195,273 +8338,287 @@ "issue:41628", "issue:41950", "issue:42175", - "issue:42907", "issue:43065", "issue:43116", "issue:43295", "issue:43299", + "issue:43335", "issue:43404", - "issue:43441", "issue:43479", + "issue:43540", "issue:43550", "issue:43643", - "issue:43673", + "issue:43644", + "issue:43761", "issue:43854", "issue:43976", "issue:44016", - "issue:44062", "issue:44077", + "issue:44079", "issue:44112", "issue:44186", "issue:44190", "issue:44220", "issue:44246", "issue:44279", - "issue:44336", + "issue:44355", "issue:44466", "issue:44496", "issue:44683", "issue:44734", + "issue:44743", "issue:44792", + "issue:44855", "issue:44871", - "issue:44877", "issue:44913", - "issue:44964", - "issue:44987", "issue:45003", "issue:45216", - "issue:45245", "issue:45276", "issue:45335", + "issue:45372", "issue:45375", "issue:45446" ], "soft_pairs": [ - "issue:44077|issue:44877", + "issue:43761|issue:44743", + "issue:44355|issue:44855", + "issue:43644|issue:44913", + "issue:44496|issue:45372", + "issue:43976|issue:44855", + "issue:43540|issue:44734", + "issue:44077|issue:44079", "issue:43550|issue:44190", - "issue:43116|issue:44190", "issue:43295|issue:45276", + "issue:43335|issue:44855", + "issue:43404|issue:45216", + "issue:43116|issue:44190", "issue:41950|issue:44016", "issue:36246|issue:41628", - "issue:43404|issue:45216", - "issue:36246|issue:38617", "issue:39692|issue:44190", + "issue:43644|issue:44871", "issue:44186|issue:44734", - "issue:44871|issue:45245", - "issue:38175|issue:43065", - "issue:44016|issue:44220", "issue:43479|issue:45375", + "issue:36246|issue:38617", + "issue:44016|issue:44220", "issue:44246|issue:45003", - "issue:42175|issue:44279", - "issue:44683|issue:45446", + "issue:38175|issue:43065", "issue:44466|issue:44913", + "issue:44683|issue:45446", + "issue:42175|issue:44279", "issue:44112|issue:44792", "issue:43854|issue:44792", "issue:43295|issue:45335", - "issue:42907|issue:43441", - "issue:43299|issue:45216", - "issue:44016|issue:44062", "issue:43643|issue:44496", - "issue:44112|issue:44964", - "issue:43673|issue:44336", - "issue:43976|issue:44987" + "issue:43299|issue:45216" ], - "prepared_review_unit_hash": "1961b55963e42d8760b594bd23656b1f2a20c392f0416cc99403b88e59533365", + "prepared_review_unit_hash": "64143b45b764f86bec0d77ee3c552500287cec0ca90c1dfadde5fd942773908b", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 47, "item_count": 47, - "soft_pair_count": 28, - "serialized_chars": 35560, - "estimated_input_tokens": 8890, - "estimated_eval_tokens": 18036 + "soft_pair_count": 30, + "serialized_chars": 36364, + "estimated_input_tokens": 9091, + "estimated_eval_tokens": 18438 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is a loose mix of unrelated Transformers issues (imports, config loading, model-specific regressions, pipelines, docs/tests). None of the soft pairs look like the same underlying bug or mergeable PR change.", - "confidence": 0.96, - "canonical_issue_reason": null, + "summary": "This set is a loose collection of unrelated Transformers issues; none of the soft-edge pairs look like the same underlying bug or change, so all are rejected. The best anchor issue is the broad v5 regression about non-persistent buffers being filled incorrectly, which is the closest to the recurring serialization/regression theme in the set.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43644 is the strongest representative of the cluster\u2019s repeated v5/regression/serialization theme and is broad enough to serve as an anchor, even though the other items are not true duplicates of it.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:43644 has the most general core-runtime impact and is more central than the narrower model- or example-specific reports elsewhere in the set.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44077", - "right": "issue:44877", + "left": "issue:43761", + "right": "issue:44743", "accept": false, - "reason": "Both involve config/model initialization, but one is about `patchtsmixer` optional `post_init` and the other is strict loading of `granite_speech`; different failure modes and code paths." + "reason": "Different models and different failures: CLIPVisionModel hidden_states regression vs Qwen3.5 recurrent-state reset with cache." }, { - "left": "issue:43550", - "right": "issue:44190", + "left": "issue:44355", + "right": "issue:44855", "accept": false, - "reason": "`torch.compile`/SDPA failure in Bamba-9B-v2 is unrelated to a local dataset loading problem in an image classification example." + "reason": "Both are Python/runtime-related, but one is about compiled Python files and the other is a DebertaV2 import parsing issue on Python 3.13; not the same bug." }, { - "left": "issue:43116", + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Both involve reload/serialization behavior, but one is about non-persistent buffers getting junk values and the other is GPTNeoX rotary_pct not persisting; different code paths." + }, + { + "left": "issue:44496", + "right": "issue:45372", + "accept": false, + "reason": "Both are model-loading/import problems, but they concern different missing symbols and different models; not the same concrete issue." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "General Python-version compatibility complaints are not enough; these point to different failure modes and different code paths." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input validation and transformers serve KV-cache indexing are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44079", + "accept": false, + "reason": "One is a config/post_init rule for patchtsmixer, the other is ModelOutput key assignment logic; different bugs." + }, + { + "left": "issue:43550", "right": "issue:44190", "accept": false, - "reason": "Empty multi-label classification outputs in `run_classification.py` is a different script/model behavior than local dataset loading in `run_image_classification_no_trainer.py`." + "reason": "torch.compile+SDPA failure in Bamba is unrelated to the image-classification no-trainer local dataset loading issue." }, { "left": "issue:43295", "right": "issue:45276", "accept": false, - "reason": "Both mention embedding/token handling, but one is a regression in processor/tokenizer image passing and the other is `resize_token_embeddings` not updating Gemma4-specific embeddings." + "reason": "Processor/tokenizer regression and resize_token_embeddings propagation are different problems." }, { - "left": "issue:41950", - "right": "issue:44016", + "left": "issue:43335", + "right": "issue:44855", "accept": false, - "reason": "Video-classification pipeline processor lookup is unrelated to a notebook syntax error." + "reason": "SwitchTransformers sparse-layer config bug is unrelated to the Python 3.13 import parsing error." }, { - "left": "issue:36246", - "right": "issue:41628", + "left": "issue:43404", + "right": "issue:45216", "accept": false, - "reason": "Both are import errors, but they concern different missing symbols (`Qwen2_5_VLImageProcessor` vs `AutoImageProcessor`) and likely different export issues." + "reason": "Both mention weight tying/checkpointing, but they affect different models and different failure modes; not mergeable as one bug." }, { - "left": "issue:43404", - "right": "issue:45216", + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Example script multi-label output bug vs local dataset loading failure in a different script." + }, + { + "left": "issue:41950", + "right": "issue:44016", "accept": false, - "reason": "One is tied `lm_head.weight` in Mistral3 generation; the other is incorrect saved checkpoints for Qwen3.5. Different models and different serialization paths." + "reason": "Video-classification pipeline processor lookup and notebook syntax error are unrelated." }, { "left": "issue:36246", - "right": "issue:38617", + "right": "issue:41628", "accept": false, - "reason": "Different missing imports from different modules; no shared concrete bug beyond being import errors." + "reason": "Both are import errors around image processor APIs, but they are different missing names and not the same underlying breakage." }, { "left": "issue:39692", "right": "issue:44190", "accept": false, - "reason": "SigLIP2 doc/example errors are about model/processor mismatch and quantization; local dataset loading is a separate training-script issue." + "reason": "Docs example errors for SigLIP2 are unrelated to the no-trainer image-classification dataset issue." + }, + { + "left": "issue:43644", + "right": "issue:44871", + "accept": false, + "reason": "Buffer serialization corruption and Gemma3 eos_token_id mismatch are different issues." }, { "left": "issue:44186", "right": "issue:44734", "accept": false, - "reason": "LayoutLMv2 tokenizer crashes on NER padding/truncation is unrelated to `transformers serve` KV-cache continuation indexing." + "reason": "LayoutLMv2Tokenizer NER/padding crash is unrelated to serve KV-cache continuation indexing." }, { - "left": "issue:44871", - "right": "issue:45245", + "left": "issue:43479", + "right": "issue:45375", "accept": false, - "reason": "Gemma eos token config inconsistency is unrelated to the category-cardinality runtime error." + "reason": "Different multimodal config bugs in different models; not the same defect." }, { - "left": "issue:38175", - "right": "issue:43065", + "left": "issue:36246", + "right": "issue:38617", "accept": false, - "reason": "SigLIP2 zero probabilities and a dummy `nn.Conv2d` in Sam3PixelDecoder are different model bugs." + "reason": "Qwen2_5_VLImageProcessor import error and layer_type_validation import error are unrelated import failures." }, { "left": "issue:44016", "right": "issue:44220", "accept": false, - "reason": "Notebook syntax error and `_torch_extract_fbank_features()` failure are unrelated." + "reason": "Notebook syntax error is unrelated to _torch_extract_fbank_features() behavior." }, { - "left": "issue:43479", - "right": "issue:45375", + "left": "issue:44246", + "right": "issue:45003", "accept": false, - "reason": "Phi4MultimodalConfig default-init bug is not the same as Qwen3_5MoeVisionConfig dropping a missing field under `@strict`." + "reason": "Import slowdown and unsafe sys.modules access are different import-time issues." }, { - "left": "issue:44246", - "right": "issue:45003", + "left": "issue:38175", + "right": "issue:43065", "accept": false, - "reason": "Import-time slowness and unsafe `sys.modules[]` access are related to imports, but not the same concrete bug." + "reason": "SigLIP2 zero probabilities and Sam3PixelDecoder dummy Conv2d are unrelated model-specific bugs." }, { - "left": "issue:42175", - "right": "issue:44279", + "left": "issue:44466", + "right": "issue:44913", "accept": false, - "reason": "TensorFlow missing from a `.[torch]` install is a dependency packaging issue, not the same as a generic transformers dependency problem." + "reason": "Device-dependent lm_head serialization inconsistency and rotary_pct reload regression are different serialization issues." }, { "left": "issue:44683", "right": "issue:45446", "accept": false, - "reason": "Compiled flex_attention on torch >= 2.9 and a PyTorch version check for AuxRequest import are separate version-compatibility bugs." + "reason": "Compiled flex_attention failure on torch>=2.9 and a wrong version check for AuxRequest import are related only by subsystem, not the same bug." }, { - "left": "issue:44466", - "right": "issue:44913", + "left": "issue:42175", + "right": "issue:44279", "accept": false, - "reason": "Both mention config/weights behavior, but one is tied-weight serialization and the other is `rotary_pct` not persisting in GPTNeoX config." + "reason": "Package extras/backend dependency issue is not the same as a generic dependency problem report." }, { "left": "issue:44112", "right": "issue:44792", "accept": false, - "reason": "CI stale device override failure in GraniteSpeech is unrelated to a janus image-generation test failure." + "reason": "CI stale test failure in GraniteSpeech and janus image-generation test failure are unrelated." }, { "left": "issue:43854", "right": "issue:44792", "accept": false, - "reason": "GLM-4.7-Flash unit-test loading failure and janus image-generation test failure are different models and different test paths." + "reason": "Unable to load a specific model in unit tests is not the same as a failed janus image-generation test." }, { "left": "issue:43295", "right": "issue:45335", "accept": false, - "reason": "Both involve embedding resizing, but the broken code paths are different model families (`processor.tokenizer` regression vs T5Gemma decoder embeddings)." - }, - { - "left": "issue:42907", - "right": "issue:43441", - "accept": false, - "reason": "Saving dequantized Ministral/Devstral models is not the same bug as FlashAttention failure in Ministral-3 loading." - }, - { - "left": "issue:43299", - "right": "issue:45216", - "accept": false, - "reason": "Both are Qwen3/Qwen3.5 regressions, but one is load-time model loading and the other is save_pretrained checkpoint correctness; not the same bug." - }, - { - "left": "issue:44016", - "right": "issue:44062", - "accept": false, - "reason": "Notebook syntax error and `AddedToken(..., special=...)` TypeError are unrelated." + "reason": "Different resize_token_embeddings propagation bugs in different models/code paths." }, { "left": "issue:43643", "right": "issue:44496", "accept": false, - "reason": "`trust_remote_code=True` dropping fields in `AutoConfig` is unrelated to an unrecognized model/config lacking `model_type`." - }, - { - "left": "issue:44112", - "right": "issue:44964", - "accept": false, - "reason": "Different models and failure contexts: CI test flake in GraniteSpeech vs loading Phi-4 multimodal with latest transformers." - }, - { - "left": "issue:43673", - "right": "issue:44336", - "accept": false, - "reason": "Generation cache missing in v5 chunked prefill is unrelated to ANSI codes emitted by `loading_report`." + "reason": "trust_remote_code missing fields and unrecognized model_type are both config-loading issues, but they are distinct root causes." }, { - "left": "issue:43976", - "right": "issue:44987", + "left": "issue:43299", + "right": "issue:45216", "accept": false, - "reason": "Python version compatibility for Transformers 5.1.0 is a broad environment issue; loading `physical-intelligence/fast` is a specific model-loading failure." + "reason": "Both are Qwen-related regressions, but one is a load-time failure and the other is a save_pretrained checkpoint corruption; not the same concrete problem." } ] }, @@ -8483,33 +8640,33 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:38175", - "issue:39692", "issue:41950", - "issue:42994", + "issue:42907", "issue:43295", "issue:43299", - "issue:43317", "issue:43441", - "issue:43450", "issue:43479", "issue:43525", "issue:43526", "issue:43550", "issue:43577", "issue:43650", + "issue:43673", "issue:43824", "issue:43854", "issue:43901", "issue:43976", "issue:44016", + "issue:44062", "issue:44075", + "issue:44112", "issue:44242", "issue:44279", - "issue:44291", "issue:44297", + "issue:44336", "issue:44351", "issue:44368", "issue:44462", @@ -8517,11 +8674,12 @@ "issue:44496", "issue:44521", "issue:44617", - "issue:44683", "issue:44779", + "issue:44871", "issue:44877", "issue:44933", "issue:44964", + "issue:44987", "issue:45020", "issue:45042", "issue:45081", @@ -8534,252 +8692,259 @@ "issue:45405" ], "soft_pairs": [ - "issue:44075|issue:44368", - "issue:43824|issue:44933", + "issue:42907|issue:43441", + "issue:44016|issue:44062", + "issue:44112|issue:44964", + "issue:44871|issue:45245", + "issue:43673|issue:44336", + "issue:43976|issue:44987", "issue:44496|issue:45042", - "issue:43976|issue:44279", - "issue:43299|issue:45357", "issue:43526|issue:43901", - "issue:44242|issue:45200", - "issue:44521|issue:45325", + "issue:43824|issue:44933", + "issue:43299|issue:45357", + "issue:43976|issue:44279", "issue:43577|issue:45216", + "issue:44075|issue:44368", + "issue:44521|issue:45325", "issue:43901|issue:44496", "issue:43854|issue:44964", - "issue:44351|issue:45278", - "issue:44464|issue:45245", - "issue:44462|issue:45081", - "issue:41950|issue:45020", + "issue:44242|issue:45200", "issue:41950|issue:43650", + "issue:41950|issue:45020", + "issue:44464|issue:45245", "issue:43976|issue:45405", - "issue:44297|issue:44462", - "issue:44279|issue:45405", + "issue:44351|issue:45278", "issue:43479|issue:43525", + "issue:44462|issue:45081", "issue:44496|issue:44877", - "issue:44462|issue:44779", - "issue:38175|issue:44617", + "issue:44297|issue:44462", + "issue:44279|issue:45405", "issue:43550|issue:44964", "issue:43299|issue:43854", + "issue:44462|issue:44779", "issue:43295|issue:44779", - "issue:39692|issue:43901", - "issue:43441|issue:44683", - "issue:43450|issue:44279", - "issue:44016|issue:44291", - "issue:42994|issue:43317" + "issue:38175|issue:44617" ], - "prepared_review_unit_hash": "170af97dc51d27f9a86823ffd2e591c5cf4edadb631b61511cb40d4a8edee46b", + "prepared_review_unit_hash": "6e57d4beb410dbc043d7a31442e98fbebf42ce6dc94e3ad7f87753ddb4194c9c", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 31, - "serialized_chars": 36330, - "estimated_input_tokens": 9083, - "estimated_eval_tokens": 18422 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37250, + "estimated_input_tokens": 9313, + "estimated_eval_tokens": 18882 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This is a loose thematic cluster of Transformers regressions, loading issues, tokenizer/config bugs, and docs problems. The soft-similarity pairs mostly share broad subsystem vocabulary, but they do not describe the same concrete bug or change, so they should not be merged as duplicates.", + "summary": "The cluster is mostly a grab bag of unrelated Transformers bug reports (model loading, tokenizer/config regressions, docs, CI, and pipeline issues). None of the soft pairs look like the same concrete bug or fix.", "confidence": 0.88, - "canonical_issue_reason": "Issue 45020 is the closest umbrella report in the set because it broadly captures recent-version breakages affecting remote_code models, but the cluster is heterogeneous and does not form a true single-duplicate target.", + "canonical_issue_reason": "issue:45020 is the broadest open, actionable report in the set and the closest umbrella for recent-version model-loading regressions involving remote_code.", "canonical_pr_reason": null, - "best_issue_reason": "Issue 45020 is the best representative issue to anchor the cluster: it is broad, recent, and open, and it can contextualize several version-regression reports better than the narrower model-specific issues.", + "best_issue_reason": "issue:45020 is the best global anchor because it is open, clearly user-facing, and broad enough to represent the regression theme without overfitting to a single model.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:44075", - "right": "issue:44368", + "left": "issue:42907", + "right": "issue:43441", "accept": false, - "reason": "Different problems: SGD optimizer args being ignored vs a Qwen warning about tie_word_embeddings." + "reason": "Different problems: saving a dequantized model vs FlashAttention load failure for Ministral-3." }, { - "left": "issue:43824", - "right": "issue:44933", + "left": "issue:44016", + "right": "issue:44062", "accept": false, - "reason": "Both are import-related, but they concern different missing symbols and different code paths." + "reason": "Notebook syntax error and AddedToken keyword conflict are unrelated bugs." }, { - "left": "issue:44496", - "right": "issue:45042", + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI stale device-override test in GraniteSpeech is not the same as Phi-4 multimodal loading failure." + }, + { + "left": "issue:44871", + "right": "issue:45245", + "accept": false, + "reason": "eos_token_id config mismatch is unrelated to the categories-limit runtime error." + }, + { + "left": "issue:43673", + "right": "issue:44336", "accept": false, - "reason": "Model config loading failure is unrelated to PIL image processors requiring torchvision." + "reason": "Chunked-prefill cache regression and ANSI logging output are unrelated." }, { "left": "issue:43976", - "right": "issue:44279", + "right": "issue:44987", "accept": false, - "reason": "Python version compatibility is not the same as a generic dependency issue." + "reason": "Python-version compatibility issue is unrelated to loading physical-intelligence/fast." }, { - "left": "issue:43299", - "right": "issue:45357", + "left": "issue:44496", + "right": "issue:45042", "accept": false, - "reason": "Both mention Qwen models, but one is a load failure and the other is a save_pretrained checkpoint-key regression." + "reason": "Unrecognized model_type and PIL backend torchvision requirement are different code paths." }, { "left": "issue:43526", "right": "issue:43901", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug is unrelated to a docs mismatch about return_all_scores." + "reason": "BeitImageProcessorFast label reduction bug is unrelated to a docs mismatch." }, { - "left": "issue:44242", - "right": "issue:45200", + "left": "issue:43824", + "right": "issue:44933", "accept": false, - "reason": "MoE load-balancing loss logic is a different bug from missing mm_token_type_ids defaults." + "reason": "Both are import-related, but they concern different symbols/modules and different failures." }, { - "left": "issue:44521", - "right": "issue:45325", + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3VL loading failure and Qwen3.5 save_pretrained key regression are different bugs." + }, + { + "left": "issue:43976", + "right": "issue:44279", "accept": false, - "reason": "Assistant mask generation for multimodal chat templates is unrelated to Qwen2.5-VL rope index scaling." + "reason": "Python version/support problem is unrelated to a vague dependency issue." }, { "left": "issue:43577", "right": "issue:45216", "accept": false, - "reason": "BLIP2 dtype loading and Qwen3.5 checkpoint saving are different model-family regressions." + "reason": "Blip2 dtype propagation and Qwen3.5 checkpoint saving are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Unused SGD args and tie_word_embeddings warning are unrelated." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Multimodal assistant mask bug and Qwen2.5-VL rope-index bug are different generation-path issues." }, { "left": "issue:43901", "right": "issue:44496", "accept": false, - "reason": "A documentation note about return_all_scores is not the same as a model config loading error." + "reason": "Docs about return_all_scores do not match an unrecognized-model loading error." }, { "left": "issue:43854", "right": "issue:44964", "accept": false, - "reason": "These are different model-loading failures for different checkpoints and failure modes." - }, - { - "left": "issue:44351", - "right": "issue:45278", - "accept": false, - "reason": "A missing HybridCache export is not the same as the broader import-errors-after-upgrade report." + "reason": "Different model families and distinct load failures; not the same bug." }, { - "left": "issue:44464", - "right": "issue:45245", + "left": "issue:44242", + "right": "issue:45200", "accept": false, - "reason": "Chunked generation inconsistency is unrelated to the categories cardinality runtime error." + "reason": "Load-balancing loss gating and mm_token_type_ids defaults are unrelated." }, { - "left": "issue:44462", - "right": "issue:45081", + "left": "issue:41950", + "right": "issue:43650", "accept": false, - "reason": "Tokenizer.json loading behavior and a Mistral regex patch crash are different tokenizer bugs." + "reason": "Video-classification processor lookup is unrelated to the vague 'ADD THE DATA' issue." }, { "left": "issue:41950", "right": "issue:45020", "accept": false, - "reason": "Video pipeline processor lookup is unrelated to remote_code model breakages." + "reason": "A pipeline processor-selection bug is not the same as remote_code model-loading regressions." }, { - "left": "issue:41950", - "right": "issue:43650", + "left": "issue:44464", + "right": "issue:45245", "accept": false, - "reason": "A pipeline bug is not the same as a placeholder issue with no clear technical content." + "reason": "Chunked generation inconsistency and category-count limit error are unrelated." }, { "left": "issue:43976", "right": "issue:45405", "accept": false, - "reason": "Python compatibility and an unreleased dependency version bump are unrelated." + "reason": "Python-version support and minimum PEFT version bump are unrelated dependency issues." }, { - "left": "issue:44297", - "right": "issue:44462", + "left": "issue:44351", + "right": "issue:45278", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch is different from AutoTokenizer ignoring tokenizer.json." + "reason": "A missing exported symbol and general import errors are related only at a very high level, not the same bug." }, { - "left": "issue:44279", - "right": "issue:45405", + "left": "issue:43479", + "right": "issue:43525", "accept": false, - "reason": "Generic dependency trouble is not the same as a PEFT minimum-version packaging issue." + "reason": "Phi4MultimodalConfig default init and Llama4Config missing pad_token_id are different config bugs." }, { - "left": "issue:43479", - "right": "issue:43525", + "left": "issue:44462", + "right": "issue:45081", "accept": false, - "reason": "Both are config-related, but they affect different multimodal model classes and different fields." + "reason": "AutoTokenizer ignoring tokenizer.json and Mistral regex patch crashing are distinct tokenizer-loading paths." }, { "left": "issue:44496", "right": "issue:44877", "accept": false, - "reason": "Unrecognized model config loading is not the same as strict config rejecting granite_speech." + "reason": "Unrecognized model_type and strict granite_speech config loading are different configuration issues." }, { - "left": "issue:44462", - "right": "issue:44779", + "left": "issue:44297", + "right": "issue:44462", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json and Deepseek tokenizer output changes are distinct tokenizer bugs." + "reason": "save_pretrained writing the wrong tokenizer_class is not the same as AutoTokenizer ignoring tokenizer.json." }, { - "left": "issue:38175", - "right": "issue:44617", + "left": "issue:44279", + "right": "issue:45405", "accept": false, - "reason": "Unexpected zero probabilities in SigLIP2 is unrelated to a SAM3Video CUDA OOM." + "reason": "Generic dependency issue and PEFT version bump are not the same underlying problem." }, { "left": "issue:43550", "right": "issue:44964", "accept": false, - "reason": "torch.compile/SDPA failure for Bamba is unrelated to Phi-4 multimodal loading." + "reason": "torch.compile + SDPA failure in Bamba is unrelated to Phi-4 multimodal loading." }, { "left": "issue:43299", "right": "issue:43854", "accept": false, - "reason": "These are different model-loading issues for different architectures and checkpoints." + "reason": "Qwen3VL model-loading regression and GLM-4.7-Flash load failure are different models and code paths." }, { - "left": "issue:43295", + "left": "issue:44462", "right": "issue:44779", "accept": false, - "reason": "Processor/tokenizer regression in v4.57.5 is not the same as Deepseek tokenizer incorrect results in v5." - }, - { - "left": "issue:39692", - "right": "issue:43901", - "accept": false, - "reason": "A broken SigLIP2 documentation example is unrelated to a pipeline docs mismatch." - }, - { - "left": "issue:43441", - "right": "issue:44683", - "accept": false, - "reason": "FlashAttention failure in one model and compiled flex_attention failure on newer torch are different attention-path bugs." - }, - { - "left": "issue:43450", - "right": "issue:44279", - "accept": false, - "reason": "Batched video processor shape handling is unrelated to a dependency issue." + "reason": "Tokenizer.json being ignored and Deepseek tokenizer incorrect results are different tokenizer regressions." }, { - "left": "issue:44016", - "right": "issue:44291", + "left": "issue:43295", + "right": "issue:44779", "accept": false, - "reason": "Notebook syntax error and init_empty_weights argument handling are unrelated." + "reason": "Processor.tokenizer regression and Deepseek tokenizer output issues are unrelated." }, { - "left": "issue:42994", - "right": "issue:43317", + "left": "issue:38175", + "right": "issue:44617", "accept": false, - "reason": "Quantized model saving failure is a different bug from dequantized model loading with device_map=auto." + "reason": "Zero probabilities in SigLIP2 and Sam3Video OOM are entirely different failures." } ] }, @@ -8801,18 +8966,21 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ + "issue:39692", "issue:41628", "issue:41950", "issue:42175", - "issue:42491", + "issue:42994", "issue:43065", "issue:43122", "issue:43262", "issue:43295", - "issue:43299", + "issue:43317", "issue:43352", + "issue:43441", + "issue:43450", "issue:43475", "issue:43525", "issue:43650", @@ -8826,265 +8994,262 @@ "issue:43994", "issue:44016", "issue:44162", - "issue:44188", "issue:44220", "issue:44246", "issue:44261", "issue:44263", - "issue:44295", + "issue:44279", + "issue:44291", "issue:44297", - "issue:44336", - "issue:44373", - "issue:44462", "issue:44466", "issue:44521", "issue:44589", "issue:44617", "issue:44623", "issue:44671", + "issue:44683", "issue:44749", "issue:44779", "issue:44898", "issue:45250", "issue:45276", "issue:45290", - "issue:45356", "issue:45362", "issue:45397" ], "soft_pairs": [ - "issue:43723|issue:44162", + "issue:39692|issue:43901", "issue:43650|issue:45290", + "issue:44016|issue:44291", + "issue:42994|issue:43317", + "issue:43441|issue:44683", + "issue:43450|issue:44279", "issue:43065|issue:44898", - "issue:44623|issue:45397", - "issue:44261|issue:44263", "issue:44466|issue:45276", "issue:43262|issue:44521", - "issue:44623|issue:44749", + "issue:43723|issue:44162", + "issue:44261|issue:44263", + "issue:44623|issue:45397", "issue:43122|issue:43650", + "issue:44623|issue:44749", "issue:44671|issue:44779", "issue:43673|issue:43761", - "issue:44297|issue:45397", - "issue:43352|issue:45250", - "issue:42175|issue:43976", "issue:43901|issue:43994", - "issue:43295|issue:44617", - "issue:43475|issue:44016", + "issue:42175|issue:43976", + "issue:44297|issue:45397", "issue:43824|issue:43976", - "issue:44246|issue:44617", - "issue:43525|issue:43881", + "issue:43475|issue:44016", + "issue:43295|issue:44617", "issue:44589|issue:45362", + "issue:43352|issue:45250", "issue:43901|issue:44246", + "issue:43525|issue:43881", "issue:41950|issue:44220", - "issue:41628|issue:43976", - "issue:44188|issue:44336", - "issue:44295|issue:45397", - "issue:44462|issue:45356", - "issue:42491|issue:43299", - "issue:44336|issue:44373" + "issue:43450|issue:44617", + "issue:41628|issue:43976" ], - "prepared_review_unit_hash": "27f1ed4a7562ee981653ded2bcc7350ae42778fe7d057c8fff2e1f68b28ec599", + "prepared_review_unit_hash": "34ed06ea54caeadd4acdee71ba53c0ae3265d1ed862a1f8ea02e37820c125a94", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 29, - "serialized_chars": 36106, - "estimated_input_tokens": 9027, - "estimated_eval_tokens": 18310 + "serialized_chars": 35995, + "estimated_input_tokens": 8999, + "estimated_eval_tokens": 18254 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "The cluster is a loose collection of mostly unrelated Transformers issues across tokenizers, multimodal chat, model loading, vision/audio pipelines, serialization, and docs/perf regressions. None of the soft pairs look like the same concrete bug/change, so I would not collapse any of them.", - "confidence": 0.88, - "canonical_issue_reason": null, + "summary": "This cluster is highly heterogeneous; the soft pairs mostly share only broad transformer-related vocabulary, not the same underlying bug. I found no safe duplicates to merge, and the only weak thematic overlap is around SigLIP2, which is still not enough to accept a pair.", + "confidence": 0.96, + "canonical_issue_reason": "If forced to anchor the cluster, issue 43994 is the closest substantive user-facing bug in the only loosely shared area (SigLIP2), but there is no true canonical duplicate set here.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "43994 is the most concrete and broadly applicable issue among this set; the others are mostly unrelated subsystems or narrow regressions. It is still only a weak representative, not a genuine duplicate hub.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43723", - "right": "issue:44162", + "left": "issue:39692", + "right": "issue:43901", "accept": false, - "reason": "Both are user-facing regressions, but one is an AutoTokenizer import/loading issue and the other is an ESM2 model breakage; different components and failure modes." + "reason": "SigLIP2 docs/example bug vs a text-classification docs mismatch; different feature area and failure mode." }, { "left": "issue:43650", "right": "issue:45290", "accept": false, - "reason": "Completely different issues: a vague non-technical report vs a concrete chat-template crash on assistant tool-call messages." + "reason": "Completely unrelated: one is a vague data request, the other is a chat-template tool-call crash." }, { - "left": "issue:43065", - "right": "issue:44898", + "left": "issue:44016", + "right": "issue:44291", "accept": false, - "reason": "Both touch vision models, but one is about a dummy Conv2d in SAM3 pixel decoding while the other is a Perceiver resolution/interpolation failure; not the same bug." + "reason": "Notebook syntax error vs init_empty_weights constructor regression; unrelated issues." }, { - "left": "issue:44623", - "right": "issue:45397", + "left": "issue:42994", + "right": "issue:43317", "accept": false, - "reason": "Processor save_pretrained missing files and gemma-4 ZeRO3 from_pretrained are unrelated serialization vs distributed-loading problems." + "reason": "Both involve quantization, but one is saving and the other is dequantized model loading/offload; different bugs." }, { - "left": "issue:44261", - "right": "issue:44263", + "left": "issue:43441", + "right": "issue:44683", "accept": false, - "reason": "Different GLM/MoE bugs: one is a missing config epsilon affecting precision, the other is torch.split return handling in a different indexer path." + "reason": "Both mention Flash Attention, but one is model support gating and the other is compiled flex_attention on newer torch; not the same bug." }, { - "left": "issue:44466", - "right": "issue:45276", + "left": "issue:43450", + "right": "issue:44279", "accept": false, - "reason": "One is inconsistent lm_head serialization behavior; the other is resize_token_embeddings not propagating to Gemma4 embedding/output layers. Similar theme, different code paths." + "reason": "Video processor batch-shape bug vs a generic dependency issue; no shared code-path." }, { - "left": "issue:43262", - "right": "issue:44521", + "left": "issue:43065", + "right": "issue:44898", "accept": false, - "reason": "Both involve apply_chat_template, but one is an audio processor default sampling-rate bug and the other is multimodal assistant mask generation. Different failures." + "reason": "Dummy Conv2d in SAM3 pixel decoder vs Perceiver interpolation failure; unrelated model internals." }, { - "left": "issue:44623", - "right": "issue:44749", + "left": "issue:44466", + "right": "issue:45276", "accept": false, - "reason": "Processor save_pretrained missing files is unrelated to the reported slowdown in filtering data after upgrading Transformers." + "reason": "Serialization of tied weights vs resize_token_embeddings not updating Gemma4 embeddings; different concrete problems." }, { - "left": "issue:43122", - "right": "issue:43650", + "left": "issue:43262", + "right": "issue:44521", "accept": false, - "reason": "Tokenization regression vs unrelated non-technical issue; no shared underlying bug." + "reason": "Audio chat-template sampling-rate default vs multimodal assistant-mask bug; same API surface, different behavior." }, { - "left": "issue:44671", - "right": "issue:44779", + "left": "issue:43723", + "right": "issue:44162", "accept": false, - "reason": "Both are v5 regressions, but one is CamemBERT masked-LM prediction quality and the other is DeepSeek tokenizer output; not the same concrete change." + "reason": "Tokenizer import/loading issue vs ESM2 regression; unrelated models and symptoms." }, { - "left": "issue:43673", - "right": "issue:43761", + "left": "issue:44261", + "right": "issue:44263", "accept": false, - "reason": "Generation cache missing during chunked prefill is unrelated to CLIPVisionModel hidden_states behavior." + "reason": "MLA layernorm epsilon precision issue vs torch.split return behavior in a different component; not the same bug." }, { - "left": "issue:44297", + "left": "issue:44623", "right": "issue:45397", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch and gemma-4 ZeRO3 loading are unrelated issues." + "reason": "processor.save_pretrained missing files vs Gemma-4 zero3 from_pretrained failure; save/load bugs but distinct." }, { - "left": "issue:43352", - "right": "issue:45250", + "left": "issue:43122", + "right": "issue:43650", "accept": false, - "reason": "One is a specific model lacking Flash Attention 2.0 support; the other is a broad flash-attention topic. Not the same fixable code-path bug." + "reason": "Tokenizer behavior regression vs unrelated data request." }, { - "left": "issue:42175", - "right": "issue:43976", + "left": "issue:44623", + "right": "issue:44749", "accept": false, - "reason": "Packaging/back-end dependency selection is unrelated to Python 3.9/3.10 compatibility." + "reason": "Processor save bug vs a Chinese report about slow filtering; unrelated." }, { - "left": "issue:43901", - "right": "issue:43994", + "left": "issue:44671", + "right": "issue:44779", "accept": false, - "reason": "Docs mismatch for return_all_scores is unrelated to SigLIP2 producing bad outputs in AutoModel/pipeline." + "reason": "CamemBERT MLM predictions vs Deepseek tokenization; different models and failure modes." }, { - "left": "issue:43295", - "right": "issue:44617", + "left": "issue:43673", + "right": "issue:43761", "accept": false, - "reason": "Processor.tokenizer regression with image inputs is unrelated to Sam3Video CUDA out-of-memory." + "reason": "Generation cache regression vs CLIPVisionModel hidden_states regression; both v5 bugs but different code paths." }, { - "left": "issue:43475", - "right": "issue:44016", + "left": "issue:43901", + "right": "issue:43994", "accept": false, - "reason": "SAM3 vision encoder attribute error and a notebook syntax error are clearly different artifacts." + "reason": "Docs wording about return_all_scores vs a SigLIP2 inference/result bug; no same underlying issue." }, { - "left": "issue:43824", + "left": "issue:42175", "right": "issue:43976", "accept": false, - "reason": "Missing Qwen2.5-VL import and Python version support are separate issues." + "reason": "Package/backend dependency mismatch vs Python version compatibility; unrelated." }, { - "left": "issue:44246", - "right": "issue:44617", + "left": "issue:44297", + "right": "issue:45397", "accept": false, - "reason": "Import-time performance regression is unrelated to Sam3Video memory exhaustion." + "reason": "Tokenizer class metadata mismatch on save vs Gemma-4 deepspeed zero3 loading failure; different layers." }, { - "left": "issue:43525", - "right": "issue:43881", + "left": "issue:43824", + "right": "issue:43976", "accept": false, - "reason": "Missing pad_token_id in Llama4Config and glm-4v-9b loading failure are different model/config problems." + "reason": "Missing Qwen2.5-VL import vs Python support issue; unrelated." }, { - "left": "issue:44589", - "right": "issue:45362", + "left": "issue:43475", + "right": "issue:44016", "accept": false, - "reason": "Float8 storage type error is unrelated to Qwen3.5-35B chat crashes." + "reason": "SAM3 video encoder output attribute bug vs notebook syntax error; unrelated." }, { - "left": "issue:43901", - "right": "issue:44246", + "left": "issue:43295", + "right": "issue:44617", "accept": false, - "reason": "A docs inconsistency and slow import are unrelated." + "reason": "Processor/tokenizer regression vs CUDA OOM in Sam3Video; not the same bug." }, { - "left": "issue:41950", - "right": "issue:44220", + "left": "issue:44589", + "right": "issue:45362", "accept": false, - "reason": "Video-classification pipeline selecting image processors is unrelated to _torch_extract_fbank_features audio feature extraction." + "reason": "Missing Float8 storage type vs Qwen3.5 chat crash; different failures." }, { - "left": "issue:41628", - "right": "issue:43976", + "left": "issue:43352", + "right": "issue:45250", "accept": false, - "reason": "AutoImageProcessor import failure and Transformers Python version support are different issues." + "reason": "Specific model lacking FlashAttention2 support vs a generic Flash Attention 2.0 issue title; too vague to merge." }, { - "left": "issue:44188", - "right": "issue:44336", + "left": "issue:43901", + "right": "issue:44246", "accept": false, - "reason": "Torch.compile attention-kernel divergence is unrelated to ANSI codes in loading_report." + "reason": "Pipeline docs mismatch vs import sometimes slow; unrelated." }, { - "left": "issue:44295", - "right": "issue:45397", + "left": "issue:43525", + "right": "issue:43881", "accept": false, - "reason": "Reading position_ids after buffer registration and gemma-4 ZeRO3 loading are unrelated." + "reason": "Llama4Config pad_token_id attribute error vs glm-4v-9b loading failure; different models and root causes." }, { - "left": "issue:44462", - "right": "issue:45356", + "left": "issue:41950", + "right": "issue:44220", "accept": false, - "reason": "AutoTokenizer ignoring tokenizer.json and Kimi-K2.5 codec/regression warnings are both tokenizer-related but not the same underlying bug." + "reason": "Video-classification pipeline image-processor lookup vs audio feature extraction bug; unrelated." }, { - "left": "issue:42491", - "right": "issue:43299", + "left": "issue:43450", + "right": "issue:44617", "accept": false, - "reason": "Both mention Qwen/MoE models and version regressions, but one is LoRA compatibility across branches while the other is a v5 model-loading breakage; not one concrete fix." + "reason": "Incorrect batched video processor shape vs SAM3Video OOM; distinct symptoms and fixes." }, { - "left": "issue:44336", - "right": "issue:44373", + "left": "issue:41628", + "right": "issue:43976", "accept": false, - "reason": "ANSI escape handling in loading_report and a wrong docstring for position_ids are unrelated." + "reason": "AutoImageProcessor import failure vs Python version compatibility; not the same problem." } ] }, @@ -9106,36 +9271,35 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:36246", "issue:41950", + "issue:42491", + "issue:42757", "issue:42886", "issue:43065", "issue:43066", - "issue:43232", "issue:43257", - "issue:43262", - "issue:43295", + "issue:43299", "issue:43408", "issue:43441", - "issue:43475", "issue:43526", "issue:43531", "issue:43540", "issue:43596", - "issue:43653", "issue:43673", "issue:43901", "issue:43994", - "issue:44060", - "issue:44077", "issue:44162", + "issue:44188", "issue:44242", "issue:44246", "issue:44295", "issue:44336", "issue:44351", + "issue:44373", + "issue:44462", "issue:44485", "issue:44496", "issue:44561", @@ -9152,250 +9316,245 @@ "issue:45229", "issue:45276", "issue:45292", + "issue:45356", "issue:45362", "issue:45375", + "issue:45397", "issue:45459" ], "soft_pairs": [ + "issue:44188|issue:44336", + "issue:44295|issue:45397", + "issue:42491|issue:43299", "issue:44246|issue:44496", - "issue:36246|issue:44351", - "issue:43531|issue:45362", + "issue:44336|issue:44373", + "issue:44462|issue:45356", "issue:43526|issue:45003", "issue:43994|issue:45276", + "issue:43531|issue:45362", + "issue:36246|issue:42757", + "issue:36246|issue:44351", "issue:43673|issue:44561", - "issue:44242|issue:44857", "issue:44336|issue:44485", "issue:36246|issue:43596", + "issue:44242|issue:44857", "issue:44162|issue:45200", - "issue:44295|issue:44749", - "issue:44987|issue:45459", - "issue:44877|issue:45375", "issue:44246|issue:44871", "issue:44779|issue:45292", + "issue:44987|issue:45459", + "issue:44295|issue:44749", + "issue:43441|issue:43540", + "issue:44877|issue:45375", "issue:41950|issue:43596", "issue:43066|issue:44779", - "issue:43441|issue:43540", - "issue:44995|issue:45200", "issue:42886|issue:45229", "issue:43901|issue:45003", - "issue:43232|issue:44336", + "issue:44995|issue:45200", "issue:43065|issue:43408", - "issue:43257|issue:45137", - "issue:43262|issue:44246", - "issue:44496|issue:45003", - "issue:43295|issue:45292", - "issue:43475|issue:43540", - "issue:43653|issue:44162", - "issue:44060|issue:44077" + "issue:43257|issue:45137" ], - "prepared_review_unit_hash": "98b32ec42db3bfcfc91d8f6f413f1dc9971777777d77088df5cb6736f1d57cf7", + "prepared_review_unit_hash": "bd82b06ebbeef5758a0a57b9cfbe00f0836e80d9a5ebf0c061cb604396c884fc", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 47, - "item_count": 47, - "soft_pair_count": 30, - "serialized_chars": 36094, - "estimated_input_tokens": 9024, - "estimated_eval_tokens": 18304 + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36191, + "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 issues across tokenizers, generation, vision/audio processors, SAM3, Qwen, Gemma, DeepSpeed, and docs/perf. None of the soft-edge pairs look like the same underlying bug or change, so all are rejected.", - "confidence": 0.88, - "canonical_issue_reason": "issue:43295 is the broadest user-facing regression in the set, centered on core processor/tokenizer behavior breaking custom code and image inputs.", + "summary": "This cluster is heterogeneous: the suggested pairs span unrelated bugs, regressions, docs, config, tokenizer, and model-loading issues. I found no soft edge that safely collapses to the same underlying change, and there are no PRs in the set.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this cluster well; the items are not one duplicate family but many distinct problems across different subsystems and models.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43295 is the strongest representative issue here: a concrete regression in a public API with clear downstream impact and reproduction value.", + "best_issue_reason": "No issue is a strong global representative for deduping here, because the cluster lacks a shared concrete code path or fix target.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Torch.compile attention-kernel branching and ANSI output/reporting are unrelated code paths." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Reading registered position_ids and Gemma-4 ZeRO3 from_pretrained are different failures." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3-MoE, but one is LoRA portability and the other is model loading; not the same concrete bug." + }, { "left": "issue:44246", "right": "issue:44496", "accept": false, - "reason": "Import slowdown vs model-loading/config error; different failure modes and code paths." + "reason": "Import latency and an unrecognized model/config error are unrelated issues." }, { - "left": "issue:36246", - "right": "issue:44351", + "left": "issue:44336", + "right": "issue:44373", "accept": false, - "reason": "Different missing symbols from unrelated APIs: Qwen2.5-VL processor import vs HybridCache import." + "reason": "Terminal ANSI output generation and a wrong docstring for position_ids do not share a bug." }, { - "left": "issue:43531", - "right": "issue:45362", + "left": "issue:44462", + "right": "issue:45356", "accept": false, - "reason": "Same broad Qwen area, but sliding_window behavior and a chat crash are distinct bugs." + "reason": "AutoTokenizer ignoring tokenizer.json and Kimi-K2.5 codec/warning regressions are different tokenizer bugs." }, { "left": "issue:43526", "right": "issue:45003", "accept": false, - "reason": "BeitImageProcessorFast label reduction bug vs unsafe sys.modules access in modeling_utils." + "reason": "BeitImageProcessorFast label reduction and unsafe sys.modules access are unrelated." }, { "left": "issue:43994", "right": "issue:45276", "accept": false, - "reason": "SIGLIP pipeline/model output issue vs Gemma4 embedding-resize propagation bug." + "reason": "Siglip2 model/pipeline output issues and resize_token_embeddings propagation are different problems." }, { - "left": "issue:43673", - "right": "issue:44561", + "left": "issue:43531", + "right": "issue:45362", "accept": false, - "reason": "Generation cache chunked_prefill regression vs removal of is_torch_fx_available affecting remote-code models." + "reason": "Both are Qwen3-family bugs, but sliding_window handling and chat crashes are distinct code paths." }, { - "left": "issue:44242", - "right": "issue:44857", + "left": "issue:36246", + "right": "issue:42757", "accept": false, - "reason": "MoE load-balancing loss omission vs LwDetrImageLoss AMP/CUDA crash." + "reason": "Different missing imports from different packages; same symptom class, not the same bug." }, { - "left": "issue:44336", - "right": "issue:44485", + "left": "issue:36246", + "right": "issue:44351", "accept": false, - "reason": "ANSI output formatting bug vs GLM-5 RoPE implementation discussion." + "reason": "Both are import errors, but they concern different symbols and likely different fixes." }, { - "left": "issue:36246", - "right": "issue:43596", + "left": "issue:43673", + "right": "issue:44561", "accept": false, - "reason": "Qwen2.5-VL import error vs DeepSpeed ZeRO3/BERT index error; unrelated subsystems." + "reason": "Generation cache chunked_prefill and removal of is_torch_fx_available are unrelated regressions." }, { - "left": "issue:44162", - "right": "issue:45200", + "left": "issue:44336", + "right": "issue:44485", "accept": false, - "reason": "ESM2 regression vs Gemma4 mm_token_type_ids defaulting issue." + "reason": "ANSI/reporting behavior and GLM-5 RoPE implementation discussion are not the same issue." }, { - "left": "issue:44295", - "right": "issue:44749", + "left": "issue:36246", + "right": "issue:43596", "accept": false, - "reason": "Buffer access error vs performance regression after upgrade; not the same bug." + "reason": "Qwen2.5-VL import failure and DeepSpeed ZeRO3/BertModel index error are unrelated." }, { - "left": "issue:44987", - "right": "issue:45459", + "left": "issue:44242", + "right": "issue:44857", "accept": false, - "reason": "Model loading failure vs tokenizer error handling around missing protobuf." + "reason": "Missing MoE load-balancing loss and an AMP/CUDA crash are different bugs." }, { - "left": "issue:44877", - "right": "issue:45375", + "left": "issue:44162", + "right": "issue:45200", "accept": false, - "reason": "Both involve config strictness, but different models and different missing fields." + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are unrelated model-specific issues." }, { "left": "issue:44246", "right": "issue:44871", "accept": false, - "reason": "Import performance issue vs Gemma eos_token_id configuration mismatch." + "reason": "Slow import timing and eos_token_id config mismatch are unrelated." }, { "left": "issue:44779", "right": "issue:45292", "accept": false, - "reason": "DeepSeek tokenizer correctness regression vs resize_token_embeddings propagation bug." + "reason": "Tokenizer regression and resize_token_embeddings not updating outputs are different code paths." }, { - "left": "issue:41950", - "right": "issue:43596", + "left": "issue:44987", + "right": "issue:45459", "accept": false, - "reason": "Video-classification processor selection bug vs DeepSpeed ZeRO3 init crash." + "reason": "Model loading failure and protobuf error masking are unrelated tokenizer/loading issues." }, { - "left": "issue:43066", - "right": "issue:44779", + "left": "issue:44295", + "right": "issue:44749", "accept": false, - "reason": "Both tokenizer-related, but decoder-type selection and DeepSeek output correctness are different problems." + "reason": "Position_ids buffer access and a general slowdown regression are not the same bug." }, { "left": "issue:43441", "right": "issue:43540", "accept": false, - "reason": "FlashAttention failure in Ministral vs video-input ValueError in Qwen3OmniMoe." + "reason": "Ministral FlashAttention failure and Qwen3OmniMoe video input errors are unrelated." }, { - "left": "issue:44995", - "right": "issue:45200", + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-validation problems, but for different models and different missing fields." + }, + { + "left": "issue:41950", + "right": "issue:43596", + "accept": false, + "reason": "Video-classification pipeline processor selection and ZeRO3/BertModel index errors are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:44779", "accept": false, - "reason": "Stale indexer-cache crash vs text-only fine-tuning token-type defaults." + "reason": "Tokenizer decoder type regression and DeepSeek tokenizer regression are not the same concrete issue." }, { "left": "issue:42886", "right": "issue:45229", "accept": false, - "reason": "Offline cache tokenizer load bug vs Gemma4 multi-GPU OOM; unrelated." + "reason": "Offline cache loading and Gemma4 multi-GPU OOM are unrelated." }, { "left": "issue:43901", "right": "issue:45003", "accept": false, - "reason": "Docs mismatch about return_all_scores vs unsafe sys.modules lookup." + "reason": "Docs mismatch and unsafe sys.modules access are not the same bug." }, { - "left": "issue:43232", - "right": "issue:44336", + "left": "issue:44995", + "right": "issue:45200", "accept": false, - "reason": "Generation kwargs sync bug vs ANSI codes in loading_report." + "reason": "Stale indexer cache on second forward and Gemma 4 text-only defaults are unrelated." }, { "left": "issue:43065", "right": "issue:43408", "accept": false, - "reason": "SAM3 pixel decoder implementation issue vs model-type mismatch warning." + "reason": "SAM3 dummy Conv2d implementation and sam3_video/sam3_tracker config warning are related only by family, not the same fix." }, { "left": "issue:43257", "right": "issue:45137", "accept": false, - "reason": "Qwen3 MoE weight conversion with accelerate/deepspeed vs generic ZeRO3 deque error." - }, - { - "left": "issue:43262", - "right": "issue:44246", - "accept": false, - "reason": "Audio chat-template sampling-rate bug vs import slowdown; unrelated." - }, - { - "left": "issue:44496", - "right": "issue:45003", - "accept": false, - "reason": "Unrecognized-model/config issue vs unsafe sys.modules access bug." - }, - { - "left": "issue:43295", - "right": "issue:45292", - "accept": false, - "reason": "Processor.tokenizer/image-input regression vs embedding resize propagation." - }, - { - "left": "issue:43475", - "right": "issue:43540", - "accept": false, - "reason": "SAM3 vision encoder attribute error vs Qwen3OmniMoe video-processing failure." - }, - { - "left": "issue:43653", - "right": "issue:44162", - "accept": false, - "reason": "BigBird tokenizer special-token registration bug vs ESM2 model breakage." - }, - { - "left": "issue:44060", - "right": "issue:44077", - "accept": false, - "reason": "Qwen3-Next tied-weights warning vs PatchTSMixer post_init policy change; different code paths." + "reason": "Qwen3-MoE weight conversion with accelerate/deepspeed and ZeRO3 deque underflow are different bugs." } ] }, @@ -9417,36 +9576,37 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:28282", "issue:30990", - "issue:36246", - "issue:41628", - "issue:42222", - "issue:42371", "issue:42673", - "issue:42994", "issue:43232", + "issue:43262", + "issue:43295", "issue:43381", "issue:43441", "issue:43475", - "issue:43525", + "issue:43540", "issue:43606", "issue:43643", "issue:43646", + "issue:43653", "issue:43673", "issue:43701", "issue:43704", - "issue:43824", "issue:43931", "issue:44038", + "issue:44060", "issue:44077", "issue:44162", + "issue:44246", "issue:44297", + "issue:44336", "issue:44361", "issue:44368", "issue:44466", + "issue:44496", "issue:44623", "issue:44734", "issue:44749", @@ -9455,9 +9615,7 @@ "issue:44977", "issue:45003", "issue:45081", - "issue:45092", "issue:45103", - "issue:45200", "issue:45230", "issue:45278", "issue:45292", @@ -9469,244 +9627,258 @@ "issue:45459" ], "soft_pairs": [ + "issue:43232|issue:44336", + "issue:44496|issue:45003", + "issue:43262|issue:44246", + "issue:43295|issue:45292", + "issue:44060|issue:44077", + "issue:43475|issue:43540", "issue:44466|issue:45335", - "issue:44933|issue:45459", + "issue:43653|issue:44162", "issue:44368|issue:44977", "issue:43673|issue:44734", + "issue:44933|issue:45459", "issue:43441|issue:43606", "issue:44869|issue:45356", "issue:43232|issue:43673", "issue:43931|issue:44368", - "issue:44297|issue:44623", "issue:43673|issue:45278", + "issue:44297|issue:44623", "issue:43646|issue:44162", "issue:43381|issue:45292", - "issue:42673|issue:43704", - "issue:44361|issue:44749", "issue:43381|issue:45335", + "issue:42673|issue:43704", "issue:28282|issue:30990", - "issue:45230|issue:45397", + "issue:44361|issue:44749", + "issue:44077|issue:45003", "issue:44749|issue:45310", + "issue:45230|issue:45397", "issue:44038|issue:44623", - "issue:36246|issue:43824", - "issue:44077|issue:45003", - "issue:42371|issue:45092", "issue:43475|issue:45381", - "issue:43701|issue:44749", - "issue:45081|issue:45103", - "issue:41628|issue:44933", "issue:43475|issue:43643", - "issue:42994|issue:43525", - "issue:42222|issue:45200" + "issue:43701|issue:44749", + "issue:45081|issue:45103" ], - "prepared_review_unit_hash": "062d9e7c1524dc5c0b2d428bade1235ae547dacc52f4d23e8a921926473b2052", + "prepared_review_unit_hash": "77fa6eb05ff4c306902569c5ce267ccea6439d1799aef53bd0cbc01368ae5df4", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 29, - "serialized_chars": 36389, - "estimated_input_tokens": 9098, - "estimated_eval_tokens": 18452 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36501, + "estimated_input_tokens": 9126, + "estimated_eval_tokens": 18508 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of Transformers bug reports. The only clear duplicate-like pair is the two Qwen3ForCausalLM VRAM-leak reports; the rest are separate regressions across imports, serialization, generation, tokenizer, and model-specific behavior.", - "confidence": 0.85, - "canonical_issue_reason": "issue:45278 is the best umbrella-style issue in this set: it is broad, explicitly about many import errors after a v4\u2192v5 upgrade, and is the most plausible triage anchor among these mostly unrelated bug reports.", + "summary": "Most candidate pairs are unrelated or only loosely similar. The only clear duplicate is the Qwen3ForCausalLM VRAM leak report; everything else appears to be separate bugs or regressions in different code paths/models.", + "confidence": 0.93, + "canonical_issue_reason": "issue:42673 is the clearer, higher-signal report for the Qwen3ForCausalLM VRAM leak, and issue:43704 appears to be the same bug reported again.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the strongest overall representative because it captures a broad, user-facing regression class that overlaps with several other import-related reports in the cluster.", + "best_issue_reason": "issue:42673 best represents the only true duplicate pair in this set and is the strongest canonical issue choice.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Different subsystems: generation kwargs vs terminal ANSI output formatting." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading error is unrelated to unsafe sys.modules access." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio sampling-rate defaulting and slow import are unrelated bugs." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Custom processor/tokenizer regression is unrelated to embedding resize behavior." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Weight-tying warning bug and patchtsmixer post_init restriction are different issues." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Both involve multimodal/video models, but they are distinct failures in different model paths." + }, { "left": "issue:44466", "right": "issue:45335", "accept": false, - "reason": "Both involve token embedding resizing, but one is a tied-weights serialization inconsistency and the other is output_embeddings/decoder.embed_tokens not updating. Different failure modes." + "reason": "Both touch tied embeddings, but one is serialization/device-dependent and the other is resize_token_embeddings propagation." }, { - "left": "issue:44933", - "right": "issue:45459", + "left": "issue:43653", + "right": "issue:44162", "accept": false, - "reason": "One is a missing import from image_utils; the other is tokenizer error handling when protobuf is absent. Different subsystems and symptoms." + "reason": "Tokenizer special-token registration and ESM2 breakage are unrelated." }, { "left": "issue:44368", "right": "issue:44977", "accept": false, - "reason": "Both mention Qwen and flash-attention-ish behavior, but one is a tie_word_embeddings warning and the other is a generation failure. Not the same bug." + "reason": "A tied-weights warning and a flash-attention generation bug are not the same concrete defect." }, { "left": "issue:43673", "right": "issue:44734", "accept": false, - "reason": "Both are generation/cache-related, but chunked_prefill cache missing and /v1/responses KV-indexing crash are distinct code paths and symptoms." + "reason": "Both concern generation/cache, but the reported failures and code paths differ." + }, + { + "left": "issue:44933", + "right": "issue:45459", + "accept": false, + "reason": "Missing image_utils import and protobuf error masking are different import/runtime problems." }, { "left": "issue:43441", "right": "issue:43606", "accept": false, - "reason": "Ministral flash-attention failure and bark-small CPU-offload device mismatch are unrelated model-specific bugs." + "reason": "FlashAttention failure for Ministral-3 is unrelated to CPU offload device mismatch for bark-small." }, { "left": "issue:44869", "right": "issue:45356", "accept": false, - "reason": "Both are tokenizer/decoding issues, but Whisper trailing replacement-character crashes and Kimi-K2.5 codec/fix_mistral_regex regressions are different problems." + "reason": "Whisper timestamp decoding crash and Kimi tokenizer codec regression are different tokenizer bugs." }, { "left": "issue:43232", "right": "issue:43673", "accept": false, - "reason": "One is a sync_gpus/_update_model_kwargs_for_generation issue; the other is a chunked_prefill cache regression. Same area, different bug." + "reason": "These are different generation-related regressions with distinct triggers and symptoms." }, { "left": "issue:43931", "right": "issue:44368", "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and a Qwen3.5 tie_word_embeddings warning are unrelated symptoms." + "reason": "Weight shape mismatch in Qwen3-VL is unrelated to a tied-embeddings warning in Qwen3.5." }, { - "left": "issue:44297", - "right": "issue:44623", + "left": "issue:43673", + "right": "issue:45278", "accept": false, - "reason": "Tokenizer save_pretrained metadata mismatch vs processor.save_pretrained missing files are adjacent but distinct save/export bugs." + "reason": "A specific chunked_prefill cache bug is not the same as broad import errors after upgrade." }, { - "left": "issue:43673", - "right": "issue:45278", + "left": "issue:44297", + "right": "issue:44623", "accept": false, - "reason": "Generation cache regression and broad import failures after upgrade are unrelated; no shared underlying bug." + "reason": "Tokenizer save_pretrained metadata mismatch and processor save_pretrained missing files are separate serialization bugs." }, { "left": "issue:43646", "right": "issue:44162", "accept": false, - "reason": "Custom model initialization breakage and ESM2 breakage are both broad, but not the same concrete defect." + "reason": "Custom model initialization breakage is not the same as the ESM2-specific failure." }, { "left": "issue:43381", "right": "issue:45292", "accept": false, - "reason": "Gradient checkpointing eval-mode restriction and resize_token_embeddings not updating output embeddings are different model-behavior bugs." + "reason": "Gradient checkpointing eval-mode behavior is unrelated to embedding resize propagation." + }, + { + "left": "issue:43381", + "right": "issue:45335", + "accept": false, + "reason": "Different features and failure modes: checkpointing vs token embedding resize." }, { "left": "issue:42673", "right": "issue:43704", "accept": true, - "reason": "These have the exact same title and describe the same Qwen3ForCausalLM VRAM leak in multiple dataloader threads." + "reason": "Exact same title and same Qwen3ForCausalLM VRAM leak in multiple dataloader threads." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch missing ImportError and Sentence Transformers hanging are not the same bug." }, { "left": "issue:44361", "right": "issue:44749", "accept": false, - "reason": "MLukeTokenizer AttributeError on tasks and a post-upgrade data filtering slowdown are unrelated." + "reason": "Tokenizer AttributeError on tasks and slowdown after upgrade are unrelated." }, { - "left": "issue:43381", - "right": "issue:45335", + "left": "issue:44077", + "right": "issue:45003", "accept": false, - "reason": "Eval-mode gradient checkpointing and t5gemma embedding resize behavior are separate issues." + "reason": "Patchtsmixer post_init policy and unsafe sys.modules access are different code issues." }, { - "left": "issue:28282", - "right": "issue:30990", + "left": "issue:44749", + "right": "issue:45310", "accept": false, - "reason": "PyTorch-missing ImportError and Sentence Transformers loading hang are clearly different failures." + "reason": "Performance regression during filtering is unrelated to Qwen3.5 MoE from_pretrained errors." }, { "left": "issue:45230", "right": "issue:45397", "accept": false, - "reason": "A generic bug report and a Gemma-4 zero3 from_pretrained issue are not the same underlying problem." - }, - { - "left": "issue:44749", - "right": "issue:45310", - "accept": false, - "reason": "Data filtering slowdown after upgrade and Qwen3.5-Moe from_pretrained error are unrelated." + "reason": "Generic bug report is too vague and does not match the specific gemma-4 zero3 failure." }, { "left": "issue:44038", "right": "issue:44623", "accept": false, - "reason": "Qwen3-VL-Moe bug and processor.save_pretrained missing files are different issues." + "reason": "Qwen3-VL-Moe loading error and processor save_pretrained missing files are separate issues." }, { - "left": "issue:36246", - "right": "issue:43824", - "accept": false, - "reason": "Both are Qwen2.5-VL import errors, but they involve different missing symbols and different import locations; not enough to treat as the same bug." - }, - { - "left": "issue:44077", - "right": "issue:45003", - "accept": false, - "reason": "Optional post_init handling in patchtsmixer and unsafe sys.modules access in modeling_utils are distinct code issues." - }, - { - "left": "issue:42371", - "right": "issue:45092", + "left": "issue:43475", + "right": "issue:45381", "accept": false, - "reason": "TF32 API guidance and InternVL2 meta-initialization incompatibility are unrelated." + "reason": "Both are vision/video-related, but they affect different models and different tensor fields." }, { "left": "issue:43475", - "right": "issue:45381", + "right": "issue:43643", "accept": false, - "reason": "SAM 3 Video missing fpn_position_embeddings and qwen2.5-vl vision_position_ids bug are different model-specific regressions." + "reason": "A missing vision encoder attribute is unrelated to missing fields from AutoConfig with trust_remote_code." }, { "left": "issue:43701", "right": "issue:44749", "accept": false, - "reason": "Checkpoint-resume key mismatch and a training-data slowdown are unrelated." + "reason": "Checkpoint key mismatch and filtering slowdown are unrelated." }, { "left": "issue:45081", "right": "issue:45103", "accept": false, - "reason": "Mistral regex/tokenizer crash and auto_docstring annotations crash are unrelated implementation bugs." - }, - { - "left": "issue:41628", - "right": "issue:44933", - "accept": false, - "reason": "Both are import-related, but AutoImageProcessor missing vs a bad image_utils import are not the same concrete failure." - }, - { - "left": "issue:43475", - "right": "issue:43643", - "accept": false, - "reason": "SAM 3 Vision output attribute error and trust_remote_code AutoConfig missing fields are unrelated." - }, - { - "left": "issue:42994", - "right": "issue:43525", - "accept": false, - "reason": "Quantized saving failure and Llama4Config missing pad_token_id are separate bugs." - }, - { - "left": "issue:42222", - "right": "issue:45200", - "accept": false, - "reason": "Broken vitpose models and Gemma 4 mm_token_type_ids defaults are different model-family regressions." + "reason": "Both are crashes in helper code, but they come from different utilities and different causes." } ] }, "evaluator_result": { "accept": true, - "feedback": "The analyst output is conservative and grounded in the packet. The umbrella issue choice is defensible from the title and cluster makeup, and the one accepted soft pair (42673/43704) is clearly duplicate-like with matching titles. The remaining soft-edge rejections stay on the cautious side and do not overstate shared causality." + "feedback": "The summary is grounded in the packet and stays conservative. The duplicate call on issue:42673 / issue:43704 is well supported by the exact same title and matching symptom, and the other soft-edge rejections are reasonable based on clearly different titles/code paths. I did not see evidence of overstatement." }, "evaluator_used": true, "retried": false, @@ -9725,13 +9897,15 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:30990", "issue:35141", "issue:36010", "issue:41628", "issue:42175", + "issue:42222", + "issue:42371", "issue:42831", "issue:42898", "issue:42994", @@ -9740,16 +9914,15 @@ "issue:43377", "issue:43475", "issue:43479", + "issue:43525", "issue:43673", - "issue:43723", "issue:43784", "issue:43824", "issue:43873", - "issue:43874", "issue:43976", "issue:44038", - "issue:44060", "issue:44079", + "issue:44164", "issue:44188", "issue:44222", "issue:44242", @@ -9759,7 +9932,6 @@ "issue:44623", "issue:44661", "issue:44734", - "issue:44779", "issue:44792", "issue:44869", "issue:44933", @@ -9767,276 +9939,268 @@ "issue:44964", "issue:45092", "issue:45103", - "issue:45125", "issue:45137", + "issue:45200", "issue:45310", - "issue:45356", "issue:45362", "issue:45375", "issue:45397", "issue:45478" ], "soft_pairs": [ + "issue:42371|issue:45092", "issue:45375|issue:45478", + "issue:41628|issue:44933", + "issue:42994|issue:43525", "issue:45310|issue:45375", - "issue:43873|issue:44945", "issue:43475|issue:44661", - "issue:45310|issue:45397", + "issue:42222|issue:45200", + "issue:43873|issue:44945", "issue:42831|issue:44222", - "issue:41628|issue:44351", - "issue:43232|issue:44188", "issue:45362|issue:45375", + "issue:45310|issue:45397", + "issue:43232|issue:44188", + "issue:41628|issue:44351", "issue:44222|issue:45310", - "issue:42994|issue:43824", + "issue:42994|issue:44164", + "issue:43232|issue:44792", "issue:44038|issue:45310", + "issue:30990|issue:35141", "issue:36010|issue:41628", "issue:43377|issue:44242", - "issue:43232|issue:44792", - "issue:30990|issue:35141", - "issue:43673|issue:44188", "issue:43479|issue:44079", + "issue:43673|issue:44188", "issue:43479|issue:44869", "issue:42898|issue:44568", + "issue:44964|issue:45137", "issue:42175|issue:44933", - "issue:43784|issue:43824", "issue:43976|issue:44933", - "issue:44964|issue:45137", - "issue:44188|issue:44373", - "issue:44623|issue:45310", + "issue:43784|issue:43824", "issue:43278|issue:45137", + "issue:44623|issue:45310", "issue:44734|issue:45103", - "issue:44060|issue:45125", - "issue:42898|issue:45356", - "issue:43723|issue:44568", - "issue:43874|issue:45092", - "issue:42898|issue:44779" + "issue:44188|issue:44373" ], - "prepared_review_unit_hash": "cc5310c70b4bb7f6f97fce13780ca9588495d8a8cf9856d15bee41520ededec2", + "prepared_review_unit_hash": "9a226759d146e854f6b650cf2ab9522a2b75b28035c4f442a7554b2b1540a255", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37346, - "estimated_input_tokens": 9337, - "estimated_eval_tokens": 18930 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36335, + "estimated_input_tokens": 9084, + "estimated_eval_tokens": 18424 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a mix of unrelated Transformers bug reports, not one duplicate cluster. No soft edge meets the duplicate threshold. The clearest anchor issue is the Qwen3.5 MoE from_pretrained regression, with issue 45310 as the best canonical representative because it is the most specific and is effectively repeated by the later identical-titled report.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 45310 is the clearest, most specific bug report in the set and best serves as the anchor for the repeated Qwen3.5 MoE from_pretrained regression.", + "summary": "This cluster is not a true duplicate set; it contains many unrelated Transformers issues spanning imports, config/model loading, generation, quantization, and serving. None of the soft edges are strong enough to merge under a strict duplicate policy.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "45310 is the best global issue candidate: it has a tight scope, a concrete failure mode, and the strongest evidence of being the recurring report in this set.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "Different problems: TF32 settings warning vs. old InternVL2 remote-code/meta-init incompatibility." + }, { "left": "issue:45375", "right": "issue:45478", "accept": false, - "reason": "Different failures: one is a missing vision-config field, the other is a Qwen3.5 MoE from_pretrained error." + "reason": "Same model family, but one is a missing config field under @strict and the other is a generic from_pretrained failure; not clearly the same bug." }, { - "left": "issue:45310", - "right": "issue:45375", + "left": "issue:41628", + "right": "issue:44933", "accept": false, - "reason": "Same model family but different bug surfaces; config-field omission is not the same as the from_pretrained crash." + "reason": "Both are import-related, but they target different symbols/modules and likely different code paths." }, { - "left": "issue:43873", - "right": "issue:44945", + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized saving failure vs. Llama4Config missing pad_token_id are unrelated." + }, + { + "left": "issue:45310", + "right": "issue:45375", "accept": false, - "reason": "Unrelated subsystems and symptoms: quantization/offloading vs pipeline-parallel output correctness." + "reason": "Both mention Qwen3.5 MoE loading, but one is a specific config field issue and the other is a broader from_pretrained error without enough evidence they are the same defect." }, { "left": "issue:43475", "right": "issue:44661", "accept": false, - "reason": "Different areas entirely: SAM3 vision output attribute vs tokenizer-mapping model registration." + "reason": "SAM 3 vision output attribute error vs. tokenizer mapping failure; unrelated subsystems." }, { - "left": "issue:45310", - "right": "issue:45397", + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Different models and different failures: VitPose breakage vs. Gemma 4 mm_token_type_ids defaulting." + }, + { + "left": "issue:43873", + "right": "issue:44945", "accept": false, - "reason": "Both are from_pretrained issues, but for different models and likely different root causes." + "reason": "Quantization offloading vs. pipeline-parallel output corruption are different issues." }, { "left": "issue:42831", "right": "issue:44222", "accept": false, - "reason": "FineGrainedFP8 accuracy regression is not the same as FP8 save_pretrained handling." + "reason": "Both involve FP8, but one is an accuracy regression and the other is save_pretrained behavior for MoE; not the same concrete bug." }, { - "left": "issue:41628", - "right": "issue:44351", + "left": "issue:45362", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is a chat crash and the other is a config-field omission; insufficient overlap." + }, + { + "left": "issue:45310", + "right": "issue:45397", "accept": false, - "reason": "Both are import errors, but for different missing symbols and unrelated code paths." + "reason": "Same broad loading area, but different model families (Qwen3.5 MoE vs Gemma-4 zero3)." }, { "left": "issue:43232", "right": "issue:44188", "accept": false, - "reason": "Generation cache update bug vs attention-kernel divergence under torch.compile are different defects." + "reason": "Both are generation-related, but the reported failures involve different mechanisms and code paths." }, { - "left": "issue:45362", - "right": "issue:45375", + "left": "issue:41628", + "right": "issue:44351", "accept": false, - "reason": "Qwen3.5 chat crash and missing deepstack_visual_indexes are distinct issues." + "reason": "Different missing imports; no evidence they stem from the same API break." }, { "left": "issue:44222", "right": "issue:45310", "accept": false, - "reason": "FP8 save_pretrained/moe serialization is unrelated to the Qwen3.5 MoE from_pretrained failure." + "reason": "FP8 save_pretrained bug vs Qwen3.5 MoE from_pretrained error are distinct." }, { "left": "issue:42994", - "right": "issue:43824", + "right": "issue:44164", + "accept": false, + "reason": "Both concern save/from_pretrained, but one is quantized model saving and the other is extra_state handling; not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:44792", "accept": false, - "reason": "Quantized model saving failure and missing Qwen2_5_VL import are unrelated." + "reason": "Generation kwargs sync issue vs. Janus image-generation test failure are unrelated." }, { "left": "issue:44038", "right": "issue:45310", "accept": false, - "reason": "Same model family, but different version and different failure mode; not the same concrete bug." + "reason": "Both mention Qwen variants, but the titles indicate different model releases and different failure modes." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "SentenceTransformers loading stall vs embedding reinit after resize are unrelated." }, { "left": "issue:36010", "right": "issue:41628", "accept": false, - "reason": "Different missing imports: GenerationMixin vs AutoImageProcessor." + "reason": "Both are import errors, but for different removed/moved names and different compatibility issues." }, { "left": "issue:43377", "right": "issue:44242", "accept": false, - "reason": "Batched-vs-single encoder mismatch is unrelated to MoE router-loss omission." + "reason": "MIMI batching/padding-mask bug vs MoE load-balancing loss omission are unrelated." }, { - "left": "issue:43232", - "right": "issue:44792", - "accept": false, - "reason": "Generation sync_gpus/cache issue is unrelated to the janus image-generation test failure." - }, - { - "left": "issue:30990", - "right": "issue:35141", + "left": "issue:43479", + "right": "issue:44079", "accept": false, - "reason": "Sentence Transformers loading stall and embedding reinit on resize are different bugs." + "reason": "Config default initialization bug vs ModelOutput key assignment bug are different classes of failure." }, { "left": "issue:43673", "right": "issue:44188", "accept": false, - "reason": "Missing GenerationMixin cache in chunked_prefill is not the same as branching divergence under torch.compile." - }, - { - "left": "issue:43479", - "right": "issue:44079", - "accept": false, - "reason": "Config default initialization bug and ModelOutput key assignment bug are unrelated." + "reason": "Chunked prefill cache issue vs torch.compile attention-kernel divergence are different code paths." }, { "left": "issue:43479", "right": "issue:44869", "accept": false, - "reason": "Multimodal config initialization and Whisper timestamp decode crash are unrelated." + "reason": "Phi4 multimodal config initialization vs Whisper timestamp decoding crash are unrelated." }, { "left": "issue:42898", "right": "issue:44568", "accept": false, - "reason": "Both involve tokenizer behavior, but one is a v5 cleanup setting change and the other is missing BOS/EOS insertion for a specific tokenizer." + "reason": "Both involve tokenizer behavior changes, but one is clean_up_tokenization_spaces and the other is add_special_tokens/BOS-EOS insertion; not the same bug." }, { - "left": "issue:42175", - "right": "issue:44933", + "left": "issue:44964", + "right": "issue:45137", "accept": false, - "reason": "Backend packaging/install issue vs nonexistent image_utils import are unrelated." + "reason": "Phi-4 multimodal loading failure vs DeepSpeed ZeRO3 deque error are unrelated." }, { - "left": "issue:43784", - "right": "issue:43824", + "left": "issue:42175", + "right": "issue:44933", "accept": false, - "reason": "Sentence-transformers import NameError and Qwen2_5_VL import error are different failures." + "reason": "Package extra/backend dependency issue vs missing image_utils import are unrelated." }, { "left": "issue:43976", "right": "issue:44933", "accept": false, - "reason": "Python version compatibility problem is unrelated to the image_utils import issue." + "reason": "Python version compatibility bug vs missing import are different." }, { - "left": "issue:44964", - "right": "issue:45137", + "left": "issue:43784", + "right": "issue:43824", "accept": false, - "reason": "Phi-4 multimodal loading and DeepSpeed ZeRO3 deque error are different code paths." + "reason": "Both are import errors, but they reference different missing symbols and likely different upstream regressions." }, { - "left": "issue:44188", - "right": "issue:44373", + "left": "issue:43278", + "right": "issue:45137", "accept": false, - "reason": "Attention-kernel divergence is unrelated to a docstring-only issue." + "reason": "Embedding dtype drift in eval vs ZeRO3 deque underflow are unrelated." }, { "left": "issue:44623", "right": "issue:45310", "accept": false, - "reason": "Processor.save_pretrained missing files is not the same as a Qwen3.5 MoE from_pretrained failure." - }, - { - "left": "issue:43278", - "right": "issue:45137", - "accept": false, - "reason": "Dtype drift between train/eval and ZeRO3 deque underflow are unrelated." + "reason": "Missing files in processor.save_pretrained vs Qwen3.5 MoE loading error are not the same issue." }, { "left": "issue:44734", "right": "issue:45103", "accept": false, - "reason": "KV-cache tensor indexing crash and auto_docstring future-annotations AttributeError are unrelated." - }, - { - "left": "issue:44060", - "right": "issue:45125", - "accept": false, - "reason": "Both involve Qwen3.5 MoE, but one is tied-weights warning logic and the other is missing tensor-parallel plan support." + "reason": "Serving KV-cache tensor indexing crash vs auto_docstring annotation AttributeError are unrelated." }, { - "left": "issue:42898", - "right": "issue:45356", - "accept": false, - "reason": "Tokenizer cleanup behavior change and Kimi-K2.5 codec regression are different tokenizer bugs." - }, - { - "left": "issue:43723", - "right": "issue:44568", - "accept": false, - "reason": "AutoTokenizer loading failure and add_special_tokens BOS/EOS omission are distinct issues." - }, - { - "left": "issue:43874", - "right": "issue:45092", - "accept": false, - "reason": "Missing image patch-count API and remote-code meta-init incompatibility are unrelated." - }, - { - "left": "issue:42898", - "right": "issue:44779", + "left": "issue:44188", + "right": "issue:44373", "accept": false, - "reason": "General v5 cleanup behavior change is not the same as the DeepSeek tokenizer regression." + "reason": "Runtime attention-kernel divergence vs docstring wording issue are clearly unrelated." } ] }, @@ -10058,26 +10222,22 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:22355", "issue:28282", "issue:30990", "issue:38175", "issue:41628", + "issue:42757", "issue:42831", - "issue:42886", "issue:42898", - "issue:42994", - "issue:43122", "issue:43295", "issue:43479", - "issue:43576", "issue:43644", "issue:43723", - "issue:43784", - "issue:43824", - "issue:44038", + "issue:43874", + "issue:44060", "issue:44077", "issue:44206", "issue:44222", @@ -10094,23 +10254,31 @@ "issue:44484", "issue:44485", "issue:44514", - "issue:44521", + "issue:44568", "issue:44623", + "issue:44779", "issue:44871", "issue:44913", "issue:44977", "issue:44991", "issue:45092", "issue:45103", + "issue:45125", "issue:45137", "issue:45276", "issue:45310", + "issue:45356", "issue:45357", "issue:45406", "issue:45459" ], "soft_pairs": [ + "issue:44060|issue:45125", + "issue:42898|issue:45356", + "issue:43723|issue:44568", + "issue:42898|issue:44779", "issue:45137|issue:45276", + "issue:43874|issue:45092", "issue:44077|issue:44242", "issue:42898|issue:44448", "issue:41628|issue:45459", @@ -10121,276 +10289,267 @@ "issue:44222|issue:44623", "issue:44206|issue:45103", "issue:42831|issue:44623", - "issue:44297|issue:45310", "issue:22355|issue:28282", - "issue:43723|issue:44977", + "issue:41628|issue:42757", "issue:44483|issue:45406", + "issue:44297|issue:45310", "issue:44222|issue:44484", - "issue:44514|issue:45103", "issue:44263|issue:44360", + "issue:43723|issue:44977", "issue:44263|issue:45092", - "issue:44295|issue:44623", + "issue:44514|issue:45103", "issue:43479|issue:44913", - "issue:44261|issue:45092", + "issue:44295|issue:44623", "issue:43723|issue:44871", - "issue:43723|issue:44442", + "issue:44261|issue:45092", "issue:43723|issue:44991", - "issue:44373|issue:44485", "issue:30990|issue:38175", - "issue:45092|issue:45357", - "issue:41628|issue:43824", - "issue:42886|issue:42994", - "issue:42831|issue:45310", - "issue:43576|issue:43784", - "issue:42831|issue:44038", - "issue:43122|issue:44521", - "issue:44263|issue:44373" + "issue:43723|issue:44442", + "issue:44373|issue:44485", + "issue:45092|issue:45357" ], - "prepared_review_unit_hash": "170be0d0e4c03e29f36b40e5c0b00564fff510db0df9ec07896f2f962842b02d", + "prepared_review_unit_hash": "bb42c38a74e3e90b84e736470e49e5c113d2a3669758073a731f3edb7f914a73", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 35, - "serialized_chars": 37869, - "estimated_input_tokens": 9468, - "estimated_eval_tokens": 19192 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 37224, + "estimated_input_tokens": 9306, + "estimated_eval_tokens": 18868 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a loose collection of Transformers regression issues, with the strongest common thread around v5 tokenizer/processor/loading behavior; there is no clear duplicate PR set. I would use the broad AutoTokenizer loading regression as the cluster representative.", - "confidence": 0.42, - "canonical_issue_reason": "issue:43723 is the broadest and most representative user-facing regression here: AutoTokenizer.from_pretrained loading breaks in v5, which overlaps thematically with several other tokenizer/processor issues in the cluster.", + "summary": "This cluster is mostly heterogeneous. I found one likely duplicate pocket around tokenizer loading regressions in v5; the rest are different bugs touching unrelated code paths.", + "confidence": 0.58, + "canonical_issue_reason": "issue:43723 is the broadest tokenizer-loading regression here and best serves as the umbrella for the one plausible duplicate-style report (44991).", "canonical_pr_reason": null, - "best_issue_reason": "issue:43723 is the best cluster anchor because it names the core v5 loading regression at a high level and can subsume several nearby tokenizer-related reports without being overly narrow.", + "best_issue_reason": "issue:43723 is the strongest global issue candidate because its title is generic enough to represent the shared tokenizer-loading failure pattern without overfitting to a single model.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen3/Qwen3.5 MoE problems: tied-weights warning vs missing _tp_plan for tensor parallelism." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but the symptoms and likely code paths differ: cleanup-space behavior vs codec handling/warning." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer-loading failure vs BOS/EOS insertion bug; different behavior and failure mode." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both involve tokenizer behavior in v5, but one is a cleanup-space regression and the other is incorrect DeepSeek outputs." + }, { "left": "issue:45137", "right": "issue:45276", "accept": false, - "reason": "DeepSpeed ZeRO3 deque failure and Gemma4 resize_token_embeddings behavior are unrelated bugs." + "reason": "Unrelated bugs: DeepSpeed ZeRO3 deque error vs Gemma4 resize_token_embeddings not updating embeddings." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Different multimodal failures: missing image-patch method vs old InternVL2 remote-code/meta-init incompatibility." }, { "left": "issue:44077", "right": "issue:44242", "accept": false, - "reason": "One is a patchtsmixer post_init validation issue; the other is a MoE load-balancing-loss bug." + "reason": "Distinct model logic bugs in different places; no clear shared concrete code-path." }, { "left": "issue:42898", "right": "issue:44448", "accept": false, - "reason": "Both are v5 behavior changes, but one is about clean_up_tokenization_spaces and the other is a model output regression with broader possible causes." + "reason": "Tokenizer-related regression vs Pegasus output mismatch; too broad and not the same bug." }, { "left": "issue:41628", "right": "issue:45459", "accept": false, - "reason": "Importing AutoImageProcessor and masking protobuf tokenizer errors are different failure modes." + "reason": "Import error for AutoImageProcessor vs tokenizer exception masking when protobuf is absent; unrelated failures." }, { "left": "issue:44222", "right": "issue:44297", "accept": false, - "reason": "FP8 save_pretrained for MoE and tokenizer_class metadata mismatch in tokenizer.save_pretrained are not the same bug." + "reason": "Save-pretrained FP8/MoE issue vs tokenizer_class metadata mismatch; different components." }, { "left": "issue:44442", "right": "issue:45459", "accept": false, - "reason": "AutoTokenizer failing on FastSpeech2ConformerTokenizer is distinct from protobuf-related tokenizer exception handling." + "reason": "AutoTokenizer failing to load a specific tokenizer vs protobuf error handling swallowing tokenizer errors; different root causes." }, { "left": "issue:42898", "right": "issue:43295", "accept": false, - "reason": "cleanup_spaces behavior and processor.tokenizer/image-passing regression affect different code paths." + "reason": "Different regressions: clean_up_tokenization_spaces vs processor.tokenizer / image-to-tokenizer regression." }, { "left": "issue:43479", "right": "issue:43644", "accept": false, - "reason": "Default vision/audio config initialization and non-persistent buffer junk are unrelated model-construction issues." + "reason": "Different config initialization bug vs non-persistent buffer corruption; not the same underlying change." }, { "left": "issue:44222", "right": "issue:44623", "accept": false, - "reason": "MoE FP8 save_pretrained and missing processor.save_pretrained files are different persistence bugs." + "reason": "Both mention save_pretrained, but one is FP8 MoE weights and the other is missing processor files." }, { "left": "issue:44206", "right": "issue:45103", "accept": false, - "reason": "FeatureExtractor arg validation and auto_docstring annotation handling are unrelated." + "reason": "LasrFeatureExtractor center-arg crash is unrelated to auto_docstring failures under future annotations." }, { "left": "issue:42831", "right": "issue:44623", "accept": false, - "reason": "FineGrainedFP8 accuracy problems do not match missing processor save files." - }, - { - "left": "issue:44297", - "right": "issue:45310", - "accept": false, - "reason": "Tokenizer class metadata mismatch and Qwen3.5 MoE from_pretrained failure are different regressions." + "reason": "Accuracy issue with FineGrainedFP8 vs processor.save_pretrained missing files; different bug reports." }, { "left": "issue:22355", "right": "issue:28282", "accept": false, - "reason": "Missing transformers.onnx module and missing PyTorch dependency are separate import errors." + "reason": "Different import failures: missing transformers.onnx module vs AutoModel requiring PyTorch." }, { - "left": "issue:43723", - "right": "issue:44977", + "left": "issue:41628", + "right": "issue:42757", "accept": false, - "reason": "Generic tokenizer loading regression and Qwen3.5 flash-attention generation bug are not the same issue." + "reason": "Different missing imports from different packages and different API surfaces." }, { "left": "issue:44483", "right": "issue:45406", "accept": false, - "reason": "Chat/completions request rejection and Gemma4Processor missing _tokenizer are different serving failures." + "reason": "Chat-completions API request rejection vs Gemma4Processor missing _tokenizer; unrelated." }, { - "left": "issue:44222", - "right": "issue:44484", + "left": "issue:44297", + "right": "issue:45310", "accept": false, - "reason": "FP8 MoE save_pretrained and max_shard_size default behavior are unrelated save_pretrained concerns." + "reason": "Tokenizer config save mismatch vs Qwen3.5 MoE from_pretrained error; not the same concrete bug." }, { - "left": "issue:44514", - "right": "issue:45103", + "left": "issue:44222", + "right": "issue:44484", "accept": false, - "reason": "Qwen2_5_VLProcessor batched chat-template crash and auto_docstring annotation crash are unrelated." + "reason": "FP8 save_pretrained bug vs max_shard_size behavior question; no duplicate evidence." }, { "left": "issue:44263", "right": "issue:44360", "accept": false, - "reason": "torch.split return handling and a missing ReLU in the DSA indexer are different implementation defects." + "reason": "Different DSA indexer issues: torch.split return values vs missing ReLU." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Both are tokenizer-ish regressions, but one is loading failure and the other is generation correctness with flash-attention." }, { "left": "issue:44263", "right": "issue:45092", "accept": false, - "reason": "GlmMoeDsaIndexer split handling does not match InternVL2 meta-init incompatibility." + "reason": "Different bugs: torch.split return handling vs remote-code/meta-initialization incompatibility." }, { - "left": "issue:44295", - "right": "issue:44623", + "left": "issue:44514", + "right": "issue:45103", "accept": false, - "reason": "position_ids buffer read error and processor.save_pretrained missing files are separate buffer/save issues." + "reason": "apply_chat_template batching crash is unrelated to auto_docstring annotation handling." }, { "left": "issue:43479", "right": "issue:44913", "accept": false, - "reason": "Phi4MultimodalConfig defaulting and GPTNeoX rotary_pct reload behavior are unrelated config bugs." + "reason": "Config defaulting for multimodal vision/audio settings vs GPTNeoX rotary_pct reload regression." }, { - "left": "issue:44261", - "right": "issue:45092", + "left": "issue:44295", + "right": "issue:44623", "accept": false, - "reason": "MLA layernorm epsilon precision drift and remote-code meta-init incompatibility do not describe the same bug." + "reason": "Buffer read error for position_ids vs processor.save_pretrained missing files; different code paths." }, { "left": "issue:43723", "right": "issue:44871", "accept": false, - "reason": "Tokenizer loading regression and Gemma-3 eos_token_id inconsistency are different configuration/loading problems." + "reason": "Tokenizer load failure vs eos_token_id inconsistency between tokenizer and model config." }, { - "left": "issue:43723", - "right": "issue:44442", + "left": "issue:44261", + "right": "issue:45092", "accept": false, - "reason": "Both are tokenizer-related, but one is a generic AutoTokenizer v5 loading regression and the other is a specific tokenizer class load failure." + "reason": "MLA q_a_layernorm epsilon mismatch vs InternVL2 meta-init incompatibility; unrelated." }, { "left": "issue:43723", "right": "issue:44991", - "accept": false, - "reason": "These are both tokenizer-load complaints in v5, but the affected tokenizer/model and concrete failure are not shown to be the same." - }, - { - "left": "issue:44373", - "right": "issue:44485", - "accept": false, - "reason": "Wrong position_ids docstring and GLM-5 RoPE implementation are not the same underlying defect." + "accept": true, + "reason": "Both describe v5 AutoTokenizer.from_pretrained tokenizer-loading failures; 44991 is a specific model reproduction of the same underlying issue." }, { "left": "issue:30990", "right": "issue:38175", "accept": false, - "reason": "Sentence Transformers hanging on load and zero probabilities in SigLIP2 are unrelated." + "reason": "Sentence Transformers loading hang vs SigLIP2 zero probabilities; different symptoms and likely causes." }, { - "left": "issue:45092", - "right": "issue:45357", - "accept": false, - "reason": "InternVL2 meta-init incompatibility and Qwen3.5 save_pretrained visual key regression are different model-family bugs." - }, - { - "left": "issue:41628", - "right": "issue:43824", - "accept": false, - "reason": "Missing AutoImageProcessor export and missing Qwen2_5_VLForConditionalGeneration export are different missing-symbol issues." - }, - { - "left": "issue:42886", - "right": "issue:42994", - "accept": false, - "reason": "Offline tokenizer cache loading and quantized model saving failure are unrelated." - }, - { - "left": "issue:42831", - "right": "issue:45310", - "accept": false, - "reason": "FineGrainedFP8 accuracy and Qwen3.5 MoE from_pretrained failure are different classes of bug." - }, - { - "left": "issue:43576", - "right": "issue:43784", - "accept": false, - "reason": "Broken transformers env command and sentence-transformers import NameError are unrelated." - }, - { - "left": "issue:42831", - "right": "issue:44038", + "left": "issue:43723", + "right": "issue:44442", "accept": false, - "reason": "FineGrainedFP8 accuracy issues do not match Qwen3-VL-Moe v5 compatibility problems." + "reason": "Both are tokenizer problems, but one is loading failure and the other is AutoTokenizer failing for a specific tokenizer class." }, { - "left": "issue:43122", - "right": "issue:44521", + "left": "issue:44373", + "right": "issue:44485", "accept": false, - "reason": "Tokenization output drift and all-zero assistant masks in multimodal chat templates are different symptoms." + "reason": "Docstring issue for position_ids vs GLM-5 RoPE implementation bug; not duplicates." }, { - "left": "issue:44263", - "right": "issue:44373", + "left": "issue:45092", + "right": "issue:45357", "accept": false, - "reason": "A torch.split bug in the DSA indexer and a wrong docstring are not the same issue." + "reason": "InternVL2 meta-init incompatibility vs Qwen3.5 save_pretrained writing wrong visual encoder keys; different bugs." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet: it treats the cluster as mostly heterogeneous, keeps the tokenizer-loading grouping narrow, and only accepts the 43723/44991 pair where the titles clearly align. The false soft-edge calls are conservative and do not overclaim duplicate status from weak similarity alone." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -10405,7 +10564,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:29127", "issue:30990", @@ -10414,29 +10573,30 @@ "issue:39401", "issue:41628", "issue:42371", + "issue:42757", "issue:42831", "issue:42886", "issue:42907", + "issue:42994", "issue:43066", "issue:43116", "issue:43122", "issue:43232", - "issue:43278", "issue:43408", "issue:43452", "issue:43475", + "issue:43576", + "issue:43784", "issue:43824", "issue:43937", "issue:43976", "issue:43992", "issue:44038", "issue:44117", - "issue:44222", "issue:44261", "issue:44263", "issue:44295", "issue:44297", - "issue:44351", "issue:44360", "issue:44373", "issue:44485", @@ -10449,274 +10609,294 @@ "issue:44779", "issue:44829", "issue:44945", - "issue:44977", - "issue:44987", "issue:45200", "issue:45245", "issue:45278", - "issue:45290" + "issue:45290", + "issue:45310" ], "soft_pairs": [ + "issue:42886|issue:42994", + "issue:41628|issue:43824", + "issue:42831|issue:45310", + "issue:43576|issue:43784", + "issue:42831|issue:44038", + "issue:43122|issue:44521", + "issue:42757|issue:45278", + "issue:44263|issue:44373", "issue:42886|issue:43066", "issue:44117|issue:44704", - "issue:43122|issue:44514", "issue:30990|issue:33357", - "issue:42371|issue:44263", + "issue:43122|issue:44514", + "issue:42907|issue:44945", "issue:44704|issue:45290", "issue:43232|issue:45245", - "issue:42907|issue:44945", - "issue:44295|issue:44297", - "issue:43232|issue:44743", + "issue:42371|issue:44263", "issue:43976|issue:44945", + "issue:43232|issue:44743", + "issue:44295|issue:44297", "issue:39401|issue:44488", "issue:42907|issue:43937", - "issue:41628|issue:45278", - "issue:43116|issue:45200", "issue:43408|issue:43475", "issue:44625|issue:44829", "issue:29127|issue:30990", + "issue:41628|issue:45278", "issue:44360|issue:44485", + "issue:43116|issue:45200", "issue:43992|issue:44704", - "issue:41628|issue:44987", + "issue:39401|issue:44521", "issue:43122|issue:44779", "issue:44038|issue:44297", - "issue:39401|issue:44521", - "issue:44038|issue:44295", "issue:44261|issue:44373", "issue:42371|issue:42831", + "issue:44038|issue:44295", "issue:36296|issue:43116", - "issue:42907|issue:43452", - "issue:44038|issue:44222", - "issue:43824|issue:44351", - "issue:43122|issue:44977", - "issue:43278|issue:44945" + "issue:42907|issue:43452" ], - "prepared_review_unit_hash": "83e5171febf2500164d6a094222cb8a8b5078d47c733bb9b253a3220c5c52594", + "prepared_review_unit_hash": "cfd69270f48ae30b2f6fe9e4a4c61306e652810cdb202081abebb4c3ff3d3dff", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 33, - "serialized_chars": 37075, - "estimated_input_tokens": 9269, - "estimated_eval_tokens": 18794 + "soft_pair_count": 36, + "serialized_chars": 37992, + "estimated_input_tokens": 9498, + "estimated_eval_tokens": 19252 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These items do not form a duplicate cluster; the soft pairs are mostly superficial overlaps around Transformers APIs but describe different bugs, regressions, or models. No PRs are present.", - "confidence": 0.93, - "canonical_issue_reason": "issue:43824 is the most concrete, widely referenced issue here: it has a clear import failure, strong user impact, and the highest inbound reference/activity signal among the set.", + "summary": "The set is mostly a grab bag of unrelated issues; a few share broad themes like imports, tokenizers, or generation, but none of the soft edges look like the same concrete bug/change. I would keep them separate.", + "confidence": 0.37, + "canonical_issue_reason": "No single issue clearly subsumes the others. If one representative is needed, issue:45278 is the broadest umbrella-style report about post-upgrade import failures, but it is not a true duplicate hub.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43824 is the best standalone triage anchor because it is specific, reproducible, and already drew follow-up references from other reports.", + "best_issue_reason": "issue:45278 is the most umbrella-like and still-open report, so it is the best single issue to represent the cluster if one must be chosen.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:42886", - "right": "issue:43066", + "right": "issue:42994", "accept": false, - "reason": "Different failures: offline cache tokenization load vs wrong tokenizer decoder type in v5." + "reason": "Different failures: offline cache loading vs quantized model saving." }, { - "left": "issue:44117", - "right": "issue:44704", + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different symbols and different model families." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "Accuracy regression in FP8 vs Qwen3.5-MoE from_pretrained error; not the same bug." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken env command vs sentence-transformers import NameError; unrelated code paths." + }, + { + "left": "issue:42831", + "right": "issue:44038", "accept": false, - "reason": "Both involve loading paths, but one is a tokenizer mapping None bug and the other is AutoProcessor not forwarding kwargs to cached_file." + "reason": "Different model-specific issues: FP8 accuracy vs Qwen3-VL-MoE bug." }, { "left": "issue:43122", - "right": "issue:44514", + "right": "issue:44521", "accept": false, - "reason": "Different code paths: tokenization output changed across versions vs batched apply_chat_template crash with padding=False." + "reason": "Tokenizer regression vs multimodal assistant mask bug; only a broad tokenizer theme overlaps." }, { - "left": "issue:30990", - "right": "issue:33357", + "left": "issue:42757", + "right": "issue:45278", "accept": false, - "reason": "Unrelated symptoms: Sentence Transformers loading hang vs MacOS bus error on a CLIP model." + "reason": "Specific missing symbol vs broad post-upgrade import failures; not the same concrete defect." }, { - "left": "issue:42371", - "right": "issue:44263", + "left": "issue:44263", + "right": "issue:44373", "accept": false, - "reason": "TF32 behavior settings and a torch.split return-value bug are unrelated." + "reason": "GLM-MoE indexer behavior vs a docstring issue; different type of problem." }, { - "left": "issue:44704", - "right": "issue:45290", + "left": "issue:42886", + "right": "issue:43066", "accept": false, - "reason": "AutoProcessor kwargs forwarding and chat-template tool-call handling are different bugs." + "reason": "Offline cache load failure vs wrong tokenizer decoder type in v5." }, { - "left": "issue:43232", - "right": "issue:45245", + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Mapping-name fallback bug vs AutoProcessor kwarg propagation; unrelated paths." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading hang vs MacOS bus error; both runtime failures but not the same bug." + }, + { + "left": "issue:43122", + "right": "issue:44514", "accept": false, - "reason": "Generation state sync issue vs category cardinality runtime error; no shared underlying defect." + "reason": "Tokenization regression vs batched chat-template crash; different surfaces and causes." }, { "left": "issue:42907", "right": "issue:44945", "accept": false, - "reason": "Saving dequantized models and incorrect pipeline-parallel output are different problems." + "reason": "Quantized model save failure vs pipeline parallelism output corruption; not one concrete issue." }, { - "left": "issue:44295", - "right": "issue:44297", + "left": "issue:44704", + "right": "issue:45290", "accept": false, - "reason": "Reading a buffer-backed position_ids tensor is unrelated to tokenizer_class metadata written by save_pretrained." + "reason": "Processor kwarg forwarding vs chat-template crash on tool-call messages; different code paths." }, { "left": "issue:43232", - "right": "issue:44743", + "right": "issue:45245", + "accept": false, + "reason": "Generation state handling vs category-count runtime limit; unrelated." + }, + { + "left": "issue:42371", + "right": "issue:44263", "accept": false, - "reason": "Both are generation-related, but one is sync_gpus/model kwargs handling and the other is recurrent-state reset with cache." + "reason": "TF32 configuration guidance vs GLM indexer split behavior; no overlap beyond being model-adjacent." }, { "left": "issue:43976", "right": "issue:44945", "accept": false, - "reason": "Python version compatibility issue vs pipeline parallelism output correctness; not the same bug." + "reason": "Python version compatibility vs pipeline parallelism output issue; different regressions." }, { - "left": "issue:39401", - "right": "issue:44488", + "left": "issue:43232", + "right": "issue:44743", "accept": false, - "reason": "Qwen3 offset_mapping regression and failure to load a specific tokenizer are distinct." + "reason": "Both are generation-related, but one is sync_gpus kwargs and the other is Qwen3.5 recurrent-state reset." }, { - "left": "issue:42907", - "right": "issue:43937", + "left": "issue:44295", + "right": "issue:44297", "accept": false, - "reason": "Dequantized save failure and invalid GenerationConfig for GLM-5 are different failure modes." + "reason": "Buffer access error vs tokenizer_class metadata mismatch; separate tokenizer/model save issues." }, { - "left": "issue:41628", - "right": "issue:45278", + "left": "issue:39401", + "right": "issue:44488", "accept": false, - "reason": "Specific AutoImageProcessor import error vs broad import errors after upgrade; too generic to merge." + "reason": "Wrong offset mapping vs failure to load a specific model; different symptoms." }, - { - "left": "issue:43116", - "right": "issue:45200", + { + "left": "issue:42907", + "right": "issue:43937", "accept": false, - "reason": "Multi-label example script returning empty results is unrelated to Gemma 4 mm_token_type_ids defaults." + "reason": "Saving dequantized models vs invalid GenerationConfig; unrelated failures." }, { "left": "issue:43408", "right": "issue:43475", "accept": false, - "reason": "Same family (SAM 3), but one is a model-type warning and the other is a missing attribute in vision encoder output." + "reason": "Model-type warning vs missing attribute on a vision output object; related family, different bug." }, { "left": "issue:44625", "right": "issue:44829", "accept": false, - "reason": "num_labels propagation and flash_attention_3 degenerate training are separate model bugs." + "reason": "num_labels propagation vs flash_attention_3 training degeneration; no shared concrete code path." }, { "left": "issue:29127", "right": "issue:30990", "accept": false, - "reason": "LayoutLMv3 error-message clarity is unrelated to a loading hang in Sentence Transformers." + "reason": "LayoutLMv3 error-message clarity vs sentence-transformers loading hang; unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Specific AutoImageProcessor import error vs broad upgrade-wide import failures; not the same issue." }, { "left": "issue:44360", "right": "issue:44485", "accept": false, - "reason": "DSA indexer activation issue and GLM-5 RoPE implementation are different subsystems and defects." + "reason": "DSA indexer lacking ReLU vs GLM-5 RoPE implementation; different model internals." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example-script multi-label output bug vs Gemma 4 token-type defaulting; different behaviors." }, { "left": "issue:43992", "right": "issue:44704", "accept": false, - "reason": "Missing embed_tokens.weight on UMT5Encoder.from_pretrained is not the same as AutoProcessor kwargs forwarding." + "reason": "Missing encoder weights in from_pretrained vs cached_file kwarg forwarding; different loader bugs." }, { - "left": "issue:41628", - "right": "issue:44987", + "left": "issue:39401", + "right": "issue:44514", "accept": false, - "reason": "Both are import/load regressions, but they affect different artifacts and code paths." + "reason": "Tokenizer offset mapping regression vs batched chat-template crash; same general area, not same defect." }, { "left": "issue:43122", "right": "issue:44779", "accept": false, - "reason": "Tokenizer regression across versions and Deepseek tokenizer wrong results are not shown to stem from the same bug." + "reason": "Two tokenizer regressions, but for different models and different failure modes." }, { "left": "issue:44038", "right": "issue:44297", "accept": false, - "reason": "Qwen3-VL-Moe loading bug and tokenizer save metadata mismatch are unrelated." - }, - { - "left": "issue:39401", - "right": "issue:44521", - "accept": false, - "reason": "Offset mapping regression and multimodal assistant_mask all-zero behavior are different tokenizer/template bugs." - }, - { - "left": "issue:44038", - "right": "issue:44295", - "accept": false, - "reason": "Qwen3-VL-Moe issue and position_ids buffer access error do not share a concrete code path." + "reason": "Qwen3-VL-MoE bug vs tokenizer save metadata mismatch; unrelated." }, { "left": "issue:44261", "right": "issue:44373", "accept": false, - "reason": "Missing rms_norm_eps in MLA config is not the same as a wrong position_ids docstring." + "reason": "Precision issue from missing eps config vs wrong docstring; not duplicates." }, { "left": "issue:42371", "right": "issue:42831", "accept": false, - "reason": "TF32 behavior API change and FineGrainedFP8 accuracy regression are unrelated." - }, - { - "left": "issue:36296", - "right": "issue:43116", - "accept": false, - "reason": "Tensor parallel training bug and multi-label classification example failure are different problems." - }, - { - "left": "issue:42907", - "right": "issue:43452", - "accept": false, - "reason": "Dequantized save failure and gguf_file breakage in from_pretrained are separate loader/saver issues." + "reason": "TF32 API change guidance vs FP8 accuracy regression; different concerns." }, { "left": "issue:44038", - "right": "issue:44222", - "accept": false, - "reason": "Qwen3-VL-Moe bug and FP8 save_pretrained moe bug are distinct issues." - }, - { - "left": "issue:43824", - "right": "issue:44351", + "right": "issue:44295", "accept": false, - "reason": "Different missing imports: Qwen2_5_VLForConditionalGeneration versus HybridCache." + "reason": "Model bug in Qwen3-VL-MoE vs position_ids buffer read error; no concrete overlap." }, { - "left": "issue:43122", - "right": "issue:44977", + "left": "issue:36296", + "right": "issue:43116", "accept": false, - "reason": "Tokenizer behavior change across versions is not the same as Qwen3.5 flash-attention generation failure." + "reason": "Tensor-parallel training bug vs example-script classification output issue; different subsystems." }, { - "left": "issue:43278", - "right": "issue:44945", + "left": "issue:42907", + "right": "issue:43452", "accept": false, - "reason": "Training/evaluate dtype mismatch and pipeline-parallel incorrect output are different defects." + "reason": "Saving dequantized models vs gguf_file from_pretrained loading breakage; separate loader/saver bugs." } ] }, @@ -10738,27 +10918,29 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:22355", "issue:29127", "issue:33453", "issue:38617", "issue:39401", - "issue:41628", "issue:42222", "issue:42371", "issue:42673", + "issue:42757", "issue:42907", "issue:43064", "issue:43116", - "issue:43421", + "issue:43122", + "issue:43278", "issue:43441", "issue:43450", "issue:43493", "issue:43550", "issue:43673", "issue:43784", + "issue:43824", "issue:43825", "issue:43827", "issue:43874", @@ -10771,7 +10953,7 @@ "issue:44075", "issue:44222", "issue:44291", - "issue:44484", + "issue:44351", "issue:44485", "issue:44779", "issue:44912", @@ -10784,247 +10966,270 @@ "issue:45083", "issue:45216", "issue:45278", - "issue:45310", - "issue:45405", "issue:45446", "issue:45478" ], "soft_pairs": [ + "issue:44038|issue:44222", + "issue:43824|issue:44351", + "issue:43278|issue:44945", + "issue:43122|issue:44977", "issue:45216|issue:45478", "issue:42907|issue:44222", - "issue:43784|issue:45446", + "issue:42757|issue:44351", + "issue:42907|issue:44912", "issue:42222|issue:43441", "issue:42673|issue:45072", - "issue:42907|issue:44912", - "issue:38617|issue:45278", - "issue:43784|issue:44933", + "issue:43784|issue:45446", "issue:43064|issue:43673", - "issue:43550|issue:43784", + "issue:43784|issue:44933", "issue:43937|issue:44485", - "issue:44075|issue:44936", + "issue:43550|issue:43784", + "issue:38617|issue:45278", "issue:22355|issue:29127", - "issue:42371|issue:43874", "issue:43116|issue:43493", + "issue:44075|issue:44936", + "issue:42371|issue:43874", "issue:43116|issue:43825", "issue:44222|issue:45478", - "issue:43976|issue:45446", "issue:44038|issue:45478", + "issue:43976|issue:45446", "issue:39401|issue:44779", - "issue:33453|issue:39401", "issue:43931|issue:44945", + "issue:33453|issue:39401", "issue:45083|issue:45446", "issue:43937|issue:44977", - "issue:42673|issue:43906", "issue:43827|issue:45446", - "issue:43450|issue:43986", - "issue:45405|issue:45446", + "issue:42673|issue:43906", "issue:44291|issue:45071", - "issue:39401|issue:43421", - "issue:38617|issue:41628", - "issue:44484|issue:45310" + "issue:43450|issue:43986" ], - "prepared_review_unit_hash": "4634d34c9da31f24f1544e3ad62650fb454a816d4de3a901109189f601904445", + "prepared_review_unit_hash": "6f5d8ccbbeaa83f67240a6095bfd8793e601534243110a06a4a600f8a610e99a", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 36896, - "estimated_input_tokens": 9224, - "estimated_eval_tokens": 18704 + "soft_pair_count": 33, + "serialized_chars": 37167, + "estimated_input_tokens": 9292, + "estimated_eval_tokens": 18840 }, "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 regression reports (imports, tokenizer/model bugs, generation issues, training issues). The broadest and most representative open item is the general import-regression report, not any of the model-specific cases.", - "confidence": 0.86, - "canonical_issue_reason": "issue:45278 is the broadest open report and best matches the cluster\u2019s recurring theme of v5 regression/import breakage across packages; it is also the most general and therefore the best anchor issue.", + "summary": "This is a noisy, heterogeneous cluster: import errors, tokenizer regressions, model-specific generation/loading bugs, training issues, docs, and quantization problems. I found no soft pair that looks like the same underlying bug/change.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue for the whole cluster; the items span unrelated failure modes and model families.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the best overall issue to keep as the cluster representative because it is current, broad in scope, and likely to absorb the surrounding import-regression complaints better than the model-specific or narrow bug reports.", + "best_issue_reason": "If forced to pick an anchor, issue 45278 is the broadest umbrella for the import-regression subset, but it is still not a good canonical target for the entire cluster.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Both mention MoE/Qwen, but one is a v5 FlashAttention/generation bug and the other is FP8 save_pretrained behavior." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they are for different symbols and code paths (`Qwen2_5_VLForConditionalGeneration` vs `HybridCache`)." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Different problems: dtype drift between train/eval vs incorrect output under pipeline parallelism." + }, + { + "left": "issue:43122", + "right": "issue:44977", + "accept": false, + "reason": "Both are v5 regressions, but tokenizer behavior and FlashAttention generation are different bugs." + }, { "left": "issue:45216", "right": "issue:45478", "accept": false, - "reason": "Same model family and version range, but one is a save_pretrained regression and the other is a from_pretrained load error; different concrete failure paths." + "reason": "Both involve Qwen3.5 and save/load, but one is a bad checkpoint from save_pretrained and the other is a from_pretrained load failure." }, { "left": "issue:42907", "right": "issue:44222", "accept": false, - "reason": "Both mention saving/quantization, but they affect different models and different save-time behaviors; not the same underlying bug." + "reason": "Both touch saving quantized models, but they concern different model families and different serialization paths." }, { - "left": "issue:43784", - "right": "issue:45446", + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import failures, but for unrelated packages (`huggingface_hub` vs `transformers`) and different missing names." + }, + { + "left": "issue:42907", + "right": "issue:44912", "accept": false, - "reason": "Unrelated import failures in different modules and on different code paths." + "reason": "Saving dequantized Ministral/Devstral models is unrelated to MXFP4 loading falling back to bf16." }, { "left": "issue:42222", "right": "issue:43441", "accept": false, - "reason": "Different subsystems and symptoms: vitpose model breakage vs FlashAttention failure in Ministral-3." + "reason": "Different model-specific regressions: vitpose breakage vs Ministral-3 FlashAttention failure." }, { "left": "issue:42673", "right": "issue:45072", "accept": false, - "reason": "VRAM leak in dataloader threads is unrelated to bfloat16 dtype mismatches during inference." + "reason": "VRAM leakage in Qwen3ForCausalLM dataloader threads is unrelated to dtype mismatches in inference." }, { - "left": "issue:42907", - "right": "issue:44912", + "left": "issue:43784", + "right": "issue:45446", "accept": false, - "reason": "Both involve loading/saving or quantization broadly, but they describe different models and different failure modes; too loose to be duplicates." + "reason": "One is a missing import in sentence-transformers integration; the other is a PyTorch version check bug for flex_attention." }, { - "left": "issue:38617", - "right": "issue:45278", + "left": "issue:43064", + "right": "issue:43673", "accept": false, - "reason": "Both are import-related, but one is a specific missing symbol while the other is a broad post-upgrade import regression; not clearly the same bug." + "reason": "Trainer/FSDP optimizer state corruption is unrelated to missing GenerationMixin cache during chunked prefill." }, { "left": "issue:43784", "right": "issue:44933", "accept": false, - "reason": "Both are import errors, but they target different missing imports and likely different root causes." + "reason": "Both are import-related, but they are different missing exports from different modules." }, { - "left": "issue:43064", - "right": "issue:43673", + "left": "issue:43937", + "right": "issue:44485", "accept": false, - "reason": "Training-state corruption under FSDP2/PEFT is unrelated to generation cache handling in chunked_prefill." + "reason": "GenerationConfig validation and GLM-5 RoPE implementation are distinct issues." }, { "left": "issue:43550", "right": "issue:43784", "accept": false, - "reason": "torch.compile/SDPA model failure is unrelated to an import-time NameError." + "reason": "torch.compile/SDPA failure is unrelated to a missing `nn` import." }, { - "left": "issue:43937", - "right": "issue:44485", + "left": "issue:38617", + "right": "issue:45278", "accept": false, - "reason": "Same model family, but invalid GenerationConfig and RoPE implementation are different concrete problems." + "reason": "A specific missing `layer_type_validation` import does not match the broader 'many import errors' report." }, { - "left": "issue:44075", - "right": "issue:44936", + "left": "issue:22355", + "right": "issue:29127", "accept": false, - "reason": "Optimizer argument handling and post-train evaluate failure are different issues." + "reason": "No module named transformers.onnx is unrelated to the LayoutLMv3 error-message clarity issue." }, { - "left": "issue:22355", - "right": "issue:29127", + "left": "issue:43116", + "right": "issue:43493", "accept": false, - "reason": "No meaningful overlap beyond both being bug reports." + "reason": "Multi-label classification output parsing and SigLIP2 HF-vs-JAX divergence are different bugs." }, { - "left": "issue:42371", - "right": "issue:43874", + "left": "issue:44075", + "right": "issue:44936", "accept": false, - "reason": "TF32 API guidance and a missing image-processor method are unrelated." + "reason": "Unused SGD args and evaluate-after-train failure do not share a concrete code path." }, { - "left": "issue:43116", - "right": "issue:43493", + "left": "issue:42371", + "right": "issue:43874", "accept": false, - "reason": "Multi-label classification empty outputs and SigLIP2 implementation discrepancy are different bugs." + "reason": "TF32 settings guidance is unrelated to a missing image-processor method in GLM46V." }, { "left": "issue:43116", "right": "issue:43825", "accept": false, - "reason": "Example-script classification output bug is unrelated to pipeline() translation-task messaging." + "reason": "Example script behavior and pipeline() error messaging are separate issues." }, { "left": "issue:44222", "right": "issue:45478", "accept": false, - "reason": "Both mention Moe/model saving/loading, but one is FP8 save_pretrained and the other is a from_pretrained load error; not the same concrete bug." + "reason": "Both mention Qwen3.5/MoE, but one is FP8 save_pretrained and the other is a from_pretrained load error." }, { - "left": "issue:43976", - "right": "issue:45446", + "left": "issue:44038", + "right": "issue:45478", "accept": false, - "reason": "Different version-check problems affecting different imports and modules." + "reason": "Qwen3.5 FlashAttention generation failure is a different code path from Qwen3.5 MoE loading failure." }, { - "left": "issue:44038", - "right": "issue:45478", + "left": "issue:43976", + "right": "issue:45446", "accept": false, - "reason": "Both involve Qwen MoE variants, but one is a general v5/Qwen3-VL-Moe bug and the other is a specific Qwen3.5 from_pretrained regression." + "reason": "Python version support regression is unrelated to the flex_attention import/version-check bug." }, { "left": "issue:39401", "right": "issue:44779", "accept": false, - "reason": "Both are tokenizer correctness issues, but they affect different tokenizers and different symptoms." + "reason": "Both are tokenizer regressions, but they affect different tokenizers and different outputs." }, { - "left": "issue:33453", - "right": "issue:39401", + "left": "issue:43931", + "right": "issue:44945", "accept": false, - "reason": "Regression in tokenizer loading is too broad to be the same as a specific offset_mapping bug." + "reason": "Shape mismatch while loading Qwen3-VL is unrelated to incorrect LLM output under pipeline parallelism." }, { - "left": "issue:43931", - "right": "issue:44945", + "left": "issue:33453", + "right": "issue:39401", "accept": false, - "reason": "Model weight-shape loading error and incorrect output under pipeline parallelism are different failures." + "reason": "Tokenizer loading regression and wrong offset mappings are different failure modes." }, { "left": "issue:45083", "right": "issue:45446", "accept": false, - "reason": "A helper-length calculation bug in qwen3_omni_moe is unrelated to a flex_attention import version check." + "reason": "A multimodal feature-length helper bug is unrelated to the flex_attention import/version issue." }, { "left": "issue:43937", "right": "issue:44977", "accept": false, - "reason": "Both concern Qwen3.5 generation, but invalid generation config and flash-attention generation failure are distinct causes." - }, - { - "left": "issue:42673", - "right": "issue:43906", - "accept": false, - "reason": "The second item is only a reproduction reference to another issue; there is no evidence it is the same VRAM leak." + "reason": "GLM-5 invalid GenerationConfig and Qwen3.5 FlashAttention generation failures are different bugs." }, { "left": "issue:43827", "right": "issue:45446", "accept": false, - "reason": "Documentation references to pipeline() and a PyTorch import version check are unrelated." + "reason": "Docs still referencing pipeline() is unrelated to the AuxRequest import/version-check issue." }, { - "left": "issue:43450", - "right": "issue:43986", + "left": "issue:42673", + "right": "issue:43906", "accept": false, - "reason": "Video batch shape handling and AutoProcessor crash without torchvision are different problems." + "reason": "The isolated reproduction of issue 38071 is not the same bug as Qwen3ForCausalLM VRAM leakage." }, { - "left": "issue:45405", - "right": "issue:45446", + "left": "issue:44291", + "right": "issue:45071", "accept": false, - "reason": "A dependency version bump and a PyTorch import guard bug are unrelated." + "reason": "Unexpected `_is_hf_initialized` during init_empty_weights is a different code path from PretrainedConfig type checking." }, { - "left": "issue:44291", - "right": "issue:45071", + "left": "issue:43450", + "right": "issue:43986", "accept": false, - "reason": "init_empty_weights/_is_hf_initialized loading error and PretrainedConfig type checking are distinct regressions." + "reason": "Batched video processor shape errors and AutoProcessor crashes without torchvision are different problems." } ] }, @@ -11046,9 +11251,10 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:38617", + "issue:39401", "issue:41628", "issue:42175", "issue:42222", @@ -11060,8 +11266,8 @@ "issue:43066", "issue:43097", "issue:43122", - "issue:43335", "issue:43408", + "issue:43421", "issue:43441", "issue:43452", "issue:43577", @@ -11069,7 +11275,6 @@ "issue:43673", "issue:43742", "issue:43874", - "issue:43931", "issue:43937", "issue:43986", "issue:44117", @@ -11084,8 +11289,6 @@ "issue:44496", "issue:44829", "issue:44933", - "issue:44964", - "issue:44987", "issue:45003", "issue:45071", "issue:45081", @@ -11093,285 +11296,276 @@ "issue:45278", "issue:45310", "issue:45341", + "issue:45405", + "issue:45446", "issue:45459", "issue:45464", "issue:45478" ], "soft_pairs": [ + "issue:45405|issue:45446", + "issue:44484|issue:45310", + "issue:39401|issue:43421", "issue:44829|issue:45216", - "issue:42673|issue:43577", + "issue:38617|issue:41628", "issue:44117|issue:45081", "issue:42673|issue:43582", - "issue:38617|issue:44220", - "issue:42175|issue:42222", + "issue:42673|issue:43577", "issue:44410|issue:45310", "issue:43408|issue:43742", - "issue:42371|issue:44261", + "issue:38617|issue:44220", "issue:43937|issue:44496", - "issue:44933|issue:45278", + "issue:42371|issue:44261", "issue:44483|issue:45464", + "issue:42175|issue:42222", "issue:42617|issue:45341", - "issue:43066|issue:43122", + "issue:44933|issue:45278", "issue:44484|issue:45478", - "issue:44964|issue:45278", - "issue:42222|issue:45459", + "issue:43066|issue:43122", "issue:43673|issue:45341", "issue:43097|issue:44493", + "issue:42222|issue:45459", "issue:43874|issue:44261", - "issue:43986|issue:44484", - "issue:43441|issue:44829", "issue:44410|issue:45478", - "issue:42222|issue:42994", + "issue:43441|issue:44829", + "issue:43986|issue:44484", "issue:43012|issue:43408", + "issue:42222|issue:42994", "issue:44263|issue:44485", "issue:41628|issue:45071", "issue:45003|issue:45341", - "issue:43097|issue:43452", - "issue:44117|issue:44987", - "issue:43986|issue:45341", - "issue:43931|issue:44410", - "issue:43874|issue:44263", - "issue:43335|issue:45071" + "issue:43097|issue:43452" ], - "prepared_review_unit_hash": "5f20c3d29837d1fd1c7d5d429075a63482e094eccdeb8bb198c84e7e0629369c", + "prepared_review_unit_hash": "3c8f8fcafa03e7dce582abc188ad29ecc04ff38aa7aab39801acdb3d53bebbf0", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37648, - "estimated_input_tokens": 9412, - "estimated_eval_tokens": 19080 + "soft_pair_count": 32, + "serialized_chars": 36992, + "estimated_input_tokens": 9248, + "estimated_eval_tokens": 18752 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a set of unrelated Transformers issues with similar surface themes (imports, model loading, version regressions). I found no soft pair that clearly describes the same underlying bug/change, so I would not merge any of them as duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "No single issue is a reliable canonical for the whole set; the items span multiple unrelated regressions and model-specific bugs.", + "summary": "This cluster is mostly a grab-bag of unrelated Transformers regressions and import/tokenizer/model-loading bugs; there is one likely true duplicate pair around tokenizer behavior in v5, but the rest are too semantically different to merge.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella report in the set: it summarizes many import errors after upgrading to 5.5.0, making it the best single issue to anchor nearby import-regression reports.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the broadest umbrella report for the import-regression subset, but it is too generic to serve as a true canonical for this mixed cluster.", + "best_issue_reason": "45278 is the most representative global issue because it is open, recent, and broad enough to cover the cluster\u2019s import-regression theme better than the many narrow one-off bug reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "Different subsystems: PEFT version bump vs a PyTorch version check bug in flex_attention." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "Both mention model loading/versioning, but one is about shard size in save_pretrained and the other is a Qwen3.5 MoE from_pretrained regression." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Both involve tokenizers, but one is a wrong offset_mapping bug and the other is runtime special-token post-processor updates." + }, { "left": "issue:44829", "right": "issue:45216", "accept": false, - "reason": "FlashAttention training degeneration and Qwen3.5 save_pretrained regression are different bugs affecting different code paths." + "reason": "FlashAttention training degeneration and Qwen3.5 checkpoint saving are different code paths and failure modes." }, { - "left": "issue:42673", - "right": "issue:43577", + "left": "issue:38617", + "right": "issue:41628", "accept": false, - "reason": "VRAM leakage in multithreaded dataloaders is unrelated to Blip2 dtype staying float32 on load." + "reason": "Both are import errors, but they concern different missing symbols and different API surfaces." }, { "left": "issue:44117", "right": "issue:45081", "accept": false, - "reason": "Tokenizer mapping returning None and Mistral regex patch crashing are separate tokenizer-loading failures." + "reason": "One is a tokenizer mapping None-handling bug; the other is a Mistral regex patch crash on tokenizer backend access." }, { "left": "issue:42673", "right": "issue:43582", "accept": false, - "reason": "A VRAM leak and an Apple Silicon warmup TypeError are not the same defect." - }, - { - "left": "issue:38617", - "right": "issue:44220", - "accept": false, - "reason": "Missing `layer_type_validation` import and `_torch_extract_fbank_features()` failure involve different modules and failure modes." + "reason": "VRAM leakage in dataloader threads is unrelated to an Apple Silicon TypeError in caching allocator warmup." }, { - "left": "issue:42175", - "right": "issue:42222", + "left": "issue:42673", + "right": "issue:43577", "accept": false, - "reason": "Package extra dependency omission is unrelated to broken vitpose model loading." + "reason": "A VRAM leak bug and a dtype propagation bug in Blip2 loading are unrelated." }, { "left": "issue:44410", "right": "issue:45310", "accept": false, - "reason": "Missing qwen3next projections and Qwen3.5 MoE from_pretrained errors are different model/config regressions." + "reason": "Both are Qwen-family reports, but one is missing projections in qwen3next and the other is a MoE from_pretrained error." }, { "left": "issue:43408", "right": "issue:43742", "accept": false, - "reason": "SAM3 tracker/video warning and MobileLLM key error are unrelated model-loading issues." + "reason": "A model-type mismatch warning for sam3_tracker is not the same bug as a KeyError loading MobileLLM-125M." }, { - "left": "issue:42371", - "right": "issue:44261", + "left": "issue:38617", + "right": "issue:44220", "accept": false, - "reason": "TF32 API warning behavior is unrelated to MLA q_a_layernorm precision/config issues." + "reason": "An import error for layer_type_validation is unrelated to _torch_extract_fbank_features() behavior." }, { "left": "issue:43937", "right": "issue:44496", "accept": false, - "reason": "GLM-5 generation-config validation and Olmo-Hybrid model recognition are distinct problems." + "reason": "GenerationConfig validation failure and unrecognized-model config loading are separate problems." }, { - "left": "issue:44933", - "right": "issue:45278", + "left": "issue:42371", + "right": "issue:44261", "accept": false, - "reason": "A specific missing import from `image_utils` is not the same as a broad upgrade-time import-error report." + "reason": "TF32 configuration guidance is unrelated to MLA q_a_layernorm missing rms_norm_eps." }, { "left": "issue:44483", "right": "issue:45464", "accept": false, - "reason": "Chat API failure on v5.3 and Qwen3.5 streaming inference failure are different request-path bugs." + "reason": "Both concern chat/completions, but one is a general API request rejection and the other is a streaming inference failure on a specific model." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow dependency packaging and vitpose model breakage are not the same underlying issue." }, { "left": "issue:42617", "right": "issue:45341", "accept": false, - "reason": "`3d_parallel.py` runtime failure and a `testing_utils.py` bug are unrelated." + "reason": "A 3d_parallel.py runtime issue and a testing_utils bug are unrelated." }, { - "left": "issue:43066", - "right": "issue:43122", + "left": "issue:44933", + "right": "issue:45278", "accept": false, - "reason": "Tokenizer decoder-type mismatch and tokenization changes across versions are related only superficially, not the same concrete bug." + "reason": "A specific missing import from image_utils is not the same concrete bug as a broad report about many import errors after upgrade." }, { "left": "issue:44484", "right": "issue:45478", "accept": false, - "reason": "`max_shard_size` default behavior and Qwen3.5 MoE from_pretrained errors are unrelated." - }, - { - "left": "issue:44964", - "right": "issue:45278", - "accept": false, - "reason": "Phi-4 multimodal model loading failure is not the same issue as a broad import-error regression report." + "reason": "Both are Qwen3.5/MoE-related only in a broad sense; one is about shard size, the other about from_pretrained loading failure." }, { - "left": "issue:42222", - "right": "issue:45459", - "accept": false, - "reason": "Broken vitpose models and tokenizer error masking without protobuf are different areas and symptoms." + "left": "issue:43066", + "right": "issue:43122", + "accept": true, + "reason": "These both describe the same tokenizer regression in v5: tokenization output changes because the decoder/type handling is wrong." }, { "left": "issue:43673", "right": "issue:45341", "accept": false, - "reason": "Generation cache missing in v5 chunked prefill does not match a testing utility bug." + "reason": "Generation cache missing in chunked_prefill is unrelated to a minor testing utility bug." }, { "left": "issue:43097", "right": "issue:44493", "accept": false, - "reason": "Removal of `tie_embeddings_and_encoder_decoder` and unexpected position-id keys are separate v5 regressions." + "reason": "A removed config field warning and unexpected position-id key errors are different regression classes." + }, + { + "left": "issue:42222", + "right": "issue:45459", + "accept": false, + "reason": "Vitpose model breakage and tokenizer error masking when protobuf is absent are unrelated." }, { "left": "issue:43874", "right": "issue:44261", "accept": false, - "reason": "Missing image-patch counting method and MLA norm-eps precision error are unrelated multimodal/model bugs." + "reason": "A missing image-patch count method and a missing rms_norm_eps config value are different model/component bugs." }, { - "left": "issue:43986", - "right": "issue:44484", + "left": "issue:44410", + "right": "issue:45478", "accept": false, - "reason": "AutoProcessor crash without torchvision and `max_shard_size` behavior are unrelated." + "reason": "Missing projections in qwen3next is not the same concrete defect as a Qwen3.5 MoE from_pretrained error." }, { "left": "issue:43441", "right": "issue:44829", "accept": false, - "reason": "FlashAttention failure for Ministral-3 and FlashAttention-3 training degeneration are different concrete bugs." + "reason": "Both mention FlashAttention, but they target different models and distinct failures, so they should not be merged." }, { - "left": "issue:44410", - "right": "issue:45478", + "left": "issue:43986", + "right": "issue:44484", "accept": false, - "reason": "qwen3next projection omission and Qwen3.5 MoE loading regression are not the same code-path failure." + "reason": "AutoProcessor/video-model torchvision crash is unrelated to shard-size behavior in save_pretrained." }, { - "left": "issue:42222", - "right": "issue:42994", + "left": "issue:43012", + "right": "issue:43408", "accept": false, - "reason": "Vitpose breakage and quantized model saving failure do not share the same underlying issue." + "reason": "A bfloat16 compilation warning and a sam3 model-type warning are unrelated." }, { - "left": "issue:43012", - "right": "issue:43408", + "left": "issue:42222", + "right": "issue:42994", "accept": false, - "reason": "BF16 compile warning and SAM3 type warning are unrelated warning/regression reports." + "reason": "Vitpose model breakage and quantized model saving failure are different bugs." }, { "left": "issue:44263", "right": "issue:44485", "accept": false, - "reason": "`torch.split` return handling and GLM-5 RoPE implementation are different GLM-family problems." + "reason": "A torch.split return-value issue and a GLM-5 RoPE implementation discussion are not the same change." }, { "left": "issue:41628", "right": "issue:45071", "accept": false, - "reason": "Importing `AutoImageProcessor` and PretrainedConfig type-checking are unrelated breakages." + "reason": "An AutoImageProcessor import failure is different from a PreTrainedConfig type-checking regression." }, { "left": "issue:45003", "right": "issue:45341", "accept": false, - "reason": "`sys.modules[]` access in modeling_utils is unrelated to a testing utility bug." + "reason": "Unsafe sys.modules access in modeling_utils is unrelated to a testing_utils bug." }, { "left": "issue:43097", "right": "issue:43452", "accept": false, - "reason": "A removed config flag and gguf_file loading breakage are different regressions." - }, - { - "left": "issue:44117", - "right": "issue:44987", - "accept": false, - "reason": "Tokenizer mapping returning None and loading `physical-intelligence/fast` failing are not clearly the same defect." - }, - { - "left": "issue:43986", - "right": "issue:45341", - "accept": false, - "reason": "AutoProcessor without torchvision and testing_utils behavior are unrelated." - }, - { - "left": "issue:43931", - "right": "issue:44410", - "accept": false, - "reason": "Qwen3-VL weight-shape mismatch and qwen3next missing projections are different model-specific loading errors." - }, - { - "left": "issue:43874", - "right": "issue:44263", - "accept": false, - "reason": "Missing image-patch API and torch.split value handling are unrelated implementation bugs." - }, - { - "left": "issue:43335", - "right": "issue:45071", - "accept": false, - "reason": "SwitchTransformers sparse-layer construction and config type-checking regressions do not share the same bug." + "reason": "Removed tie_embeddings_and_encoder_decoder support is unrelated to gguf_file breakage in from_pretrained." } ] }, - "evaluator_result": null, - "evaluator_used": false, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the cluster as mostly heterogeneous, selects 45278 as the broad import-regression anchor, and its rejected soft-pairs are conservative. The one accepted tokenizer pair is plausibly mergeable as a single v5 tokenizer regression fix." + }, + "evaluator_used": true, "retried": false, - "accepted_nontrivial_soft_edge": false, + "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null }, @@ -11386,15 +11580,16 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:38617", + "issue:42757", "issue:42886", "issue:43012", "issue:43064", "issue:43066", "issue:43097", - "issue:43208", + "issue:43335", "issue:43352", "issue:43381", "issue:43408", @@ -11402,14 +11597,12 @@ "issue:43531", "issue:43550", "issue:43582", - "issue:43824", "issue:43825", "issue:43873", + "issue:43874", "issue:43901", "issue:43906", "issue:43931", - "issue:43937", - "issue:43957", "issue:43976", "issue:43986", "issue:43992", @@ -11417,6 +11610,7 @@ "issue:44062", "issue:44117", "issue:44261", + "issue:44263", "issue:44315", "issue:44351", "issue:44368", @@ -11427,270 +11621,284 @@ "issue:44485", "issue:44655", "issue:44704", - "issue:44792", "issue:44829", + "issue:44987", "issue:45071", "issue:45103", "issue:45245", + "issue:45290", "issue:45341", "issue:45372", - "issue:45375", - "issue:45479" + "issue:45375" ], "soft_pairs": [ + "issue:44117|issue:44987", + "issue:43986|issue:45341", + "issue:43931|issue:44410", + "issue:43335|issue:45071", "issue:43408|issue:44368", "issue:43931|issue:44038", "issue:43986|issue:45372", + "issue:43874|issue:44263", "issue:44829|issue:45245", - "issue:43064|issue:44315", "issue:43976|issue:45071", "issue:44464|issue:45071", "issue:43531|issue:45341", - "issue:44704|issue:45103", + "issue:43064|issue:44315", + "issue:38617|issue:42757", "issue:44261|issue:44485", - "issue:38617|issue:44351", - "issue:44062|issue:44117", - "issue:43986|issue:44117", "issue:44410|issue:45375", + "issue:44704|issue:45103", "issue:43502|issue:43992", + "issue:44062|issue:44117", + "issue:43986|issue:44117", + "issue:38617|issue:44351", "issue:43582|issue:45341", - "issue:38617|issue:43097", "issue:44038|issue:44410", - "issue:43957|issue:44351", + "issue:43097|issue:43873", "issue:43992|issue:44484", - "issue:43824|issue:43986", + "issue:43986|issue:45290", "issue:43066|issue:43906", - "issue:43097|issue:43873", - "issue:43352|issue:44380", "issue:43825|issue:44655", + "issue:38617|issue:43097", + "issue:43352|issue:44380", "issue:43097|issue:43901", "issue:43012|issue:43550", "issue:43012|issue:44368", - "issue:42886|issue:43381", - "issue:43208|issue:43550", - "issue:43937|issue:44792", - "issue:44380|issue:45479" + "issue:42886|issue:43381" ], - "prepared_review_unit_hash": "e40da513adabe104c5c47a1ad5aeec26070b823370a8387d359ca21b75864398", + "prepared_review_unit_hash": "26b5ed52c893d95a760af23e2b1b323c442b944cecd0431c6fe67f2990de8383", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 32, - "serialized_chars": 37279, - "estimated_input_tokens": 9320, - "estimated_eval_tokens": 18896 + "soft_pair_count": 34, + "serialized_chars": 37881, + "estimated_input_tokens": 9471, + "estimated_eval_tokens": 19198 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly a heterogeneous set of Transformers issues with a few loose v5-regression similarities, but no soft pair looks like the same underlying bug/change. No PRs are present.", - "confidence": 0.86, - "canonical_issue_reason": "issue:43097 is the clearest representative of the cluster\u2019s broad theme: a concrete v5 breaking-change regression with a well-scoped API removal and inbound references.", + "summary": "This cluster is heterogeneous: most pairs are only superficially similar (imports, warnings, tokenizer/Qwen/GLM mentions) but do not appear to describe the same underlying bug or change.", + "confidence": 0.9, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43097 is the strongest standalone issue for triage because it is specific, actionable, and centrally captures the recurring v5 compatibility-break pattern seen across the set.", + "best_issue_reason": "issue:43873 is the most active unresolved bug here, with the broadest practical impact and the clearest ongoing discussion, even though it is not a true duplicate hub for the rest of the set.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading-related, but one is about TOKENIZER_MAPPING_NAMES returning None and the other about a transformers>=5.1.0 model load failure; different failure modes and code paths." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "AutoProcessor/video-model dependency crash vs a testing_utils bug; no shared underlying bug." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Both involve Qwen-family models, but one is a Qwen3-VL weight-shape mismatch and the other is missing layer projections in qwen3next; distinct bugs." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug vs PretrainedConfig type checking regression; unrelated." + }, { "left": "issue:43408", "right": "issue:44368", "accept": false, - "reason": "Different problems: a model-type mismatch warning for sam3_video/sam3_tracker vs a tie_word_embeddings warning during Qwen3.5 LoRA fine-tuning." + "reason": "Both are warnings, but they concern different model families and different config mismatches." }, { "left": "issue:43931", "right": "issue:44038", "accept": false, - "reason": "Different model-loading failures: Qwen3-VL-30B weight-shape mismatch vs a generic transformers 5.0/Qwen3-VL-Moe bug." + "reason": "Same broad Qwen3-VL area, but one is a shape mismatch and the other is a MoE load bug in transformers 5.0; not the same issue." }, { "left": "issue:43986", "right": "issue:45372", "accept": false, - "reason": "Both are loading/import issues, but one is AutoProcessor without torchvision and the other is a mistral_common ReasoningEffort import break in Gemma 4 processor loading." + "reason": "Both are processor-loading failures, but the missing dependency/import is different and the affected processors are unrelated." }, { - "left": "issue:44829", - "right": "issue:45245", + "left": "issue:43874", + "right": "issue:44263", "accept": false, - "reason": "Unrelated failures: flash_attention_3 training degeneration vs a categorical-cardinality runtime limit." + "reason": "Both are GLM-related, but one is a missing processor method and the other is a torch.split/indexer bug; different code paths." }, { - "left": "issue:43064", - "right": "issue:44315", + "left": "issue:44829", + "right": "issue:45245", "accept": false, - "reason": "Different training-state bugs: FSDP2/PEFT cpu_ram_efficient_loading optimizer-state corruption vs Liger Kernel not being applied with model_init." + "reason": "FlashAttention training degeneration and category-count limit are unrelated bugs." }, { "left": "issue:43976", "right": "issue:45071", "accept": false, - "reason": "Both mention version compatibility, but one is a Python support mismatch in 5.1.0 and the other is PreTrainedConfig type-checking breakage in 5.4.0." + "reason": "Python version compatibility issue vs config type-checking regression; no shared fix." }, { "left": "issue:44464", "right": "issue:45071", "accept": false, - "reason": "Chunked generation inconsistency with compiled forward is a runtime-generation bug, not the same as PreTrainedConfig type-checking." + "reason": "Compiled chunked generation inconsistency vs PretrainedConfig typing; unrelated." }, { "left": "issue:43531", "right": "issue:45341", "accept": false, - "reason": "Qwen3-MoE sliding_window behavior is model-specific; the other is a general testing_utils bug." + "reason": "Qwen3-MoE sliding_window behavior vs a testing utility typo/bug; not the same bug." }, { - "left": "issue:44704", - "right": "issue:45103", + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption vs Liger kernel not applied with model_init; different training paths." + }, + { + "left": "issue:38617", + "right": "issue:42757", "accept": false, - "reason": "Both are loader/helper bugs, but one is AutoProcessor kwargs not reaching cached_file and the other is an auto_docstring crash with future annotations." + "reason": "Both are ImportErrors, but they are for different missing symbols in different modules." }, { "left": "issue:44261", "right": "issue:44485", "accept": false, - "reason": "Different RoPE/norm precision issues: missing rms_norm_eps in MLA q_a_layernorm vs GLM-5 RoPE implementation discussion." + "reason": "MLA rms_norm_eps precision issue vs GLM-5 RoPE implementation; unrelated model internals." }, { - "left": "issue:38617", - "right": "issue:44351", + "left": "issue:44410", + "right": "issue:45375", "accept": false, - "reason": "Both are import errors, but for different symbols: layer_type_validation vs HybridCache." + "reason": "Both are Qwen config/model bugs, but one is missing attention projections and the other is a missing vision config field." }, { - "left": "issue:44062", - "right": "issue:44117", + "left": "issue:44704", + "right": "issue:45103", "accept": false, - "reason": "Different tokenizer-loading failures: AddedToken special-argument collision vs TOKENIZER_MAPPING_NAMES sometimes returning None." + "reason": "AutoProcessor kwargs forwarding bug vs auto_docstring crash under future annotations; different layers and fixes." }, { - "left": "issue:43986", - "right": "issue:44117", + "left": "issue:43502", + "right": "issue:43992", "accept": false, - "reason": "AutoProcessor missing torchvision is unrelated to TOKENIZER_MAPPING_NAMES returning None in from_pretrained." + "reason": "Local-files-only network requests vs missing embed_tokens weight on load; separate loading bugs." }, { - "left": "issue:44410", - "right": "issue:45375", + "left": "issue:44062", + "right": "issue:44117", "accept": false, - "reason": "Different Qwen variants and different bugs: missing attn projections vs missing deepstack_visual_indexes config field." + "reason": "Tokenizer AddedToken argument duplication vs TOKENIZER_MAPPING_NAMES assuming non-None; both tokenizer-related but not the same defect." }, { - "left": "issue:43502", - "right": "issue:43992", + "left": "issue:43986", + "right": "issue:44117", "accept": false, - "reason": "Both are loading-related, but one is unwanted API requests with local_files_only and the other is UMT5Encoder missing embed_tokens.weight." + "reason": "Processor-loading crash without torchvision vs tokenizer mapping None-handling; different subsystems and symptoms." }, { - "left": "issue:43582", - "right": "issue:45341", + "left": "issue:38617", + "right": "issue:44351", "accept": false, - "reason": "AppleSilicon caching_allocator_warmup TypeError is unrelated to the testing_utils bug." + "reason": "Both are import-name errors, but they concern different missing public symbols and likely separate API regressions." }, { - "left": "issue:38617", - "right": "issue:43097", + "left": "issue:43582", + "right": "issue:45341", "accept": false, - "reason": "Different regressions: missing internal import vs removal of tie_embeddings_and_encoder_decoder." + "reason": "Apple Silicon caching allocator TypeError vs testing_utils bug; unrelated." }, { "left": "issue:44038", "right": "issue:44410", "accept": false, - "reason": "Both are Qwen-family issues, but one is a general transformers 5.0/Qwen3-VL-Moe bug and the other is a specific missing projection problem in qwen3next." + "reason": "Both reference Qwen3, but one is a model load failure and the other is missing layer projections; not the same underlying bug." }, { - "left": "issue:43957", - "right": "issue:44351", + "left": "issue:43097", + "right": "issue:43873", "accept": false, - "reason": "Meta-device model loading breakage is not the same as the HybridCache import failure." + "reason": "Config deprecation/removal warning issue vs quantization offloading bug; unrelated." }, { "left": "issue:43992", "right": "issue:44484", "accept": false, - "reason": "Missing embed_tokens.weight during load is distinct from the save_pretrained max_shard_size default question." + "reason": "Missing encoder weight on load vs default max_shard_size question; different concerns." }, { - "left": "issue:43824", - "right": "issue:43986", + "left": "issue:43986", + "right": "issue:45290", "accept": false, - "reason": "Different import/load failures: missing Qwen2_5_VLForConditionalGeneration export vs AutoProcessor crashing without torchvision." + "reason": "AutoProcessor video-model dependency crash vs chat-template tokenization crash with tool calls; different code paths." }, { "left": "issue:43066", "right": "issue:43906", "accept": false, - "reason": "Wrong tokenizer decoder type in v5 is unrelated to the isolated reproduction of issue 38071." + "reason": "A title referencing an isolated reproduction is not enough to match a specific tokenizer-decoder bug; the concrete failure descriptions differ." }, { - "left": "issue:43097", - "right": "issue:43873", + "left": "issue:43825", + "right": "issue:44655", "accept": false, - "reason": "API removal in v5 is not the same as offloading failing under quantization; they may both be version regressions, but not the same bug." + "reason": "Pipeline task messaging vs saving Pipeline objects; separate pipeline issues." }, { - "left": "issue:43352", - "right": "issue:44380", + "left": "issue:38617", + "right": "issue:43097", "accept": false, - "reason": "Different model/backends: NemotronHForCausalLM lacking Flash Attention 2.0 support vs GPT2 attention scaling ignored under SDPA/FlashAttention." + "reason": "ImportError for a missing config symbol vs removal of a config option; not the same bug." }, { - "left": "issue:43825", - "right": "issue:44655", + "left": "issue:43352", + "right": "issue:44380", "accept": false, - "reason": "Different pipeline/config messaging issue vs inability to save Pipeline objects with save_pretrained." + "reason": "FlashAttention 2 support error for Nemotron vs GPT2 attention scaling ignored under SDPA/FlashAttention; different models and fixes." }, { "left": "issue:43097", "right": "issue:43901", "accept": false, - "reason": "API removal in v5 is unrelated to docs still mentioning return_all_scores." + "reason": "Removed tie_embeddings setting vs text-classification docs mismatch; unrelated." }, { "left": "issue:43012", "right": "issue:43550", "accept": false, - "reason": "Bfloat16 compilation warning is unrelated to Bamba-9B-v2 torch.compile + SDPA failure." + "reason": "bfloat16 compile warning vs Bamba torch.compile SDPA failure; different symptoms and likely different fixes." }, { "left": "issue:43012", "right": "issue:44368", "accept": false, - "reason": "PyTorch bfloat16 precision warning is not the same as the tie_word_embeddings warning in LoRA fine-tuning." + "reason": "PyTorch precision warning during compile vs tie_word_embeddings warning; both warnings but not the same behavior." }, { "left": "issue:42886", "right": "issue:43381", "accept": false, - "reason": "Offline cache-tokenizer loading failure is different from gradient checkpointing being unusable in eval mode." - }, - { - "left": "issue:43208", - "right": "issue:43550", - "accept": false, - "reason": "xLSTM training blockers are model-specific and unrelated to the Bamba SDPA compile failure." - }, - { - "left": "issue:43937", - "right": "issue:44792", - "accept": false, - "reason": "Invalid GenerationConfig for GLM-5 is not the same as a janus image-generation test failure." - }, - { - "left": "issue:44380", - "right": "issue:45479", - "accept": false, - "reason": "GPT2 attention scaling under SDPA/FlashAttention is unrelated to the zero-loss sequence-classification issue." + "reason": "Offline cache-loading failure vs gradient checkpointing in eval mode; unrelated runtime paths." } ] }, @@ -11712,14 +11920,14 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ + "issue:36010", "issue:36683", "issue:38617", - "issue:41669", + "issue:42757", "issue:42907", "issue:42994", - "issue:43010", "issue:43012", "issue:43064", "issue:43066", @@ -11730,23 +11938,22 @@ "issue:43502", "issue:43525", "issue:43526", + "issue:43550", "issue:43582", "issue:43618", "issue:43756", "issue:43761", + "issue:43824", "issue:43867", "issue:43901", "issue:43937", "issue:43976", - "issue:43992", "issue:44038", "issue:44077", "issue:44117", "issue:44188", "issue:44220", "issue:44230", - "issue:44242", - "issue:44246", "issue:44263", "issue:44380", "issue:44483", @@ -11755,290 +11962,290 @@ "issue:44514", "issue:44655", "issue:44683", + "issue:44792", "issue:44938", "issue:45216", - "issue:45278", "issue:45290", - "issue:45310", "issue:45381", - "issue:45440" + "issue:45440", + "issue:45479" ], "soft_pairs": [ + "issue:43208|issue:43550", + "issue:43937|issue:44792", "issue:42907|issue:44655", + "issue:44380|issue:45479", "issue:43525|issue:44117", "issue:43352|issue:43976", "issue:43064|issue:44380", "issue:43208|issue:43526", - "issue:43618|issue:44380", "issue:44483|issue:44514", + "issue:43618|issue:44380", "issue:36683|issue:43582", "issue:43618|issue:43761", "issue:43012|issue:44683", - "issue:38617|issue:43502", "issue:44483|issue:45290", + "issue:42757|issue:43824", + "issue:38617|issue:43502", "issue:44038|issue:44230", "issue:44483|issue:45381", "issue:43299|issue:43867", - "issue:43756|issue:44485", "issue:43901|issue:44655", - "issue:43867|issue:45216", "issue:42994|issue:44655", + "issue:43756|issue:44485", "issue:36683|issue:43475", - "issue:43352|issue:43937", + "issue:43867|issue:45216", "issue:44220|issue:44655", + "issue:43352|issue:43937", "issue:43066|issue:45381", - "issue:43064|issue:44492", "issue:44077|issue:44380", "issue:36683|issue:44077", "issue:44263|issue:45440", + "issue:43064|issue:44492", "issue:43976|issue:44938", "issue:36683|issue:43525", + "issue:36010|issue:42757", "issue:44188|issue:44380", - "issue:43992|issue:45310", - "issue:43618|issue:45216", - "issue:36683|issue:44242", - "issue:43010|issue:45278", - "issue:41669|issue:44246", - "issue:36683|issue:43352" + "issue:43618|issue:45216" ], - "prepared_review_unit_hash": "53c26a1423fd700cd60273543702b485dc431172a74e7d9f09ee4799c356b2bb", + "prepared_review_unit_hash": "07e4f140947f9edc3fd706ededa0158057371a599de9b9bc7ba3e83f0e5a3663", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, + "node_count": 47, + "item_count": 47, "soft_pair_count": 36, - "serialized_chars": 37989, - "estimated_input_tokens": 9498, - "estimated_eval_tokens": 19252 + "serialized_chars": 37554, + "estimated_input_tokens": 9389, + "estimated_eval_tokens": 19034 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous: config attribute errors, saving/loading regressions, chat-template/API crashes, import-time issues, and model-specific behavior bugs. A few pairs are thematically close, but none are strong enough to treat as the same underlying bug/change.", - "confidence": 0.87, - "canonical_issue_reason": null, + "summary": "The cluster is highly heterogeneous: most items are distinct model-specific regressions, import errors, save/load bugs, or unrelated runtime issues. None of the suggested soft edges look like the same underlying bug or change, so I would reject all of them.", + "confidence": 0.97, + "canonical_issue_reason": "Issue 44188 is the strongest single anchor because it is the most active open issue and a core runtime regression, but the set is too diverse for any issue to truly represent the whole cluster.", "canonical_pr_reason": null, - "best_issue_reason": "issue:45278 is the broadest umbrella-style report in the set, covering many import errors after the 4.57.0 -> 5.5.0 update, so it is the most representative issue here.", + "best_issue_reason": "Issue 44188 is the best available cluster representative: it is open, high-activity, and centered on core execution behavior. Even so, it does not subsume the other items well.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "Different bugs and model families: xLSTM training issues vs Bamba torch.compile/SDPA failure." + }, + { + "left": "issue:43937", + "right": "issue:44792", + "accept": false, + "reason": "GLM-5 invalid GenerationConfig vs janus test failure; unrelated failure modes." + }, { "left": "issue:42907", "right": "issue:44655", "accept": false, - "reason": "Both are save-related, but one is about dequantized model saving and the other about saving Pipeline objects; different code paths and failures." + "reason": "Both involve saving, but one is dequantized Ministrals/Devstrals and the other is Pipeline.save_pretrained; different code paths." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "GPT2 attention scaling under SDPA/FlashAttention vs degenerate zero loss in sequence classification; unrelated." }, { "left": "issue:43525", "right": "issue:44117", "accept": false, - "reason": "One is a missing config attribute, the other is tokenizer mapping returning None during from_pretrained; not the same bug." + "reason": "Missing pad_token_id on Llama4Config is unrelated to TOKENIZER_MAPPING_NAMES returning None." }, { "left": "issue:43352", "right": "issue:43976", "accept": false, - "reason": "Nemotron Flash Attention support and a Python version compatibility break are unrelated problems." + "reason": "Flash Attention 2 support issue vs Python version compatibility issue; not the same bug." }, { "left": "issue:43064", "right": "issue:44380", "accept": false, - "reason": "Distributed optimizer-state corruption in Trainer/FSDP2/PEFT is not the same as GPT-2 attention scaling being ignored in SDPA/FlashAttention." + "reason": "FSDP2/PEFT optimizer-state mismatch is unrelated to GPT2 attention scaling behavior." }, { "left": "issue:43208", "right": "issue:43526", "accept": false, - "reason": "xLSTM training bugs and BEiT image processor label reduction are separate model/component issues." + "reason": "xLSTM training bugs vs BeitImageProcessorFast label reduction bug; different subsystems." }, { - "left": "issue:43618", - "right": "issue:44380", + "left": "issue:44483", + "right": "issue:44514", "accept": false, - "reason": "CLIP output assignment and GPT-2 attention scaling are different regressions in different model families." + "reason": "Both are chat-related, but one is /v1/chat/completions request handling and the other is apply_chat_template batching/padding." }, { - "left": "issue:44483", - "right": "issue:44514", + "left": "issue:43618", + "right": "issue:44380", "accept": false, - "reason": "One is an API request rejection, the other is a chat-template crash on batched input; related area but not the same defect." + "reason": "CLIPOutput attentions regression is unrelated to GPT2 scaling under SDPA/FlashAttention." }, { "left": "issue:36683", "right": "issue:43582", "accept": false, - "reason": "Gemma3Config missing vocab_size and Apple Silicon allocator TypeError are unrelated." + "reason": "Gemma3Config missing vocab_size vs Apple Silicon TypeError in caching_allocator_warmup; unrelated." }, { "left": "issue:43618", "right": "issue:43761", "accept": false, - "reason": "Both involve CLIP, but one is missing attentions and the other is missing hidden_states; too specific to merge as one bug." + "reason": "Both involve CLIP, but one is attentions not assigned and the other is hidden_states missing; different concrete failures." }, { "left": "issue:43012", "right": "issue:44683", "accept": false, - "reason": "A precision warning during compile is not the same as compiled flex_attention failing on torch >= 2.9." + "reason": "PyTorch warning on bfloat16 compilation is not the same as flex_attention failing on torch>=2.9." }, { - "left": "issue:38617", - "right": "issue:43502", + "left": "issue:44483", + "right": "issue:45290", "accept": false, - "reason": "ImportError for layer_type_validation and network calls despite local_files_only=True are unrelated." + "reason": "Different chat surfaces: API request rejection vs apply_chat_template crash on assistant tool-call messages." }, { - "left": "issue:44483", - "right": "issue:45290", + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Import errors, but for different symbols and packages: huggingface_hub.is_offline_mode vs transformers.Qwen2_5_VLForConditionalGeneration." + }, + { + "left": "issue:38617", + "right": "issue:43502", "accept": false, - "reason": "Both touch chat/completions, but one is an endpoint acceptance failure and the other is an apply_chat_template crash; not clearly the same underlying bug." + "reason": "Import failure from configuration_utils is unrelated to local_files_only still triggering network requests." }, { "left": "issue:44038", "right": "issue:44230", "accept": false, - "reason": "Qwen3-VL-Moe regression and fp8 support for Qwen3-VL/Qwen3.5 MoE are different issues." + "reason": "Same model family, but one is a loading regression and the other is an fp8 support request; not the same fix." }, { "left": "issue:44483", "right": "issue:45381", "accept": false, - "reason": "Chat-completions request handling and qwen2.5-vl video vision_position_ids are separate bugs." + "reason": "Chat-completions request rejection is unrelated to Qwen2.5-VL video vision_position_ids being wrong." }, { "left": "issue:43299", "right": "issue:43867", "accept": false, - "reason": "Qwen3.5 MoE loading regression and a generic sorted-state_dict load error are not the same concrete failure." - }, - { - "left": "issue:43756", - "right": "issue:44485", - "accept": false, - "reason": "Both mention RoPE, but one is a Smollm3 layer-count issue and the other is a GLM-5 RoPE implementation discussion." + "reason": "Qwen3VL-Moe loading breakage is not the same as a sorted-state_dict load error." }, { "left": "issue:43901", "right": "issue:44655", "accept": false, - "reason": "A docs/behavior mismatch for TextClassificationPipeline is unrelated to saving Pipeline objects." + "reason": "Docs mismatch for return_all_scores is unrelated to inability to save Pipeline objects." }, { - "left": "issue:43867", - "right": "issue:45216", + "left": "issue:42994", + "right": "issue:44655", "accept": false, - "reason": "Generic load-model failure with sorted state_dict is not the same as the Qwen3.5 save_pretrained regression." + "reason": "Both involve saving, but quantized model save failure and Pipeline.save_pretrained failure are different artifacts and bugs." }, { - "left": "issue:42994", - "right": "issue:44655", + "left": "issue:43756", + "right": "issue:44485", "accept": false, - "reason": "Quantized model saving failure and pipeline save_pretrained failure are different save paths." + "reason": "Smollm3 RoPE-layer issue and GLM-5 RoPE implementation discussion are different model-specific problems." }, { "left": "issue:36683", "right": "issue:43475", "accept": false, - "reason": "Missing Gemma3Config vocab_size and missing Sam3VisionEncoderOutput.fpn_position_embeddings are separate attribute errors on different models." + "reason": "Different missing-attribute regressions on different models: Gemma3Config.vocab_size vs Sam3VisionEncoderOutput.fpn_position_embeddings." }, { - "left": "issue:43352", - "right": "issue:43937", + "left": "issue:43867", + "right": "issue:45216", "accept": false, - "reason": "Flash Attention support for Nemotron is unrelated to GLM-5 GenerationConfig validation failure." + "reason": "Both concern saving/loading checkpoints, but the sorted state_dict load error and the Qwen3.5 save_pretrained regression are distinct." }, { "left": "issue:44220", "right": "issue:44655", "accept": false, - "reason": "_torch_extract_fbank_features() and saving Pipeline objects are unrelated components." + "reason": "Audio feature extraction bug vs Pipeline saving bug; no shared code path." }, { - "left": "issue:43066", - "right": "issue:45381", + "left": "issue:43352", + "right": "issue:43937", "accept": false, - "reason": "Tokenizer decoder type regression and qwen2.5-vl video position IDs are different model-loading/processing bugs." + "reason": "Flash Attention support for NemotronH is unrelated to GLM-5 GenerationConfig validation." }, { - "left": "issue:43064", - "right": "issue:44492", + "left": "issue:43066", + "right": "issue:45381", "accept": false, - "reason": "Wrong optimizer state handling and a typo in cache strategies are unrelated." + "reason": "Wrong tokenizer decoder type in v5 is unrelated to Qwen2.5-VL video position IDs." }, { "left": "issue:44077", "right": "issue:44380", "accept": false, - "reason": "Optional post_init policy and GPT-2 attention scaling are not the same bug." + "reason": "patchtsmixer post_init validation and GPT2 attention scaling are unrelated." }, { "left": "issue:36683", "right": "issue:44077", "accept": false, - "reason": "A missing config field in Gemma3 and a patchtsmixer post_init policy change are unrelated." + "reason": "Gemma3Config missing vocab_size is unrelated to patchtsmixer post_init allowance." }, { "left": "issue:44263", "right": "issue:45440", "accept": false, - "reason": "A torch.split return-value issue in GlmMoeDsaIndexer is not the same as DeepseekV3MoE diverging from remote implementation." + "reason": "Both are MoE-related, but they affect different models and different internals." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Optimizer-state corruption and a cache-strategy typo are unrelated." }, { "left": "issue:43976", "right": "issue:44938", "accept": false, - "reason": "Python 3.9/3.10 compatibility and Python 3.14 loading failure are different environment regressions." + "reason": "Both are Python compatibility issues, but they target different interpreter versions and likely different causes." }, { - "left": "issue:36683", - "right": "issue:43525", + "left": "issue:36010", + "right": "issue:42757", "accept": false, - "reason": "Gemma3Config missing vocab_size and Llama4Config missing pad_token_id are separate missing-attribute bugs." + "reason": "Different import failures from different modules and symbols." }, { "left": "issue:44188", "right": "issue:44380", "accept": false, - "reason": "torch.compile branching in attention kernels and GPT-2 attention scaling under SDPA/FlashAttention are different issues." - }, - { - "left": "issue:43992", - "right": "issue:45310", - "accept": false, - "reason": "UMT5 missing embed_tokens.weight and Qwen3.5 MoE from_pretrained error are unrelated loading failures." + "reason": "Different attention/backend regressions: allow_is_bidirectional_skip under torch.compile vs GPT2 scaling under SDPA/FlashAttention." }, { "left": "issue:43618", "right": "issue:45216", "accept": false, - "reason": "CLIP attentions no longer assigned and Qwen3.5 saved checkpoint regression are unrelated model-specific bugs." - }, - { - "left": "issue:43010", - "right": "issue:45278", - "accept": false, - "reason": "Cache update no_grad decoration and broad import errors after upgrade are different problem areas." - }, - { - "left": "issue:41669", - "right": "issue:44246", - "accept": false, - "reason": "Both involve import performance, but one is a specific import-* cleanup and the other is a generic intermittent slow import report; not enough to merge." - }, - { - "left": "issue:36683", - "right": "issue:43352", - "accept": false, - "reason": "Gemma3Config missing vocab_size and Nemotron Flash Attention support are unrelated." + "reason": "CLIP attentions regression is unrelated to the Qwen3.5 save_pretrained checkpoint bug." } ] }, @@ -12060,14 +12267,13 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:28282", - "issue:30990", - "issue:31795", "issue:33453", "issue:36296", "issue:36683", + "issue:41669", "issue:42548", "issue:42890", "issue:43010", @@ -12089,21 +12295,21 @@ "issue:44077", "issue:44112", "issue:44222", - "issue:44315", + "issue:44242", + "issue:44246", "issue:44351", - "issue:44355", "issue:44360", "issue:44387", "issue:44484", - "issue:44485", "issue:44556", - "issue:44749", "issue:44857", "issue:44938", "issue:44964", "issue:44977", "issue:45071", "issue:45125", + "issue:45278", + "issue:45310", "issue:45341", "issue:45357", "issue:45375", @@ -12112,10 +12318,15 @@ "issue:45478" ], "soft_pairs": [ + "issue:43992|issue:45310", + "issue:36683|issue:44242", + "issue:43010|issue:45278", + "issue:36683|issue:43352", + "issue:41669|issue:44246", "issue:43352|issue:43931", - "issue:43992|issue:45478", "issue:43867|issue:45375", "issue:36683|issue:43064", + "issue:43992|issue:45478", "issue:44938|issue:45071", "issue:43867|issue:45357", "issue:43010|issue:43976", @@ -12125,260 +12336,242 @@ "issue:43644|issue:44938", "issue:36296|issue:43575", "issue:43867|issue:44484", - "issue:42890|issue:44112", - "issue:42890|issue:44964", "issue:43618|issue:44857", + "issue:42890|issue:44964", + "issue:42890|issue:44112", "issue:28282|issue:33453", "issue:43618|issue:44977", "issue:43502|issue:44062", "issue:43010|issue:44075", "issue:42890|issue:45341", "issue:43723|issue:44556", - "issue:42548|issue:44351", "issue:44387|issue:44556", - "issue:43010|issue:43065", "issue:43010|issue:43232", "issue:44222|issue:45125", - "issue:44355|issue:44749", - "issue:45125|issue:45478", - "issue:45125|issue:45375", - "issue:30990|issue:31795", - "issue:43502|issue:44485", - "issue:44315|issue:44360", - "issue:36296|issue:44075" + "issue:43010|issue:43065", + "issue:42548|issue:44351", + "issue:45125|issue:45478" ], - "prepared_review_unit_hash": "fad5508a8b64926321f70926e4f50f49a666f3c6e1228b068a27b8bd0f83a3cd", + "prepared_review_unit_hash": "f2149912b8b6170e28f17d509ed7cd0e0d5a0188463ad8cbd902664207207d0c", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 48, - "item_count": 48, - "soft_pair_count": 34, - "serialized_chars": 37303, - "estimated_input_tokens": 9326, - "estimated_eval_tokens": 18908 + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36385, + "estimated_input_tokens": 9097, + "estimated_eval_tokens": 18450 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This set is mostly a heterogeneous mix of unrelated Transformers issues, with only a loose cluster around loading/regression reports. None of the soft pairs look like the same concrete bug or change, so no duplicate merges are warranted.", - "confidence": 0.44, - "canonical_issue_reason": "issue:43723 is the clearest representative of the loading-regression subset: it is generic, user-facing, and closest to the few other load-path reports in this set.", + "summary": "Most items are unrelated standalone issues; the only obvious exact duplicate in the set is the Qwen3.5 MoE from_pretrained regression (45310/45478), but it is not one of the listed soft edges. No soft pair meets a strict same-bug/same-change duplicate threshold.", + "confidence": 0.58, + "canonical_issue_reason": "issue:45310 is the clearest representative issue in this set: it is a concrete, narrowly scoped regression report for Qwen3.5 MoE from_pretrained, and issue:45478 is an exact-title duplicate of it.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43723 is the best overall anchor because it describes a broad v5 loading failure without being overly model-specific.", + "best_issue_reason": "issue:45310 is the best global representative because it describes a specific reproducible bug, is well-scoped, and has a near-identical duplicate, making it a strong canonical triage target.", "best_pr_reason": null, "soft_edge_verdicts": [ { - "left": "issue:43352", - "right": "issue:43931", + "left": "issue:43992", + "right": "issue:45310", "accept": false, - "reason": "Different models and different failures: flash-attention support gating vs weight-shape mismatch during load." + "reason": "Different code paths: UMT5Encoder missing embed_tokens.weight vs Qwen3.5 MoE from_pretrained error." }, { - "left": "issue:43992", - "right": "issue:45478", + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Gemma3Config vocab_size attribute error is unrelated to load-balancing loss not being added." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Decorator/no_grad cache update bug is unrelated to broad import errors after upgrading transformers." + }, + { + "left": "issue:36683", + "right": "issue:43352", + "accept": false, + "reason": "Config attribute bug and Flash Attention 2 unsupported error are different problems." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Both mention import slowdowns, but one is specifically about import-star usage in models while the other is a generic intermittent import-time complaint; not clearly the same underlying bug." + }, + { + "left": "issue:43352", + "right": "issue:43931", "accept": false, - "reason": "Both are from_pretrained/load failures, but one is a missing embedding weight and the other is a Qwen3.5 MoE load error; not the same bug." + "reason": "Flash Attention support check vs Qwen3-VL weight shape mismatch are unrelated." }, { "left": "issue:43867", "right": "issue:45375", "accept": false, - "reason": "A sorted-state_dict load error and a missing config field are different root causes." + "reason": "State_dict sorting/load error is not the same as a missing config field dropped by strict parsing." }, { "left": "issue:36683", "right": "issue:43064", "accept": false, - "reason": "Unrelated: Gemma3 config attribute error vs distributed training optimizer-state corruption." + "reason": "Gemma3Config vocab_size bug and FSDP2 optimizer-state mismatch are unrelated." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "UMT5Encoder missing embedding weights is unrelated to the Qwen3.5 MoE from_pretrained regression." }, { "left": "issue:44938", "right": "issue:45071", "accept": false, - "reason": "Both are version-related, but Python 3.14 load failure and PretrainedConfig type-checking breakage are separate issues." + "reason": "Python 3.14 import/load failure is not the same as PretrainedConfig type-checking breakage." }, { "left": "issue:43867", "right": "issue:45357", "accept": false, - "reason": "One is a load-order/state_dict problem; the other is incorrect save_pretrained key serialization." + "reason": "Model load error from state_dict ordering is unrelated to incorrect save_pretrained visual encoder keys." }, { "left": "issue:43010", "right": "issue:43976", "accept": false, - "reason": "Cache update no_grad behavior is unrelated to Python-version compatibility in 5.1.0." + "reason": "Generation kwargs/cache update bug is unrelated to a Python-version compatibility issue." }, { "left": "issue:44938", "right": "issue:45381", "accept": false, - "reason": "Python 3.14 import/load failure and Qwen2.5-VL vision_position_ids are unrelated." + "reason": "Interpreter load failure and Qwen2.5-VL video position-id bug are different issues." }, { "left": "issue:44360", "right": "issue:45440", "accept": false, - "reason": "DSA indexer/ReLU discussion is unrelated to DeepseekV3MoE divergence." + "reason": "DSA indexer missing ReLU and DeepseekV3MoE divergence are unrelated code paths." }, { "left": "issue:43618", "right": "issue:44077", "accept": false, - "reason": "CLIP attentions regression and SGD args not being used are different code paths." + "reason": "CLIPOutput attentions regression is unrelated to patchtsmixer post_init eligibility." }, { "left": "issue:43644", "right": "issue:44938", "accept": false, - "reason": "Non-persistent buffer junk and Python 3.14 loading failure are different bugs." + "reason": "Non-persistent buffer junk fill is unrelated to Python 3.14 import failure." }, { "left": "issue:36296", "right": "issue:43575", "accept": false, - "reason": "Tensor-parallel training bug and tp-related OOM while loading a Qwen model are not the same concrete failure." + "reason": "Tensor parallel training bug and tp-related OOM on Qwen2-57B are not the same concrete defect." }, { "left": "issue:43867", "right": "issue:44484", "accept": false, - "reason": "Load error from sorted state_dict is unrelated to the 50GB max_shard_size default question." + "reason": "State_dict sorting load error and save_pretrained shard-size question are unrelated." }, { - "left": "issue:42890", - "right": "issue:44112", + "left": "issue:43618", + "right": "issue:44857", "accept": false, - "reason": "Both are flaky tests, but they target different models and different causes." + "reason": "Missing CLIPOutput attentions and LwDetrImageLoss AMP crash are different bugs." }, { "left": "issue:42890", "right": "issue:44964", "accept": false, - "reason": "Test seeding flakiness is unrelated to loading microsoft/Phi-4-multimodal-instruct." - }, - { - "left": "issue:43618", - "right": "issue:44857", - "accept": false, - "reason": "Attention field regression and AMP/CUDA loss crash are unrelated." + "reason": "SamHQ test seed flakiness is unrelated to Phi-4 multimodal model loading failure." }, { "left": "issue:28282", "right": "issue:33453", "accept": false, - "reason": "ImportError for missing PyTorch and tokenizer-loading regression are different problems." + "reason": "PyTorch-missing AutoModel ImportError and tokenizer loading regression are different issues." }, { "left": "issue:43618", "right": "issue:44977", "accept": false, - "reason": "CLIP attentions disappearing and Qwen3.5 flash-attention generation issues do not share a concrete bug." + "reason": "CLIPOutput attentions regression is unrelated to Qwen3.5 flash-attention generation issues." }, { "left": "issue:43502", "right": "issue:44062", "accept": false, - "reason": "local_files_only still making requests is unrelated to AddedToken special-argument duplication." + "reason": "local_files_only network leakage and AddedToken keyword collision are unrelated." }, { "left": "issue:43010", "right": "issue:44075", "accept": false, - "reason": "Cache update decoration and optimizer argument handling are unrelated." + "reason": "Cache update no_grad behavior and SGD args being ignored are unrelated." }, { "left": "issue:42890", "right": "issue:45341", "accept": false, - "reason": "A specific SamHQ test seed issue and a generic testing_utils bug are not the same underlying defect." + "reason": "SamHQ test flakiness is unrelated to the testing_utils bug." }, { "left": "issue:43723", "right": "issue:44556", "accept": false, - "reason": "Both concern loading, but one is tokenizer.from_pretrained and the other is checkpoint reload after upgrade; too different to merge." - }, - { - "left": "issue:42548", - "right": "issue:44351", - "accept": false, - "reason": "Import errors for different symbols (PreTrainedModel vs HybridCache) are separate API/export issues." + "reason": "Tokenizer loading regression and checkpoint reload regression after version upgrade are different load paths." }, { "left": "issue:44387", "right": "issue:44556", "accept": false, - "reason": "int4 quantization memory growth and checkpoint reload failure are different failure modes." - }, - { - "left": "issue:43010", - "right": "issue:43065", - "accept": false, - "reason": "Cache no_grad behavior and Sam3PixelDecoder dummy Conv2d are unrelated." + "reason": "Int4 CUDA reserved-memory OOM and checkpoint reload failure are unrelated." }, { "left": "issue:43010", "right": "issue:43232", "accept": false, - "reason": "Both touch generation internals, but the described bugs are different and not clearly the same code-path issue." + "reason": "Cache update no_grad bug and generation kwargs after sync_gpus are different generation issues." }, { "left": "issue:44222", "right": "issue:45125", "accept": false, - "reason": "FP8 save_pretrained behavior and missing tensor-parallel plan are separate model-feature bugs." + "reason": "FP8 save_pretrained moe and missing _tp_plan for tensor parallelism are different Qwen/MoE problems." }, { - "left": "issue:44355", - "right": "issue:44749", + "left": "issue:43010", + "right": "issue:43065", "accept": false, - "reason": "Compiled Python file loading errors and data filtering slowness after upgrade are unrelated." + "reason": "Cache/layer update decoration bug is unrelated to the Sam3PixelDecoder dummy Conv2d issue." }, { - "left": "issue:45125", - "right": "issue:45478", + "left": "issue:42548", + "right": "issue:44351", "accept": false, - "reason": "Missing _tp_plan and from_pretrained error in Qwen3.5 MoE are related to the same model family, but not clearly the same concrete bug." + "reason": "cannot import PreTrainedModel and cannot import HybridCache are different export/import failures." }, { "left": "issue:45125", - "right": "issue:45375", - "accept": false, - "reason": "Tensor-parallel planning and a missing vision config field are different defects." - }, - { - "left": "issue:30990", - "right": "issue:31795", - "accept": false, - "reason": "Sentence-transformers loading hang and forward-doc confusion are unrelated." - }, - { - "left": "issue:43502", - "right": "issue:44485", - "accept": false, - "reason": "local_files_only network requests and GLM-5 RoPE implementation are unrelated." - }, - { - "left": "issue:44315", - "right": "issue:44360", - "accept": false, - "reason": "Liger kernel application and DSA indexer/ReLU discussion are unrelated." - }, - { - "left": "issue:36296", - "right": "issue:44075", + "right": "issue:45478", "accept": false, - "reason": "Tensor-parallel training bug and SGD args not used are not the same issue." + "reason": "Missing tensor-parallel plan and Qwen3.5 MoE from_pretrained error are different problems." } ] }, @@ -12400,8 +12593,9 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ + "issue:30990", "issue:31795", "issue:33357", "issue:33453", @@ -12410,8 +12604,8 @@ "issue:38175", "issue:38617", "issue:41628", - "issue:41669", "issue:42548", + "issue:42757", "issue:42831", "issue:42890", "issue:43010", @@ -12426,313 +12620,312 @@ "issue:43827", "issue:43874", "issue:43901", - "issue:44038", + "issue:44075", "issue:44222", "issue:44230", - "issue:44261", "issue:44273", + "issue:44315", "issue:44351", "issue:44355", "issue:44360", "issue:44368", "issue:44373", - "issue:44393", "issue:44485", "issue:44492", "issue:44556", - "issue:44623", "issue:44655", + "issue:44749", + "issue:44871", "issue:44908", - "issue:44938", "issue:44961", "issue:45003", "issue:45071", "issue:45125", - "issue:45230", + "issue:45375", "issue:45468" ], "soft_pairs": [ + "issue:44355|issue:44749", + "issue:45125|issue:45375", + "issue:30990|issue:31795", + "issue:43502|issue:44485", "issue:43010|issue:44368", - "issue:38175|issue:43502", + "issue:44075|issue:44871", + "issue:44315|issue:44360", "issue:42890|issue:44373", + "issue:38175|issue:43502", "issue:43299|issue:44556", "issue:43502|issue:44961", "issue:44355|issue:44655", - "issue:44222|issue:44230", "issue:44355|issue:45071", + "issue:44222|issue:44230", "issue:42831|issue:43874", "issue:42890|issue:43232", + "issue:42757|issue:43576", "issue:33453|issue:34689", "issue:33453|issue:36296", "issue:43827|issue:44961", - "issue:43576|issue:45071", "issue:43576|issue:43824", + "issue:43576|issue:45071", "issue:43673|issue:44492", "issue:38617|issue:43576", "issue:33357|issue:33453", - "issue:41628|issue:42548", "issue:44961|issue:45468", "issue:31795|issue:38617", + "issue:41628|issue:42548", "issue:43576|issue:44351", "issue:44908|issue:44961", - "issue:43576|issue:44938", + "issue:36296|issue:43901", "issue:44485|issue:44961", "issue:44273|issue:45003", - "issue:36296|issue:43901", "issue:36296|issue:43526", - "issue:41628|issue:44355", "issue:36296|issue:43012", - "issue:44623|issue:45230", - "issue:44261|issue:44360", - "issue:43576|issue:44556", - "issue:44393|issue:45071", - "issue:41669|issue:43673", - "issue:44038|issue:45125" + "issue:41628|issue:44355" ], - "prepared_review_unit_hash": "963e8654bcbd0eeb34b687502ea664d69793651205251dd68364583a5f342289", + "prepared_review_unit_hash": "6c5616a16fd5287a8357d223f5e30adc7ba0b02decc7abcf7694edf7fe2df22f", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, "soft_pair_count": 36, - "serialized_chars": 37720, - "estimated_input_tokens": 9430, - "estimated_eval_tokens": 19116 + "serialized_chars": 37906, + "estimated_input_tokens": 9477, + "estimated_eval_tokens": 19210 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "Mostly heterogeneous issue reports with no PRs present. The cluster\u2019s closest representative is the top-level `transformers` import regression, but the soft pairs are all distinct bugs/changes and should be rejected.", - "confidence": 0.42, - "canonical_issue_reason": "Issue 42548 is the broadest representative of the cluster\u2019s import/export breakage theme: a top-level `transformers` import failure that best subsumes the other symbol-import regressions.", + "summary": "This cluster is mostly heterogeneous: loading/import regressions, doc issues, model-specific bugs, and unrelated test/perf reports. None of the soft pairs look like the same underlying bug or change, so all soft edges are rejected.", + "confidence": 0.91, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Issue 42548 is the best single issue to anchor this cluster because it is a user-facing, high-impact root import regression rather than a narrow model- or test-specific problem.", + "best_issue_reason": "issue:43576 is the broadest v5-facing regression report and has external references, so it is the most reasonable representative if one issue must be picked.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Different bugs: compiled-file import/runtime errors vs slowdown when filtering after version upgrade." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Different model-parallel/config problems: missing _tp_plan vs a config field being dropped." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Loading hang in Sentence Transformers vs documentation confusion in forward() are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only causing network calls is unrelated to GLM-5 RoPE implementation." + }, { "left": "issue:43010", "right": "issue:44368", "accept": false, - "reason": "Different bugs: cache/layer `update()` no-grad decoration vs a tie-word-embeddings warning in Qwen3.5 fine-tuning." + "reason": "Cache update no_grad decoration vs tie_word_embeddings warning are different code paths." }, { - "left": "issue:38175", - "right": "issue:43502", + "left": "issue:44075", + "right": "issue:44871", + "accept": false, + "reason": "SGD args ignored is unrelated to Gemma-3 eos_token_id configuration mismatch." + }, + { + "left": "issue:44315", + "right": "issue:44360", "accept": false, - "reason": "Different symptoms and code paths: SigLIP2 zero probabilities vs `local_files_only=True` still triggering network requests." + "reason": "Liger Kernel application issue vs DSA indexer ReLU discussion are unrelated." }, { "left": "issue:42890", "right": "issue:44373", "accept": false, - "reason": "Unrelated: a flaky SAM HQ integration test missing seeds vs a docstring fix for `position_ids`." + "reason": "Missing set_seed in tests vs wrong docstring for position_ids are not the same bug." + }, + { + "left": "issue:38175", + "right": "issue:43502", + "accept": false, + "reason": "SIGLIP2 zero probabilities vs unexpected API requests are unrelated." }, { "left": "issue:43299", "right": "issue:44556", "accept": false, - "reason": "Both are version/loading regressions, but they target different checkpoints and failure modes; not the same concrete bug." + "reason": "Both mention loading, but one is Qwen3VL MoE loading and the other is checkpoint reload compatibility after upgrade; not the same fix." }, { "left": "issue:43502", "right": "issue:44961", "accept": false, - "reason": "`local_files_only` network access bug vs an unrelated placeholder/nonsense issue title; no shared underlying change." + "reason": "local_files_only network leakage is unrelated to the 'racoon' issue." }, { "left": "issue:44355", "right": "issue:44655", "accept": false, - "reason": "Compiled Python file errors are unrelated to saving `Pipeline` objects with `save_pretrained`." + "reason": "Compiled-file errors vs saving Pipeline objects are different failures." }, { - "left": "issue:44222", - "right": "issue:44230", + "left": "issue:44355", + "right": "issue:45071", "accept": false, - "reason": "Both mention FP8/MoE, but one is about `save_pretrained` and the other about inference support; different fixes." + "reason": "Compiled-file execution errors are unrelated to PretrainedConfig type-checking regressions." }, { - "left": "issue:44355", - "right": "issue:45071", + "left": "issue:44222", + "right": "issue:44230", "accept": false, - "reason": "Compiled-file loading errors vs `PretrainedConfig` type-checking breakage are distinct issues." + "reason": "FP8 save_pretrained bug vs FP8 inference/support request are related only broadly to fp8, not the same concrete issue." }, { "left": "issue:42831", "right": "issue:43874", "accept": false, - "reason": "FP8 accuracy regression is unrelated to a missing `get_number_of_image_patches` method in an image processor." + "reason": "FP8 accuracy issue vs missing image-processor method are different problems." }, { "left": "issue:42890", "right": "issue:43232", "accept": false, - "reason": "A flaky test seed issue is unrelated to generation kwargs handling after `sync_gpus`." + "reason": "Seeded test flakiness vs generation kwargs after sync_gpus are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing huggingface_hub import vs broken transformers env command are different failures." }, { "left": "issue:33453", "right": "issue:34689", "accept": false, - "reason": "Tokenizer loading regression vs a Llama 3.2 vision model-loading breakage are different concrete failures." + "reason": "Tokenizer loading regression vs specific Llama 3.2 Vision model loading breakage are not the same bug." }, { "left": "issue:33453", "right": "issue:36296", "accept": false, - "reason": "Tokenizer loading regression and tensor-parallel training bug are not the same code-path problem." + "reason": "Tokenizer loading regression is unrelated to tensor-parallel training behavior." }, { "left": "issue:43827", "right": "issue:44961", "accept": false, - "reason": "Docs still referencing `pipeline()` vs an unrelated placeholder issue title; no common bug/change." + "reason": "Docs still referencing pipeline() vs 'racoon' are unrelated." }, { "left": "issue:43576", - "right": "issue:45071", + "right": "issue:43824", "accept": false, - "reason": "Broken `transformers env` command is unrelated to `PretrainedConfig` type-checking changes." + "reason": "Broken v5 env command and a missing model import are both import/API regressions but not the same concrete issue." }, { "left": "issue:43576", - "right": "issue:43824", + "right": "issue:45071", "accept": false, - "reason": "CLI environment command breakage is not the same as a missing top-level export for `Qwen2_5_VLForConditionalGeneration`." + "reason": "CLI/env breakage is unrelated to PretrainedConfig type-checking behavior." }, { "left": "issue:43673", "right": "issue:44492", "accept": false, - "reason": "Missing GenerationMixin cache in chunked prefill is unrelated to a cache-strategy typo." + "reason": "Missing GenerationMixin cache in chunked_prefill vs a typo in cache strategies are different issues." }, { "left": "issue:38617", "right": "issue:43576", "accept": false, - "reason": "Missing `layer_type_validation` export and the broken `transformers env` command are different regressions." + "reason": "ImportError for layer_type_validation is unrelated to the env command regression." }, { "left": "issue:33357", "right": "issue:33453", "accept": false, - "reason": "MacOS bus error with CLIP is unrelated to tokenizer loading regression." - }, - { - "left": "issue:41628", - "right": "issue:42548", - "accept": false, - "reason": "Both are import errors, but they are different missing symbols (`AutoImageProcessor` vs `PreTrainedModel`) and do not look like one fix." + "reason": "MacOS bus error with a community CLIP model vs tokenizer loading regression are unrelated." }, { "left": "issue:44961", "right": "issue:45468", "accept": false, - "reason": "Unrelated issues; one is a placeholder title, the other is a Gemma-4 audio positional encoding bug." + "reason": "'racoon' provides no evidence of being the same as the Gemma-4 audio positional encoding bug." }, { "left": "issue:31795", "right": "issue:38617", "accept": false, - "reason": "Documentation confusion for generation args is unrelated to the `layer_type_validation` import error." + "reason": "Doc confusion in model.forward vs missing layer_type_validation import are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Different missing exports: AutoImageProcessor vs PreTrainedModel." }, { "left": "issue:43576", "right": "issue:44351", "accept": false, - "reason": "Broken `transformers env` CLI and missing `HybridCache` export are distinct problems." + "reason": "Broken env command and missing HybridCache import are separate API/export problems." }, { "left": "issue:44908", "right": "issue:44961", "accept": false, - "reason": "Scheduler kwargs bug is unrelated to the placeholder issue." + "reason": "Scheduler kwargs bug vs 'racoon' are unrelated." }, { - "left": "issue:43576", - "right": "issue:44938", + "left": "issue:36296", + "right": "issue:43901", "accept": false, - "reason": "CLI breakage in v5 and Python 3.14 load failure are different issues." + "reason": "Tensor-parallel training bug vs TextClassificationPipeline docs are unrelated." }, { "left": "issue:44485", "right": "issue:44961", "accept": false, - "reason": "GLM-5 RoPE implementation discussion is unrelated to the placeholder issue." + "reason": "GLM-5 RoPE implementation issue vs 'racoon' are unrelated." }, { "left": "issue:44273", "right": "issue:45003", "accept": false, - "reason": "Lazy loading malfunction and unsafe `sys.modules` access are related only at a very broad level, not the same concrete bug." - }, - { - "left": "issue:36296", - "right": "issue:43901", - "accept": false, - "reason": "Tensor-parallel training bug vs docs mentioning `return_all_scores` are unrelated." + "reason": "Lazy loading malfunction and unsafe sys.modules access may be adjacent, but there is no evidence they are the same concrete bug." }, { "left": "issue:36296", "right": "issue:43526", "accept": false, - "reason": "Tensor-parallel training bug is unrelated to `reduce_labels` returning only one label in BeitImageProcessorFast." - }, - { - "left": "issue:41628", - "right": "issue:44355", - "accept": false, - "reason": "Top-level import error and compiled Python file failures are different regressions." + "reason": "Tensor-parallel training bug vs BeitImageProcessorFast reduce_labels bug are unrelated." }, { "left": "issue:36296", "right": "issue:43012", "accept": false, - "reason": "Tensor-parallel training bug is not the same as a bfloat16 compilation warning." - }, - { - "left": "issue:44623", - "right": "issue:45230", - "accept": false, - "reason": "Missing files from `processor.save_pretrained` is unrelated to a generic bug report placeholder." - }, - { - "left": "issue:44261", - "right": "issue:44360", - "accept": false, - "reason": "Missing `rms_norm_eps` precision issue and DSA indexer ReLU omission are unrelated." - }, - { - "left": "issue:43576", - "right": "issue:44556", - "accept": false, - "reason": "Broken `transformers env` command is unrelated to checkpoint reload failures after upgrading versions." - }, - { - "left": "issue:44393", - "right": "issue:45071", - "accept": false, - "reason": "Qwen3-VL 2D bbox output issue is unrelated to `PretrainedConfig` type-checking breakage." - }, - { - "left": "issue:41669", - "right": "issue:43673", - "accept": false, - "reason": "Removing `import *` usage is a codebase hygiene/performance change, not the same as missing GenerationMixin cache state." + "reason": "Tensor-parallel training bug vs bfloat16 compilation warning are unrelated." }, { - "left": "issue:44038", - "right": "issue:45125", + "left": "issue:41628", + "right": "issue:44355", "accept": false, - "reason": "Qwen3-VL-Moe loading bug and missing `_tp_plan` for tensor parallelism are different concrete model-loading issues." + "reason": "Missing AutoImageProcessor import vs compiled-file execution errors are different problems." } ] }, @@ -12754,7 +12947,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:31795", "issue:33453", @@ -12763,18 +12956,18 @@ "issue:41669", "issue:42371", "issue:42890", - "issue:42913", - "issue:43066", "issue:43352", "issue:43408", "issue:43502", "issue:43519", "issue:43576", "issue:43606", + "issue:43673", "issue:43704", "issue:43723", "issue:43827", "issue:43976", + "issue:44038", "issue:44077", "issue:44246", "issue:44261", @@ -12796,9 +12989,9 @@ "issue:44861", "issue:44908", "issue:44945", - "issue:44998", "issue:45071", "issue:45092", + "issue:45125", "issue:45230", "issue:45310", "issue:45341", @@ -12806,23 +12999,29 @@ "issue:45478" ], "soft_pairs": [ + "issue:44261|issue:44360", + "issue:44623|issue:45230", + "issue:43576|issue:44556", + "issue:41669|issue:43673", + "issue:44393|issue:45071", + "issue:44038|issue:45125", "issue:44556|issue:45092", "issue:42371|issue:44492", "issue:42890|issue:44908", "issue:44704|issue:44908", "issue:44246|issue:44273", - "issue:44297|issue:45230", "issue:43408|issue:44315", + "issue:44297|issue:45230", "issue:43576|issue:44908", "issue:43352|issue:44393", + "issue:44908|issue:45071", "issue:43976|issue:44393", "issue:44295|issue:45230", - "issue:44908|issue:45071", "issue:41669|issue:44351", "issue:41669|issue:45341", - "issue:44315|issue:44829", "issue:44492|issue:44945", "issue:41669|issue:44492", + "issue:44315|issue:44829", "issue:44393|issue:45478", "issue:44077|issue:44861", "issue:31795|issue:33453", @@ -12839,7 +13038,7 @@ "issue:43519|issue:44908", "issue:31795|issue:34689", "issue:31795|issue:37428", - "issue:44315|issue:44393", + "issue:44315|issue:45092", "issue:44273|issue:45071", "issue:43704|issue:44485", "issue:44485|issue:45468", @@ -12850,324 +13049,313 @@ "issue:44263|issue:45468", "issue:44261|issue:45468", "issue:43606|issue:45468", - "issue:42913|issue:43066", - "issue:42890|issue:43704", - "issue:44908|issue:44998" + "issue:42890|issue:43704" ], - "prepared_review_unit_hash": "c78556109a901afe34fa4270267e5f8523ddd7a8bd31402a3885fa28de127045", + "prepared_review_unit_hash": "507882254f35c684cb25de9cb34cd424665c6dd06c48c5b01b8cc02315062963", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { "node_count": 48, "item_count": 48, - "soft_pair_count": 47, - "serialized_chars": 41120, - "estimated_input_tokens": 10280, - "estimated_eval_tokens": 20816 + "soft_pair_count": 51, + "serialized_chars": 42381, + "estimated_input_tokens": 10596, + "estimated_eval_tokens": 21448 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly heterogeneous, but there is a small tokenizer/v5 regression subtheme. I would keep one tokenizer decoder regression as the canonical issue and reject the rest as unrelated or only loosely similar.", - "confidence": 0.55, - "canonical_issue_reason": "issue:43066 is the most concrete tokenizer-v5 regression here: it names a specific wrong decoder type rather than a broad symptom, making it the best anchor for the tokenizer-related subset.", + "summary": "This cluster is mostly a false-positive similarity group: the items share broad themes like v5 regressions, loading, imports, tokenizer/processor APIs, or model-family names, but they describe different bugs and change requests. No soft edge looks like the same underlying issue, so all candidate merges should be rejected.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 44556 is the closest representative of the cluster because it is a concrete, version-bounded v5 regression with a clear reload failure and actionable reproduction. The rest of the items are too heterogeneous to share one true canonical issue.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43066 is the strongest representative issue for the cluster\u2019s only coherent duplicate theme; the others are either broader symptoms or unrelated subsystems.", + "best_issue_reason": "Issue 44556 is the strongest standalone issue here: specific symptom, explicit upgrade boundary, and a reproducible compatibility regression. It is more actionable than the broader or more ambiguous reports.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:44261", + "right": "issue:44360", + "accept": false, + "reason": "Different bugs: one is a precision issue from missing rms_norm_eps, the other is a ReLU omission in a DSA indexer." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "One is a specific save_pretrained file omission; the other is an uninformative generic bug report with no concrete matching symptom." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Different failure modes: broken transformers env command versus checkpoint reload incompatibility after upgrading to v5." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-star performance regressions and missing GenerationMixin cache are unrelated code paths." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bbox hallucination is unrelated to a PretrainedConfig type-checking break in v5.4.0." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Different model-specific issues: Qwen3-VL-Moe versus missing _tp_plan for Qwen3_5Moe tensor parallelism." + }, { "left": "issue:44556", "right": "issue:45092", "accept": false, - "reason": "Both are v5 loading failures, but one is a generic checkpoint reload problem and the other is a remote-code/meta-init incompatibility; not the same concrete bug." + "reason": "Both concern v5 loading compatibility, but they affect different checkpoint types and stem from different initialization paths." }, { "left": "issue:42371", "right": "issue:44492", "accept": false, - "reason": "TF32 API guidance and cache-strategy typo are unrelated." + "reason": "TF32 behavior settings and a cache-strategy typo are unrelated." }, { "left": "issue:42890", "right": "issue:44908", "accept": false, - "reason": "Test flakiness from missing set_seed and an inverse_sqrt scheduler kwargs bug are unrelated." + "reason": "Test seeding failures in sam_hq do not match an inverse_sqrt scheduler kwargs bug." }, { "left": "issue:44704", "right": "issue:44908", "accept": false, - "reason": "AutoProcessor kwargs passthrough and scheduler kwargs handling are different code paths." + "reason": "AutoProcessor kwarg forwarding to cached_file is unrelated to scheduler argument handling." }, { "left": "issue:44246", "right": "issue:44273", "accept": false, - "reason": "Import slowness and lazy-loading failure overlap only loosely; not enough evidence of the same bug." + "reason": "Import-time slowness and lazy-loading malfunction are adjacent but not the same bug." }, { - "left": "issue:44297", - "right": "issue:45230", + "left": "issue:43408", + "right": "issue:44315", "accept": false, - "reason": "Tokenizer_class mismatch in save_pretrained is unrelated to a generic bug report." + "reason": "sam3_video/sam3_tracker mismatch is unrelated to Liger Kernel not being applied via model_init." }, { - "left": "issue:43408", - "right": "issue:44315", + "left": "issue:44297", + "right": "issue:45230", "accept": false, - "reason": "Model-type warning and Liger kernel/model_init behavior are unrelated." + "reason": "Tokenizer_class mismatch on save_pretrained is not the same as a vague generic bug report." }, { "left": "issue:43576", "right": "issue:44908", "accept": false, - "reason": "Broken env command and scheduler kwargs are unrelated." + "reason": "A broken env command and scheduler kwargs being ignored are unrelated regressions." }, { "left": "issue:43352", "right": "issue:44393", "accept": false, - "reason": "Flash Attention support error and Qwen3-VL box-output issue are different problems." + "reason": "Flash Attention 2 support error for NemotronH is unrelated to Qwen3-VL bbox output errors." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs handling and PretrainedConfig type checking are different subsystems and bugs." }, { "left": "issue:43976", "right": "issue:44393", "accept": false, - "reason": "Python version compatibility and Qwen3-VL output errors are unrelated." + "reason": "Python version compatibility and Qwen3-VL output errors have no concrete overlap." }, { "left": "issue:44295", "right": "issue:45230", "accept": false, - "reason": "position_ids buffer error is not the same as a generic bug report." - }, - { - "left": "issue:44908", - "right": "issue:45071", - "accept": false, - "reason": "Scheduler kwargs handling and PretrainedConfig type checking are unrelated." + "reason": "Reading position_ids after buffer registration is unrelated to a generic bug report." }, { "left": "issue:41669", "right": "issue:44351", "accept": false, - "reason": "Import-star performance issue and HybridCache import error are different import regressions." + "reason": "Import-star cleanup and HybridCache import failure are different import-related problems." }, { "left": "issue:41669", "right": "issue:45341", "accept": false, - "reason": "Import-star performance issue and a testing_utils bug are unrelated." - }, - { - "left": "issue:44315", - "right": "issue:44829", - "accept": false, - "reason": "Liger-kernel application and flash_attention_3 training degeneration are different model-training issues." + "reason": "Import performance work and a testing_utils bug do not describe the same defect." }, { "left": "issue:44492", "right": "issue:44945", "accept": false, - "reason": "Cache-strategy typo and pipeline-parallelism output bug are unrelated." + "reason": "A cache-strategy typo is unrelated to incorrect LLM output under pipeline parallelism." }, { "left": "issue:41669", "right": "issue:44492", "accept": false, - "reason": "Import performance and cache-strategy typo are unrelated." + "reason": "Removing import * and a typo in cache strategies are not the same issue." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application during model_init is unrelated to flash_attention_3 training collapse." }, { "left": "issue:44393", "right": "issue:45478", "accept": false, - "reason": "Different Qwen model families and different failure modes; not a duplicate." + "reason": "Different model families and symptoms: Qwen3-VL bbox errors versus Qwen3.5 Moe from_pretrained failure." }, { "left": "issue:44077", "right": "issue:44861", "accept": false, - "reason": "post_init validation and tied-weight-key attribute handling are different correctness issues." + "reason": "post_init policy for patchtsmixer and _get_tied_weight_keys list handling are different code paths." }, { "left": "issue:31795", "right": "issue:33453", "accept": false, - "reason": "Documentation confusion for forward args and tokenizer loading regression are unrelated." + "reason": "Documentation confusion around forward args is unrelated to a tokenizer loading regression." }, { "left": "issue:43827", "right": "issue:44908", "accept": false, - "reason": "Pipeline-removal docs cleanup and inverse_sqrt scheduler kwargs are unrelated." + "reason": "Docs still mentioning pipeline() and scheduler kwargs handling are unrelated." }, { "left": "issue:41669", "right": "issue:44908", "accept": false, - "reason": "Import-star performance and scheduler kwargs are unrelated." + "reason": "Import-star usage and inverse_sqrt scheduler kwargs are unrelated regressions." }, { "left": "issue:43723", "right": "issue:44393", "accept": false, - "reason": "Tokenizer loading in v5 and Qwen3-VL bounding-box output are unrelated." + "reason": "Tokenizer loading in v5 is unrelated to Qwen3-VL bounding-box output issues." }, { "left": "issue:44655", "right": "issue:44861", "accept": false, - "reason": "Pipeline save_pretrained and tied-weight-key crash are unrelated." + "reason": "Saving Pipeline objects and tied-weights-key handling are different APIs with different failures." }, { "left": "issue:43576", "right": "issue:44273", "accept": false, - "reason": "Broken env command and lazy-loading failure are unrelated." + "reason": "Broken env command and lazy loading malfunction are both v5-related but not the same concrete bug." }, { "left": "issue:44360", "right": "issue:45468", "accept": false, - "reason": "DSA indexer activation bug and Gemma-4 audio positional encoding are unrelated." + "reason": "DSA indexer ReLU omission is unrelated to Gemma4AudioRelPositionalEncoding." }, { "left": "issue:45230", "right": "issue:45310", "accept": false, - "reason": "Generic bug report and Qwen3.5 MoE from_pretrained error are not the same issue." + "reason": "A placeholder bug report does not establish the same root cause as the Qwen3.5 Moe from_pretrained error." }, { "left": "issue:44273", "right": "issue:44908", "accept": false, - "reason": "Lazy loading and scheduler kwargs are unrelated." + "reason": "Lazy loading problems and scheduler kwargs handling are unrelated." }, { "left": "issue:41669", "right": "issue:43704", "accept": false, - "reason": "Import-star performance and VRAM leak in dataloader threads are unrelated." + "reason": "Import-time cleanup and VRAM leakage in multithreaded loading are different bugs." }, { "left": "issue:43519", "right": "issue:44485", "accept": false, - "reason": "Qwen3VL timestamp math and GLM-5 RoPE implementation are different model bugs." - }, - { - "left": "issue:42913", - "right": "issue:43066", - "accept": true, - "reason": "Both are tokenizer-v5 regressions; the 'unexpected behavior difference' and 'wrong decoder type' look like the same underlying tokenizer config/decoder bug." - }, - { - "left": "issue:42890", - "right": "issue:43704", - "accept": false, - "reason": "Missing set_seed in tests and VRAM leak across threads are unrelated." - }, - { - "left": "issue:43502", - "right": "issue:43519", - "accept": false, - "reason": "local_files_only network leakage and Qwen3VL timestamp calculation are unrelated." - }, - { - "left": "issue:43519", - "right": "issue:44623", - "accept": false, - "reason": "Timestamp calculation and missing files on processor save are different processor bugs." + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are unrelated model/processor issues." }, { "left": "issue:43519", - "right": "issue:43704", - "accept": false, - "reason": "Timestamp calculation and VRAM leak are unrelated." - }, - { - "left": "issue:44263", - "right": "issue:45468", - "accept": false, - "reason": "torch.split return values and audio positional encoding are unrelated." - }, - { - "left": "issue:44261", - "right": "issue:45468", - "accept": false, - "reason": "MLA q_a_layernorm precision issue and audio positional encoding are unrelated." - }, - { - "left": "issue:43606", - "right": "issue:45468", + "right": "issue:44908", "accept": false, - "reason": "CPU offload device mismatch and Gemma-4 audio positional encoding are unrelated." + "reason": "Timestamp math in a processor is unrelated to scheduler configuration." }, { "left": "issue:31795", "right": "issue:34689", "accept": false, - "reason": "Forward-API docs confusion and Llama 3.2 model loading breakage are unrelated." + "reason": "Forward-arg documentation confusion is unrelated to a Llama 3.2 model loading regression." }, { "left": "issue:31795", "right": "issue:37428", "accept": false, - "reason": "Forward-API docs confusion and flash-attention import error are unrelated." + "reason": "Documentation wording and an ImportError for flash attention support are different classes of issue." }, { "left": "issue:44315", - "right": "issue:44393", + "right": "issue:45092", "accept": false, - "reason": "Liger-kernel application and Qwen3-VL bounding-box output are unrelated." + "reason": "Model_init not applying Liger Kernel is not the same as old remote-code checkpoints failing under meta initialization." }, { "left": "issue:44273", "right": "issue:45071", "accept": false, - "reason": "Lazy loading and PretrainedConfig type checking are unrelated." + "reason": "Lazy loading malfunction and PretrainedConfig type checking are unrelated." }, { "left": "issue:43704", "right": "issue:44485", "accept": false, - "reason": "VRAM leak and GLM-5 RoPE implementation are unrelated." + "reason": "VRAM leakage in dataloader threads is unrelated to GLM-5 RoPE math." }, { "left": "issue:44485", "right": "issue:45468", "accept": false, - "reason": "GLM-5 RoPE implementation and Gemma-4 audio positional encoding are unrelated." + "reason": "GLM-5 RoPE implementation and Gemma4AudioRelPositionalEncoding are different model-specific bugs." }, { "left": "issue:42371", "right": "issue:43704", "accept": false, - "reason": "TF32 API guidance and VRAM leak are unrelated." + "reason": "TF32 API migration advice is unrelated to a VRAM leak in Qwen3ForCausalLM." }, { - "left": "issue:44908", - "right": "issue:44998", + "left": "issue:43502", + "right": "issue:43519", "accept": false, - "reason": "Scheduler kwargs bug and an unrelated placeholder issue are not duplicates." + "reason": "local_files_only network requests and Qwen3VL timestamp errors are unrelated failures." } ] }, - "evaluator_result": { - "accept": true, - "feedback": "The assessment is grounded in the packet and stays conservative overall. The accepted tokenizer pair is plausible from the titles alone, and the rejected pairs mostly link clearly different subsystems or failure modes. The summary is a little strong in calling it the cluster\u2019s only coherent duplicate theme, but the follow-up sentence tempers that appropriately." - }, - "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 }, @@ -13182,7 +13370,7 @@ "source": "llm", "cache_hit": false, "model": "gpt-5.4-mini?service_tier=flex", - "cluster_id": "cluster-15354-359", + "cluster_id": "cluster-15354-362", "nodes": [ "issue:31515", "issue:31795", @@ -13190,6 +13378,8 @@ "issue:36296", "issue:36683", "issue:38617", + "issue:42913", + "issue:43066", "issue:43298", "issue:43525", "issue:43572", @@ -13199,12 +13389,15 @@ "issue:44556", "issue:44623", "issue:44861", + "issue:44908", "issue:44998", "issue:45290", "issue:45310", "issue:45341" ], "soft_pairs": [ + "issue:42913|issue:43066", + "issue:44908|issue:44998", "issue:44861|issue:45290", "issue:44998|issue:45341", "issue:44492|issue:44998", @@ -13215,130 +13408,142 @@ "issue:31515|issue:38617", "issue:44230|issue:44623", "issue:31515|issue:36296", - "issue:44230|issue:45310", "issue:31515|issue:31795", + "issue:44230|issue:45310", "issue:31515|issue:33453", - "issue:43298|issue:44230", - "issue:31515|issue:36683" + "issue:31515|issue:36683", + "issue:43298|issue:44230" ], - "prepared_review_unit_hash": "fd165d3c1aec28a570268b53d5a521af4d7cb0256340fb58cd96b7ac67116dca", + "prepared_review_unit_hash": "0bfa4f0163b55a2ca1d55f16044d944c4291f1439bbf2e4d99123e9095f93e13", "component_budget": { - "node_count": 359, - "item_count": 359, - "soft_pair_count": 1335, - "serialized_chars": 689788, - "estimated_input_tokens": 172447, - "estimated_eval_tokens": 345150 + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 }, "budget": { - "node_count": 19, - "item_count": 19, - "soft_pair_count": 15, - "serialized_chars": 15101, - "estimated_input_tokens": 3776, - "estimated_eval_tokens": 7808 + "node_count": 22, + "item_count": 22, + "soft_pair_count": 17, + "serialized_chars": 17367, + "estimated_input_tokens": 4342, + "estimated_eval_tokens": 8940 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "No true duplicate cluster here: the paired items are mostly unrelated issues spanning loading performance, config attribute errors, docs, audio support, and typo/cleanup reports. None look like the same underlying bug or change.", - "confidence": 0.93, - "canonical_issue_reason": "No single canonical issue: the soft-linked items are heterogeneous and do not share the same concrete failure mode or code path.", + "summary": "Mostly a loose thematically related set of loading/tokenizer/config regression issues, but the soft pairs do not describe the same underlying bug. No PRs are present.", + "confidence": 0.83, + "canonical_issue_reason": "issue:44556 is the clearest, most self-contained regression report in the set: checkpoint reload breaks after upgrading from v4.57 to v5.2/v5.3. It best represents the broad loading/version-compatibility theme, even though the rest are not true duplicates.", "canonical_pr_reason": null, - "best_issue_reason": "issue:31515 is the closest broad anchor because it is the clearest from_pretrained/checkpoint-loading report, but it still does not match the rest closely enough to serve as a real duplicate canonical.", + "best_issue_reason": "issue:44556 is the strongest standalone triage candidate because it has a concrete versioned repro and a specific failure mode. The other issues are either more specialized, more vague, or clearly unrelated despite surface similarity.", "best_pr_reason": null, "soft_edge_verdicts": [ + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both mention tokenizer v5 behavior, but one is a general behavior change from v4 to v5 and the other is a specific wrong decoder type problem. Not the same bug." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Completely unrelated; scheduler kwargs vs a nonsensical unrelated issue title. No shared concrete bug." + }, { "left": "issue:44861", "right": "issue:45290", "accept": false, - "reason": "Different bugs: tied-weights key handling vs chat template crashes on assistant tool-call messages." + "reason": "Different code paths: tied-weight key extraction crash vs chat template handling of tool-call assistant messages." }, { "left": "issue:44998", "right": "issue:45341", "accept": false, - "reason": "Unrelated: a spammy 'Unemployment' issue versus a testing_utils bug." + "reason": "No meaningful overlap; unrelated generic issue versus a testing_utils bug." }, { "left": "issue:44492", "right": "issue:44998", "accept": false, - "reason": "A typo in cache strategy docs/config is not the same as the unrelated unemployment issue." + "reason": "A typo in cache strategies is not the same as the other issue; no shared underlying defect." }, { "left": "issue:43618", "right": "issue:44998", "accept": false, - "reason": "CLIPOutput attentions regression is unrelated to the junk unemployment issue." + "reason": "CLIP attentions assignment regression is unrelated to the other issue." }, { "left": "issue:44556", "right": "issue:44998", "accept": false, - "reason": "Checkpoint reload regression and an unrelated non-technical issue are clearly different." + "reason": "Checkpoint reload/version upgrade regression does not match the unrelated issue." }, { "left": "issue:43525", "right": "issue:44861", "accept": false, - "reason": "Missing pad_token_id in Llama4Config is a config attribute bug, not tied-weight key crashing." + "reason": "Different model/config attribute errors in different subsystems; not the same bug." }, { "left": "issue:43572", "right": "issue:44861", "accept": false, - "reason": "StableLmConfig pad_token_idx removal/mismatch is a different config problem than tied-weight key handling." + "reason": "StableLmConfig pad_token_idx regression and tied-weights crash are different failures." }, { "left": "issue:31515", "right": "issue:38617", "accept": false, - "reason": "Slow from_pretrained checkpoint loading is unrelated to an ImportError for layer_type_validation." + "reason": "Checkpoint loading slowness and an ImportError for layer_type_validation are unrelated." }, { "left": "issue:44230", "right": "issue:44623", "accept": false, - "reason": "FP8 inference support for Qwen models is not the same as processor.save_pretrained missing files." + "reason": "fp8 inference support for a vision-language model is not the same as missing files from processor.save_pretrained." }, { "left": "issue:31515", "right": "issue:36296", "accept": false, - "reason": "Checkpoint loading slowdown and tensor-parallel training bug are different failure modes." + "reason": "from_pretrained load performance and tensor-parallel training are different problems." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Loading slow checkpoint and documentation confusion in model.forward are not the same underlying issue." }, { "left": "issue:44230", "right": "issue:45310", "accept": false, - "reason": "FP8 inference support request is not the same concrete bug as Qwen3.5 Moe from_pretrained failure." + "reason": "Both involve Qwen-family models, but one is fp8 support and the other is a from_pretrained error; not the same concrete bug." }, { "left": "issue:31515", - "right": "issue:31795", + "right": "issue:33453", "accept": false, - "reason": "Checkpoint loading performance is unrelated to forward() documentation confusion." + "reason": "Checkpoint loading slowness and tokenizer loading regression are different failure modes." }, { "left": "issue:31515", - "right": "issue:33453", + "right": "issue:36683", "accept": false, - "reason": "Loading slowdown and tokenizer-loading regression are both loading-related but not the same bug or fix." + "reason": "from_pretrained loading slowdown is unrelated to Gemma3Config missing vocab_size." }, { "left": "issue:43298", "right": "issue:44230", "accept": false, - "reason": "Continuous batching audio-model support and Qwen FP8 inference support are different feature gaps." - }, - { - "left": "issue:31515", - "right": "issue:36683", - "accept": false, - "reason": "from_pretrained loading slowdown is not the same as a Gemma3Config missing vocab_size attribute error." + "reason": "Continuous batching for audio models and fp8 Qwen VL inference support are distinct issues with different code paths." } ] }, @@ -13368,12 +13573,12 @@ "soft_pairs": [ "issue:26350|issue:33260" ], - "prepared_review_unit_hash": "aaf6c048daf322a888ee8aa4206847c55185a3f69683e0795283f8b25dc47495", + "prepared_review_unit_hash": "f9a3a55792236d149b1577ae1deb073396836987c469fd94f49585aee6d8da0e", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -13381,7 +13586,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1589, + "serialized_chars": 1590, "estimated_input_tokens": 398, "estimated_eval_tokens": 1052 }, @@ -13390,18 +13595,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate community feature requests: one for Flash Attention 2 support on more architectures, and one for GGUF support on more architectures. They are thematically similar but target different integrations and code paths, so they should not be clustered as duplicates.", - "confidence": 0.97, - "canonical_issue_reason": null, + "summary": "Two distinct feature-request issues about adding support for more architectures, but they concern different features (Flash Attention 2 vs GGUF) and should not be deduplicated.", + "confidence": 0.98, + "canonical_issue_reason": "issue:26350 is the more established thread (older and much higher discussion activity), but it is not a duplicate of issue:33260.", "canonical_pr_reason": null, - "best_issue_reason": null, + "best_issue_reason": "issue:26350 is the better representative of the pair because it has substantially more discussion and visibility.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:26350", "right": "issue:33260", "accept": false, - "reason": "Both are broad \u2018support for more architectures\u2019 requests, but they concern different features (Flash Attention 2 vs GGUF) and would require different implementation work. Not the same underlying issue." + "reason": "They share only a generic \"support for more architectures\" theme; one is about Flash Attention 2 and the other about GGUF, so they are different feature requests and not the same underlying issue." } ] }, @@ -13438,12 +13643,12 @@ "issue:43834|issue:44365", "issue:44365|issue:44370" ], - "prepared_review_unit_hash": "cc91f282234355fa2b4aaa9604c184bfd8f6e8e836d922aa67b3b39faf4d4a00", + "prepared_review_unit_hash": "29e19d4a4c3511a0f79f4ad09609f48e8b520fa651a382a95a5096271f83b0ab", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4455, + "serialized_chars": 4454, "estimated_input_tokens": 1114, "estimated_eval_tokens": 2484 }, @@ -13451,7 +13656,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4455, + "serialized_chars": 4454, "estimated_input_tokens": 1114, "estimated_eval_tokens": 2484 }, @@ -13460,48 +13665,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are all docs-translation request issues, but they refer to different language targets. The Turkish issue is concrete; the others are generic template-like placeholders, so they are not duplicates of the Turkish task or of each other.", - "confidence": 0.97, - "canonical_issue_reason": "issue:27088 is the only concrete, language-specific report ([i18n-TR] Turkish). It is the clearest representative of this cluster\u2019s actual content.", + "summary": "These are all documentation translation request issues, but they are not the same underlying task: #27088 is specifically for Turkish, while the others are generic template issues for unspecified languages. The high text similarity appears to come from the shared issue template, not duplicate content.", + "confidence": 0.96, + "canonical_issue_reason": "#27088 is the only concrete, language-specific issue (Turkish). The other issues are generic placeholders and do not describe the same exact request.", "canonical_pr_reason": null, - "best_issue_reason": "issue:27088 is the most specific and actionable issue, with an explicit target language instead of a generic template title.", + "best_issue_reason": "#27088 is the best representative issue because it is specific and actionable, whereas the others are broad template entries without a concrete target language.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:27088", "right": "issue:44370", "accept": false, - "reason": "Different language targets; these are separate translation tasks, not the same underlying issue." + "reason": "Same translation-request template, but #27088 is Turkish and #44370 is unspecified; different underlying requests." }, { "left": "issue:27088", "right": "issue:44365", "accept": false, - "reason": "Different language targets; the generic placeholder issue is not a duplicate of the Turkish translation request." + "reason": "Shared template wording only; one is a Turkish docs translation request and the other is a generic language placeholder." }, { "left": "issue:27088", "right": "issue:43834", "accept": false, - "reason": "Different language targets; same template pattern but not the same concrete translation task." + "reason": "Both are translation issues, but they target different languages / one is generic, so they are not duplicates." }, { "left": "issue:43834", "right": "issue:44370", "accept": false, - "reason": "Both are generic template issues, but there is no evidence they refer to the same actual language or task." + "reason": "Near-identical placeholders, but they do not identify the same concrete language task." }, { "left": "issue:43834", "right": "issue:44365", "accept": false, - "reason": "Both are generic template issues and may be placeholders, but they do not identify the same concrete translation work." + "reason": "Template similarity only; no evidence they refer to the same specific translation work." }, { "left": "issue:44365", "right": "issue:44370", "accept": false, - "reason": "Both are generic template issues; insufficient basis to treat them as the same underlying request." + "reason": "Both are generic translation placeholders, but there is no concrete shared language target to establish duplication." } ] }, @@ -13566,124 +13771,124 @@ "issue:43837|issue:45027", "issue:43837|issue:45177" ], - "prepared_review_unit_hash": "da96ffbf8df2ec21aa5d985f6f0762ce616a0a92b88dd9ca0f0cf032be085ebb", + "prepared_review_unit_hash": "f71a878ee9e7b8b02ded57ea5d61c29897926f71e62ca549d747943458c1407a", "component_budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 25, - "serialized_chars": 16006, - "estimated_input_tokens": 4002, - "estimated_eval_tokens": 8260 + "serialized_chars": 16010, + "estimated_input_tokens": 4003, + "estimated_eval_tokens": 8262 }, "budget": { "node_count": 13, "item_count": 13, "soft_pair_count": 25, - "serialized_chars": 16006, - "estimated_input_tokens": 4002, - "estimated_eval_tokens": 8260 + "serialized_chars": 16010, + "estimated_input_tokens": 4003, + "estimated_eval_tokens": 8262 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All items are distinct model-support or model-addition requests; the soft pairs share broad theme only and do not look like the same underlying issue/change. No duplicate merge is warranted.", - "confidence": 0.94, + "summary": "These issues are all broad model-addition requests, but they target different models and capabilities (BERT variants, audio/TTS/ASR, OCR, sparse transformers, etc.). They share a template-level similarity, not the same underlying change, so none should be deduplicated or merged into one canonical issue.", + "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "issue:43671 is the clearest representative request: it is concrete, specific to one model family (Qwen3-TTS), and has more discussion/inbound attention than most others, but it is not a duplicate of the rest.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:36817", "right": "issue:41211", "accept": false, - "reason": "Both are model-addition requests, but EuroBERT and DEIMv2 are unrelated models with different targets." + "reason": "EuroBert vs DEIMv2 are different model requests; same broad \u201cadd model\u201d theme only." }, { "left": "issue:36817", "right": "issue:43671", "accept": false, - "reason": "EuroBERT config support and Qwen3-TTS support are different model families and different capabilities." + "reason": "EuroBert config addition and Qwen3-TTS support are unrelated model types and code paths." }, { "left": "issue:37015", "right": "issue:43837", "accept": false, - "reason": "NeoBERT and Qwen3-ASR are unrelated requests; one is a BERT variant, the other is speech recognition." + "reason": "NeoBERT and Qwen3-ASR are distinct models; not the same feature or fix." }, { "left": "issue:37015", "right": "issue:44121", "accept": false, - "reason": "NeoBERT and OpenAI weight-sparse transformer/circuitgpt are different model requests with no shared concrete change." + "reason": "NeoBERT and weight-sparse transformer support are different architectures and implementation tasks." }, { "left": "issue:37015", "right": "issue:43909", "accept": false, - "reason": "NeoBERT and LFM2.5 Audio are different architectures and different feature areas." + "reason": "NeoBERT vs LFM2.5 Audio is unrelated model support, despite similar request wording." }, { "left": "issue:37015", "right": "issue:42503", "accept": false, - "reason": "NeoBERT and ModernVBERT are both BERT-like, but they are distinct model requests, not the same change." + "reason": "NeoBERT and ModernVBERT are different model additions, not one duplicate request." }, { "left": "issue:37015", "right": "issue:45027", "accept": false, - "reason": "NeoBERT and Voxtral TTS support are unrelated model additions in different domains." + "reason": "NeoBERT and Voxtral TTS support target different modalities and code paths." }, { "left": "issue:41211", "right": "issue:43671", "accept": false, - "reason": "DEIMv2 and Qwen3-TTS are different model requests with no common implementation path." + "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests; only the request template is similar." }, { "left": "issue:42503", "right": "issue:43837", "accept": false, - "reason": "ModernVBERT models and Qwen3-ASR are separate requests; one is text/encoder-centric, the other speech ASR." + "reason": "ModernVBERT models and Qwen3-ASR support are different features and model families." }, { "left": "issue:42503", "right": "issue:43671", "accept": false, - "reason": "ModernVBERT and Qwen3-TTS are different model families and would not be merged as one PR." + "reason": "ModernVBERT and Qwen3-TTS are separate model-support requests with no shared concrete change." }, { "left": "issue:42503", "right": "issue:42738", "accept": false, - "reason": "ModernVBERT models and BERT-like models with RoPE overlap only broadly; the requested additions are not the same concrete model." + "reason": "ModernVBERT and BERT-like models with RoPE are related only at a high level; not the same requested addition." }, { "left": "issue:42503", "right": "issue:45027", "accept": false, - "reason": "ModernVBERT and Voxtral TTS are unrelated model-support requests." + "reason": "ModernVBERT and Voxtral TTS support are different model families and implementation areas." }, { "left": "issue:42738", "right": "issue:45027", "accept": false, - "reason": "BERT-like RoPE models and Voxtral TTS are too broad and target different model types." + "reason": "BERT-like RoPE support and Voxtral TTS support are not the same underlying change." }, { "left": "issue:43671", "right": "issue:43837", "accept": false, - "reason": "Same vendor/family naming, but TTS and ASR are different concrete capabilities and would require different changes." + "reason": "Qwen3-TTS and Qwen3-ASR are both Qwen3 requests, but they are different modalities and separate support work." }, { "left": "issue:43671", "right": "issue:43909", "accept": false, - "reason": "Qwen3-TTS and LFM2.5 Audio 1.5B are distinct audio-model support requests." + "reason": "Qwen3-TTS support and LFM2.5 Audio support are different models, not duplicates." }, { "left": "issue:43671", @@ -13695,55 +13900,55 @@ "left": "issue:43671", "right": "issue:45027", "accept": false, - "reason": "Both are audio-related, but Qwen3-TTS and Voxtral TTS are different concrete model integrations." + "reason": "Qwen3-TTS and Voxtral TTS are distinct model integrations with different code paths." }, { "left": "issue:43671", "right": "issue:45177", "accept": false, - "reason": "Qwen3-TTS support and DeepSeek-OCR2 are unrelated modalities and code paths." + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model additions." }, { "left": "issue:43837", "right": "issue:43909", "accept": false, - "reason": "Qwen3-ASR and LFM2.5 Audio are different speech/audio model additions." + "reason": "Qwen3-ASR and LFM2.5 Audio are separate audio-model requests, not one concrete fix." }, { "left": "issue:43837", "right": "issue:45322", "accept": false, - "reason": "Qwen3-ASR and EUPE are unrelated model requests; one is ASR, the other is an encoder." + "reason": "Qwen3-ASR and EUPE are different model families and different implementation targets." }, { "left": "issue:43837", "right": "issue:44121", "accept": false, - "reason": "Qwen3-ASR and weight-sparse transformer support are different problems." + "reason": "Qwen3-ASR support and weight-sparse transformer support do not describe the same change." }, { "left": "issue:43909", "right": "issue:44205", "accept": false, - "reason": "LFM2.5 Audio and SAM3-LiteText are unrelated model requests." + "reason": "LFM2.5 Audio and SAM3-LiteText are unrelated model requests across different modalities." }, { "left": "issue:44121", "right": "issue:45322", "accept": false, - "reason": "OpenAI weight-sparse transformer support and EUPE are different model additions with no shared concrete bug/change." + "reason": "Weight-sparse transformer support and EUPE addition are different model-architecture requests." }, { "left": "issue:43837", "right": "issue:45027", "accept": false, - "reason": "Qwen3-ASR and Voxtral-4B-TTS are both speech-related, but they are different models and different integrations." + "reason": "Qwen3-ASR and Voxtral TTS are separate speech-model support requests." }, { "left": "issue:43837", "right": "issue:45177", "accept": false, - "reason": "Qwen3-ASR and DeepSeek-OCR2 are unrelated model requests." + "reason": "Qwen3-ASR and DeepSeek-OCR2 are unrelated model additions." } ] }, @@ -13784,82 +13989,82 @@ "issue:43668|issue:44031", "issue:43668|issue:43813" ], - "prepared_review_unit_hash": "adfee87fd10a6f609a0cf5f5dd8b7fccee791790523526507cf7bc368d336366", + "prepared_review_unit_hash": "f1e950ab26a8e3227ccedd5060213ad47cbee1171dddcad42dcb1a1bab340343", "component_budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6292, - "estimated_input_tokens": 1573, - "estimated_eval_tokens": 3402 + "serialized_chars": 6293, + "estimated_input_tokens": 1574, + "estimated_eval_tokens": 3404 }, "budget": { "node_count": 6, "item_count": 6, "soft_pair_count": 8, - "serialized_chars": 6292, - "estimated_input_tokens": 1573, - "estimated_eval_tokens": 3402 + "serialized_chars": 6293, + "estimated_input_tokens": 1574, + "estimated_eval_tokens": 3404 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The cluster is heterogeneous: these issues cover unrelated model behavior, config typing, a typo, and tokenizer warnings, so they are not duplicates.", - "confidence": 0.97, - "canonical_issue_reason": "No clear canonical issue: the items describe different bugs in different components, not one shared underlying defect.", + "summary": "No duplicate cluster here: these issues cover unrelated problems across model configs, tokenization warnings, loss wiring, and a typo. None of the soft-similarity links look like the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": "No clear canonical issue; the reports are heterogeneous and do not describe the same defect or code path.", "canonical_pr_reason": null, - "best_issue_reason": "If one issue must anchor the cluster, #43240 is the broadest framework-level bug, but it still does not match the other reports.", + "best_issue_reason": "No single best issue for deduping because this set does not form a duplicate cluster; each item appears to be a distinct bug or typo.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:37934", "right": "issue:43668", "accept": false, - "reason": "Different components and symptoms: Llama4 norm implementation vs ModernBERT config type hint." + "reason": "Different model areas and bug types: Llama4TextL2Norm behavior question vs ModernBERT config type hint." }, { "left": "issue:42936", "right": "issue:43240", "accept": false, - "reason": "Mask2former ignore_value handling is unrelated to loss kwargs propagation." + "reason": "Mask2Former ignore_value wiring is unrelated to kwargs not being passed into a loss function." }, { "left": "issue:42936", "right": "issue:44031", "accept": false, - "reason": "Model-specific ignore_value bug vs tokenizer regex warning after a version bump." + "reason": "Mask2Former loss/ignore_value behavior is unrelated to tokenizer regex warnings." }, { "left": "issue:42936", "right": "issue:43813", "accept": false, - "reason": "Mask2former loss parameter issue is unrelated to a typo in a quantization attribute name." + "reason": "A model loss-parameter issue is not the same as a typo in a quantization attribute name." }, { "left": "issue:42936", "right": "issue:43668", "accept": false, - "reason": "Different model areas: unused ignore_value vs incorrect config type hint." + "reason": "These are separate issues: one about ignore_value usage, the other about a config type annotation." }, { "left": "issue:43240", "right": "issue:43668", "accept": false, - "reason": "Passing kwargs to loss calculation is a runtime behavior bug, not a type-hint issue." + "reason": "Loss-function kwargs propagation and config type hints do not describe the same bug." }, { "left": "issue:43668", "right": "issue:44031", "accept": false, - "reason": "ModernBERT config typing and tokenizer regex warnings are separate problems." + "reason": "Incorrect type hint in a config class is unrelated to tokenizer regex warnings after a version bump." }, { "left": "issue:43668", "right": "issue:43813", "accept": false, - "reason": "Config type annotation mismatch does not match a simple typo in a different code path." + "reason": "Type-hint correction and a spelling error in a quantization field are distinct fixes." } ] }, @@ -13889,12 +14094,12 @@ "soft_pairs": [ "issue:41084|issue:43037" ], - "prepared_review_unit_hash": "2c3661ea485823452c8fda7c4dc3cc85d434aed57fd1b4466414ff0aadf1e892", + "prepared_review_unit_hash": "152c7c4502d4952876d0b1c12aedba74685aa45e5052ccf05fe1d4b9fcfa9072", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -13902,7 +14107,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 1479, + "serialized_chars": 1480, "estimated_input_tokens": 370, "estimated_eval_tokens": 996 }, @@ -13911,18 +14116,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two issues are not duplicates: one asks for generic block decoding support, while the other requests DeepSeek v3.2 support. They describe different scopes and likely different implementation work.", - "confidence": 0.95, - "canonical_issue_reason": "No single canonical issue: 41084 is a broad decoding feature request, while 43037 is a model-specific support request for DeepSeek v3.2.", + "summary": "These two issues do not look like duplicates: one is a terse request about block decoding, the other is a specific request for DeepSeek v3.2 support. The only linkage is weak textual similarity, not a shared underlying problem.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue: the issues describe different topics and do not appear to report the same bug or feature request.", "canonical_pr_reason": null, - "best_issue_reason": "43037 is the more concrete and actionable issue; 41084 is too vague to serve as a strong representative.", + "best_issue_reason": "Issue 43037 is the better representative of the cluster only because it is more specific and actionable, but it is not a duplicate of 41084.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:41084", "right": "issue:43037", "accept": false, - "reason": "Different underlying requests and code paths: generic block decoding vs DeepSeek v3.2 support. Similarity is only superficial, so they should not be merged." + "reason": "Different subjects: block decoding implementation vs DeepSeek v3.2 support. Similar wording is too weak to indicate the same underlying issue." } ] }, @@ -13952,7 +14157,7 @@ "soft_pairs": [ "issue:43598|issue:43600" ], - "prepared_review_unit_hash": "06fca2f379985d728236dd16de09c917f28e1c9b3756c5a128045ccccbbeb088", + "prepared_review_unit_hash": "4827160b2fc43dac15415836da2e6ae3728ad9186c7e56ce47b13be177f9d6f6", "component_budget": { "node_count": 2, "item_count": 2, @@ -13974,18 +14179,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two issues are both about Trainer extensibility, but they describe different problems: one is about `remove_unused_column` behavior/customizability, the other about FLOPS estimation for custom trainers. They do not look like duplicates.", - "confidence": 0.94, - "canonical_issue_reason": "No clear canonical issue: the issues target different Trainer behaviors and are not the same bug.", + "summary": "These are related only in the broad sense that both mention Trainer customization, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. I would not cluster them as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "Neither issue is a strong duplicate representative, but `#43598` is the more concrete API-behavior request and better serves as a standalone issue.", + "best_issue_reason": "Neither issue is a duplicate of the other; if forced to pick one representative, #43598 is the clearer, more specific customization request around `remove_unused_column`.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43598", "right": "issue:43600", "accept": false, - "reason": "Different underlying problems: one concerns `remove_unused_column` customizability, the other concerns FLOPS estimation for custom trainers. Shared trainer context is too broad to treat as the same issue." + "reason": "Different underlying bugs/features: #43598 is about `remove_unused_column` in Trainer, while #43600 is about estimating FLOPS for custom trainers. Shared Trainer context is too broad to treat as the same issue." } ] }, @@ -14015,7 +14220,7 @@ "soft_pairs": [ "issue:43684|issue:44405" ], - "prepared_review_unit_hash": "c228be1521990ed314b34580fe1c9670fddaf80e7c0c2225989e6d136e529a82", + "prepared_review_unit_hash": "3ed92323afdf1c8dd0bd0747c10a07d0c6fe83f048fb3157bdc7b41bca4e84ab", "component_budget": { "node_count": 2, "item_count": 2, @@ -14037,18 +14242,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items are closed issues about adding Hugging Face model support, but they target different models and different registration paths: Qwen3-Omni AutoModel/AutoModelForConditionalGeneration vs Qwen3.5 AutoModelForSequenceClassification. They are related in theme but not duplicates.", - "confidence": 0.93, - "canonical_issue_reason": "Issue 43684 is the broader request and names the explicit registration gap for Qwen3-Omni in AutoModel and AutoModelForConditionalGeneration, making it the better umbrella issue for this cluster.", + "summary": "These issues are about different Qwen model capability gaps: one is model registration for Qwen3-Omni, the other is sequence classification support for Qwen3.5. They are not the same underlying bug/change, so the soft duplicate link should be rejected.", + "confidence": 0.97, + "canonical_issue_reason": null, "canonical_pr_reason": null, - "best_issue_reason": "43684 is the best representative because it captures the more general model-registration support problem; 44405 is a narrower, model-specific classification-head request.", + "best_issue_reason": null, "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43684", "right": "issue:44405", "accept": false, - "reason": "Reject: they concern different model families (Qwen3-Omni vs Qwen3.5/Qwen3_5Config) and different APIs (general/conditional generation vs sequence classification). Shared theme is only model support registration, not the same underlying bug/request." + "reason": "Different model families and different requested functionality (AutoModel registration vs AutoModelForSequenceClassification support); not mergeable as one duplicate fix." } ] }, @@ -14085,80 +14290,80 @@ "issue:43845|issue:43859", "issue:43846|issue:43859" ], - "prepared_review_unit_hash": "326262ddbfa7659213b0f549b61914c101946f82c629fc23b7c2f69999ecb1bc", + "prepared_review_unit_hash": "af74e354797a6e25483db37f945d0059fa1d72af2f9f9e138305c8d4b010815c", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4266, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 4266, - "estimated_input_tokens": 1067, - "estimated_eval_tokens": 2390 + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Three near-identical closed issues titled \"huggingface\" look like duplicates of each other; the \"Fett\" issue appears unrelated. I\u2019d keep the most active/recent huggingface report as the representative issue.", - "confidence": 0.79, - "canonical_issue_reason": "issue:43859 matches the same exact-title huggingface cluster and has the only discussion activity, so it is the best representative to retain.", + "summary": "The three \"huggingface \" issues form a weak duplicate-looking cluster, but the evidence is only soft similarity with no deterministic accept flags. \"Fett\" still looks separate. Keep the cluster representative as 43845 if needed, but do not treat the pairwise links as confirmed duplicates.", + "confidence": 0.73, + "canonical_issue_reason": "issue:43845 is the earliest of the near-identical low-information \"huggingface \" reports, so it is the best representative for that tentative cluster, though the linkage evidence is weak.", "canonical_pr_reason": null, - "best_issue_reason": "issue:43859 is the most useful survivor of the duplicate huggingface trio because it has at least some discussion activity and is otherwise identical in title.", + "best_issue_reason": "issue:43845 is the strongest global issue representative because it is the earliest of the repeated low-context reports and fits the tentative duplicate set best.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:43835", "right": "issue:43845", "accept": false, - "reason": "Different titles and no body details linking them to the same report; likely unrelated." + "reason": "Only soft similarity and different titles; not enough evidence they are the same underlying issue." }, { "left": "issue:43835", "right": "issue:43846", "accept": false, - "reason": "Different titles and no concrete evidence they describe the same issue." + "reason": "Only soft similarity with no shared concrete details or target; keep separate." }, { "left": "issue:43835", "right": "issue:43859", "accept": false, - "reason": "Different titles with no shared context beyond being closed issues." + "reason": "Different titles and no substantive overlapping evidence of the same bug." }, { "left": "issue:43845", "right": "issue:43846", - "accept": true, - "reason": "Same exact title and very similar metadata; these look like duplicate reports of the same issue." + "accept": false, + "reason": "They look similar, but the packet only provides soft similarity and deterministic_accept is false, so this should stay tentative at most." }, { "left": "issue:43845", "right": "issue:43859", - "accept": true, - "reason": "Same exact title and highly similar cluster signals; likely duplicate reports of the same underlying issue." + "accept": false, + "reason": "High textual similarity, but no concrete evidence beyond soft similarity; do not accept as a confirmed duplicate pair." }, { "left": "issue:43846", "right": "issue:43859", - "accept": true, - "reason": "Same exact title and highly similar cluster signals; likely duplicate reports of the same underlying issue." + "accept": false, + "reason": "Indistinguishable low-information titles, but still only soft similarity with no deterministic acceptance; reject as unconfirmed." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The clustering is grounded in the packet: 43845, 43846, and 43859 share the exact title and have the strongest soft-similarity signals, while 43835 (\u201cFett\u201d) is clearly distinct by title and should stay out. Choosing 43859 as the canonical survivor is conservative because it is the only one with discussion activity. The soft-edge acceptances are also cautious enough given the identical titles and high similarity, and the rejections involving 43835 are well supported." + "accept": false, + "feedback": "The cluster summary is mostly grounded, but the soft-edge verdicts are too strong. All pairwise links are supported only by soft_similarity, and the packet marks every pair as deterministic_accept=false, so accepting issue:43845 with issue:43846 and issue:43859 overstates the evidence. The conservative choice here would be to keep those pairs rejected or at most tentative, not accepted." }, "evaluator_used": true, - "retried": false, - "accepted_nontrivial_soft_edge": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, "error_kind": null, "error_message": null }, @@ -14192,88 +14397,88 @@ "issue:44572|issue:44937", "issue:44573|issue:44937" ], - "prepared_review_unit_hash": "3c9726f83b8d85e5cb7d0cd87c486f7412e87becc21ffc47a643fdac9a8dd38a", + "prepared_review_unit_hash": "aa2ee6e0cc0fc5d4a7fa47dca1935fec1664025e226a1d2469f7387c5bce3958", "component_budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5791, - "estimated_input_tokens": 1448, - "estimated_eval_tokens": 3152 + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "budget": { "node_count": 5, "item_count": 5, "soft_pair_count": 9, - "serialized_chars": 5791, - "estimated_input_tokens": 1448, - "estimated_eval_tokens": 3152 + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This cluster looks like unrelated low-information/spam issues rather than duplicates of a single bug or request. No pair shows enough shared substance to merge.", - "confidence": 0.95, - "canonical_issue_reason": "Issue #44418 is the only non-placeholder title ('Documentation Enhancement Suggestion'); the others are generic spam placeholders or unrelated naming spam.", + "summary": "All items are low-signal and appear unrelated spam or a generic documentation request. No soft pair has enough evidence to be treated as the same underlying bug or change.", + "confidence": 0.23, + "canonical_issue_reason": "issue:44418 is the only item with a non-placeholder, somewhat substantive title; the others are clearly spam-like and provide no usable bug/change signal.", "canonical_pr_reason": null, - "best_issue_reason": "#44418 is the least spam-like and most plausible as a real issue, though it still lacks enough detail to strongly justify duplication grouping.", + "best_issue_reason": "issue:44418 is the best available issue for triage because it is the most descriptive and least obviously spammy, even though it still lacks concrete technical detail.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:44371", "right": "issue:44572", "accept": false, - "reason": "Both are generic spam placeholders with no concrete shared problem or target." + "reason": "Both are generic '' issues with no concrete bug, code path, or change described; identical placeholder titles are not enough to infer the same underlying problem." }, { "left": "issue:44371", "right": "issue:44573", "accept": false, - "reason": "Generic spam placeholders only; no evidence they describe the same underlying issue." + "reason": "Both look like separate spam submissions rather than the same bug or change; there is no shared technical subject beyond the placeholder title." }, { "left": "issue:44371", "right": "issue:44418", "accept": false, - "reason": "One is spam-like, the other is a vague documentation suggestion; not the same issue." + "reason": "A spam placeholder issue and a documentation enhancement request are not the same underlying problem." }, { "left": "issue:44371", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder versus unrelated quoted-name title; no shared bug/change." + "reason": "Generic spam versus a separate spam-like promotional title; no evidence of one concrete issue or fix." }, { "left": "issue:44418", "right": "issue:44572", "accept": false, - "reason": "A vague documentation request and a spam placeholder are not the same underlying report." + "reason": "Documentation enhancement suggestion and placeholder spam are unrelated." }, { "left": "issue:44418", "right": "issue:44573", "accept": false, - "reason": "No concrete overlap beyond both being low-signal issues." + "reason": "Documentation enhancement suggestion and placeholder spam are unrelated." }, { "left": "issue:44572", "right": "issue:44573", "accept": false, - "reason": "Both are spam placeholders, but there is no evidence they refer to the same actual request." + "reason": "Both are '' issues, but there is no concrete shared bug/change to justify treating them as duplicates." }, { "left": "issue:44572", "right": "issue:44937", "accept": false, - "reason": "Spam placeholder versus unrelated title; no shared target or substance." + "reason": "Placeholder spam and a promotional-looking title do not describe the same underlying defect or change." }, { "left": "issue:44573", "right": "issue:44937", "accept": false, - "reason": "No matching problem description or code-path evidence." + "reason": "Placeholder spam and a promotional-looking title do not describe the same underlying defect or change." } ] }, @@ -14303,7 +14508,7 @@ "soft_pairs": [ "issue:45145|issue:45151" ], - "prepared_review_unit_hash": "b0f2278ebda48cfb39cce3644c42d5871fcb0958bfdb4225404c4ac67b84ef31", + "prepared_review_unit_hash": "b7cc7e33e64627c0dafc580a2ec2edef1e41dd4c17ce47cc53a41d9cf21e186f", "component_budget": { "node_count": 2, "item_count": 2, @@ -14325,24 +14530,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The pair looks very similar and may be duplicates, but the packet only shows identical titles, similar body sizes, and a soft-similarity signal. That is enough to choose a canonical issue, but not enough to strongly assert overlapping content or to accept the soft edge with high confidence.", - "confidence": 0.74, - "canonical_issue_reason": "#45145 is the better canonical issue because it is earlier and has an inbound reference, but the evidence is limited to title/metadata similarity rather than visible body overlap.", + "summary": "The two closed issues have identical titles and a soft-similarity signal, so they are plausibly related and may be duplicates, but the available packet does not expose body content to confirm that they describe the same issue.", + "confidence": 0.79, + "canonical_issue_reason": "#45145 is the better canonical issue because it has an inbound reference, giving it slightly more linkage context than #45151.", "canonical_pr_reason": null, - "best_issue_reason": "#45145 is the strongest issue representative on available metadata: it predates #45151 and already has an inbound reference.", + "best_issue_reason": "#45145 is the strongest representative of the issue cluster because it has the same title as #45151 and slightly more reference context.", "best_pr_reason": null, "soft_edge_verdicts": [ { "left": "issue:45145", "right": "issue:45151", "accept": false, - "reason": "Identical titles and similar sizes make them plausible duplicates, but the packet does not expose the bodies or any explicit overlap, so this is not strong enough to confidently merge them as the same issue." + "reason": "Likely related/possible duplicate based on identical titles and soft similarity, but the packet does not show body content or other concrete evidence to confirm they are the same underlying issue." } ] }, "evaluator_result": { "accept": false, - "feedback": "The canonical choice of #45145 is plausible because it is earlier and has an inbound reference, but the writeup overstates the evidence. The packet only supports identical titles, similar body sizes, and a soft-similarity signal; it does not show the issue bodies or prove \u201chighly overlapping content.\u201d The soft-edge accept is therefore too strong, and the 0.99 confidence is not justified by the packet." + "feedback": "The duplicate-issue inference is directionally plausible, but the analyst overstates the evidence. The packet only supports that the two closed issues have identical titles, similar body lengths, and a soft-similarity signal; it does not expose body content, so claims like \"strongly overlapping body/theme\" and \"same title/content\" are not fully grounded. Also, the soft pair was not deterministically accepted, so the verdict should be phrased more conservatively (e.g., \"likely related/possible duplicate\" rather than definitive duplicate). The confidence of 0.98 is too strong for the available evidence." }, "evaluator_used": true, "retried": true, @@ -14358,8 +14563,8 @@ "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-45295-2", "nodes": [ @@ -14407,7 +14612,7 @@ { "label": "pull_request", "component_index": 1, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -14462,211 +14667,211 @@ "pull_request:43823|pull_request:45489", "pull_request:43251|pull_request:43254" ], - "prepared_review_unit_hash": "3a09d96ebcbe511a783ca186610be5017dff65d573d6229288fc7aa791663ced", + "prepared_review_unit_hash": "e77ec4c3ee556115959682801a06929e14789ded98d08ec7fc9209aba713ea5e", "component_budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 28, - "serialized_chars": 41382, - "estimated_input_tokens": 10346, - "estimated_eval_tokens": 20948 + "serialized_chars": 41462, + "estimated_input_tokens": 10366, + "estimated_eval_tokens": 20988 }, "budget": { "node_count": 14, "item_count": 14, "soft_pair_count": 28, - "serialized_chars": 41382, - "estimated_input_tokens": 10346, - "estimated_eval_tokens": 20948 + "serialized_chars": 41462, + "estimated_input_tokens": 10366, + "estimated_eval_tokens": 20988 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "This set is mostly a heterogeneous mix of unrelated PRs (new-model additions, processor/tokenizer fixes, Gemma/Falcon/Mistral changes, docs/tests). The only clear duplicate pair is the two loss-utils PRs (43251 and 43254), which both implement the same fixed_cross_entropy kwargs support.", - "confidence": 0.95, + "summary": "Only one soft-edge pair is a true duplicate: 43251 and 43254 both patch `fixed_cross_entropy` in `src/transformers/loss/loss_utils.py` to forward supported `nn.functional.cross_entropy` kwargs. All other pairs are unrelated changes across different models, utilities, or tests; in particular, 43823 is a MobileLLM addition and should not be described as Gemma/Mistral-related.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43251 is the best canonical representative of the only true duplicate pair: it directly targets issue 43240 and cleanly updates fixed_cross_entropy to pass supported cross-entropy kwargs.", + "canonical_pr_reason": "PR 43251 is the clearest representative of the duplicate loss fix: it directly patches `fixed_cross_entropy`, is explicitly tied to issue 43240, and makes the same concrete kwargs pass-through change as 43254 with minimal scope.", "best_issue_reason": null, - "best_pr_reason": "PR 43251 is the strongest global representative because it has the explicit issue link, the clearest scoped fix, and is the more complete version of the duplicated loss-utils change; the rest of the items are not close duplicates of one another.", + "best_pr_reason": "PR 43251 is the best standalone representative in this cluster because it addresses the concrete cross-entropy kwargs bug directly, with an explicit issue link and the narrowest, most focused patch.", "soft_edge_verdicts": [ { "left": "pull_request:39895", "right": "pull_request:43424", "accept": false, - "reason": "Different changes: adding VideoPrism support vs an Executorch dynamic-shape test." + "reason": "Different features in unrelated areas: VideoPrism model addition vs executorch dynamic-shape export testing." }, { "left": "pull_request:42668", "right": "pull_request:43823", "accept": false, - "reason": "Unrelated scopes: processor-from-pretrained robustness vs adding MobileLLM model code." + "reason": "Different changes: processor-pretrained robustness work vs a new MobileLLM model addition." }, { "left": "pull_request:42668", "right": "pull_request:45078", "accept": false, - "reason": "Different code paths: processor loading robustness vs tokenizer auto-selection/error handling." + "reason": "Both involve auto/processor code, but they fix different behaviors: processor-from-pretrained robustness vs tokenizer conversion handling." }, { "left": "pull_request:43251", "right": "pull_request:43823", "accept": false, - "reason": "Loss-function kwargs fix is unrelated to adding a new model implementation." + "reason": "Cross-entropy kwargs forwarding is unrelated to adding the MobileLLM model." }, { "left": "pull_request:43251", "right": "pull_request:44827", "accept": false, - "reason": "Different fixes: cross-entropy kwargs vs Mistral4 test/model cleanup." + "reason": "Different code paths and purposes: loss utility kwargs vs Mistral4 test/model fixes." }, { "left": "pull_request:43424", "right": "pull_request:44827", "accept": false, - "reason": "Executorch export dynamic-shape test is unrelated to Mistral4 test fixes." + "reason": "Executorch export tests are unrelated to Mistral4 model/test changes." }, { "left": "pull_request:43424", "right": "pull_request:45435", "accept": false, - "reason": "Different subsystems: Executorch export testing vs Whisper tokenization bugfix." + "reason": "Different targets: executorch dynamic-shape export tests vs a Whisper tokenization bounds fix." }, { "left": "pull_request:43424", "right": "pull_request:43823", "accept": false, - "reason": "No shared underlying bug/change; test coverage vs new model addition." + "reason": "Executorch export testing and MobileLLM model addition are unrelated changes." }, { "left": "pull_request:43823", "right": "pull_request:45078", "accept": false, - "reason": "Adding MobileLLM and tokenizer auto error handling are not the same change." + "reason": "MobileLLM model addition and tokenizer conversion error handling are different changes with no shared bug." }, { "left": "pull_request:43823", "right": "pull_request:45454", "accept": false, - "reason": "MobileLLM model addition is unrelated to Gemma4 text-only training behavior." + "reason": "MobileLLM addition is unrelated to the Gemma4 text-only training fix; they touch different models and code paths." }, { "left": "pull_request:43823", "right": "pull_request:45435", "accept": false, - "reason": "New model addition vs Whisper unicode-tokenization fix; different bugs." + "reason": "MobileLLM model addition is unrelated to the Whisper special-token indexing fix." }, { "left": "pull_request:43823", "right": "pull_request:43995", "accept": false, - "reason": "MobileLLM addition and Falcon output-collection refactor are unrelated." + "reason": "New model addition vs Falcon output-collection refactor are different artifacts." }, { "left": "pull_request:43995", "right": "pull_request:45078", "accept": false, - "reason": "Falcon refactor does not fix the tokenizer auto-selection issue." + "reason": "Falcon output-interface refactoring is unrelated to tokenizer conversion logic." }, { "left": "pull_request:43995", "right": "pull_request:45435", "accept": false, - "reason": "Different code paths: Falcon forward signature/output refactor vs Whisper tokenization fix." + "reason": "Falcon refactoring and Whisper tokenization fixes address different code paths." }, { "left": "pull_request:43995", "right": "pull_request:45213", "accept": false, - "reason": "Falcon model refactor is unrelated to the model-creation skill PR." + "reason": "Falcon model refactoring and the model-creation skill PR are unrelated." }, { "left": "pull_request:43995", "right": "pull_request:44827", "accept": false, - "reason": "Falcon output refactor and Mistral4 test fixes are different changes." + "reason": "Falcon refactor is unrelated to the Mistral4 test/model fixes." }, { "left": "pull_request:44827", "right": "pull_request:45489", "accept": false, - "reason": "Mistral4 test fixes and Gemma3n/Gemma4 cache-sharing alignment are distinct model changes." + "reason": "Both are model PRs, but they fix different concrete behaviors in different model families and are not mergeable as one change." }, { "left": "pull_request:44827", "right": "pull_request:45170", "accept": false, - "reason": "Different fixes: Mistral4 test cleanup vs CLIP-like layernorm rename mapping." + "reason": "Mistral4 test/model changes are unrelated to the CLIP-like `layrnorm` renaming fix." }, { "left": "pull_request:43823", "right": "pull_request:44827", "accept": false, - "reason": "New MobileLLM model work and Mistral4 test fixes are unrelated." + "reason": "MobileLLM addition and Mistral4 fixes are unrelated changes." }, { "left": "pull_request:45078", "right": "pull_request:45454", "accept": false, - "reason": "Tokenizer conversion/error handling is unrelated to Gemma4 training text-only samples." + "reason": "Tokenizer conversion handling is unrelated to Gemma4 text-only training behavior." }, { "left": "pull_request:45078", "right": "pull_request:45435", "accept": false, - "reason": "Different tokenizer-related bugs: auto conversion logic vs Whisper unicode split bounds check." + "reason": "Different tokenizer issues: conversion-required handling vs special-token Unicode splitting bounds." }, { "left": "pull_request:45213", "right": "pull_request:45435", "accept": false, - "reason": "Model-creation skill scaffolding is unrelated to Whisper tokenization correctness." + "reason": "A model-creation skill PR is unrelated to a Whisper tokenizer bug fix." }, { "left": "pull_request:44827", "right": "pull_request:45213", "accept": false, - "reason": "Mistral4 test fixes and model-creation skill content are not the same change." + "reason": "Mistral4 fixes are unrelated to the model-creation skill scaffolding." }, { "left": "pull_request:43823", "right": "pull_request:45213", "accept": false, - "reason": "MobileLLM model addition and the model-creation skill PR are different artifacts entirely." + "reason": "MobileLLM model addition and the model-creation skill are different artifacts, not the same change." }, { "left": "pull_request:45078", "right": "pull_request:45213", "accept": false, - "reason": "Tokenizer auto behavior fix vs model-creation skill documentation/scaffolding." + "reason": "Tokenizer behavior changes are unrelated to the skill/documentation PR." }, { "left": "pull_request:45454", "right": "pull_request:45489", "accept": false, - "reason": "Gemma4 text-only training and Gemma3n/Gemma4 cache-sharing alignment are related only by model family, not the same bug." + "reason": "Gemma4 text-only training behavior and Gemma3n/Gemma4 cache-sharing alignment are separate concrete fixes." }, { "left": "pull_request:43823", "right": "pull_request:45489", "accept": false, - "reason": "MobileLLM model addition is unrelated to Gemma cache-sharing behavior." + "reason": "MobileLLM addition is unrelated to Gemma3n/Gemma4 cache-sharing changes." }, { "left": "pull_request:43251", "right": "pull_request:43254", "accept": true, - "reason": "Both PRs fix the same concrete bug in fixed_cross_entropy by forwarding supported kwargs to nn.functional.cross_entropy in the same file and issue context." + "reason": "Same file and same underlying bug: both forward supported kwargs to `nn.functional.cross_entropy` inside `fixed_cross_entropy`." } ] }, "evaluator_result": { - "accept": true, - "feedback": "The analysis is grounded in the packet and stays conservative. The only accepted pair, 43251/43254, is well supported by the shared file, shared issue target, and near-identical fixed_cross_entropy kwargs change. The rejected pairs are also plausibly unrelated and do not look mergeable into one bugfix PR." + "accept": false, + "feedback": "The 43251/43254 duplicate call is well grounded, but several soft-edge reasons are not conservative enough. In particular, PR 43823 is a MobileLLM addition, so descriptions like \"Gemma/Mistral-related code\" or \"Both touch Gemma-family models\" are factually off for pairs such as 43823|44827 and 43823|45454. Since the instructions require precise, conservative verdicts, I can\u2019t accept this output as-is." }, "evaluator_used": true, - "retried": false, + "retried": true, "accepted_nontrivial_soft_edge": true, "error_kind": null, "error_message": null @@ -14674,7 +14879,7 @@ { "label": "pull_request", "component_index": 2, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -14690,7 +14895,7 @@ "soft_pairs": [ "pull_request:40546|pull_request:42781" ], - "prepared_review_unit_hash": "6258f3819f6acfa2fdab78d7e7a3247eaeb585d87fb5e74413ff5b199f334567", + "prepared_review_unit_hash": "165162363feb23aa86a51b40e0ca400a1ea9cbb4de8adb5dce4eb9f8b182e2c9", "component_budget": { "node_count": 2, "item_count": 2, @@ -14712,18 +14917,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs add VibeVoice support, but they are not the same change: #40546 implements the core model integration, while #42781 extends it for realtime/audio-tokenizer work and touches additional generation paths. They overlap on shared VibeVoice plumbing, but they should not be merged as duplicates.", - "confidence": 0.87, + "summary": "The two pull requests are related VibeVoice work, but they are not duplicates: 40546 is the base VibeVoice implementation, while 42781 adds a realtime variant plus the acoustic tokenizer and generation changes. They share many files, but they target different concrete capabilities.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "#40546 is the better canonical PR because it is the earlier, broader core VibeVoice implementation with substantial review activity and a more complete base integration.", + "canonical_pr_reason": "40546 is the foundational VibeVoice implementation and the broader, more complete base change set; 42781 is a draft extension for realtime support that depends on and expands the core model work.", "best_issue_reason": null, - "best_pr_reason": "#40546 is the best single representative of the cluster since it covers the main VibeVoice model support; #42781 is a later, narrower realtime extension built on top of that work.", + "best_pr_reason": "40546 is the best canonical PR for the underlying VibeVoice addition because it represents the main initial implementation rather than a specialized follow-on feature.", "soft_edge_verdicts": [ { "left": "pull_request:40546", "right": "pull_request:42781", "accept": false, - "reason": "Shared VibeVoice files indicate related work, but #42781 is a separate realtime/audio-tokenizer extension rather than the same concrete fix as #40546. They overlap in subsystem, not in a single mergeable change." + "reason": "Reject: 40546 implements core VibeVoice support, while 42781 adds realtime-specific functionality and an acoustic tokenizer. Shared filenames alone do not make them the same underlying change." } ] }, @@ -14737,7 +14942,7 @@ { "label": "pull_request", "component_index": 3, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -14753,12 +14958,12 @@ "soft_pairs": [ "pull_request:40755|pull_request:41224" ], - "prepared_review_unit_hash": "ca08cfafffb81953b0a8dd69f72b038e36d33eb0ee0c1977f4bceb0c4fd8a2ea", + "prepared_review_unit_hash": "4d1fe8af026fa7dcb39a3c5cd04bddc45fe43e4ee16b2d67322e0d37441df2e2", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -14766,7 +14971,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4674, + "serialized_chars": 4675, "estimated_input_tokens": 1169, "estimated_eval_tokens": 2594 }, @@ -14775,18 +14980,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "The two PRs are unrelated: one adds TimesFM covariate forecasting, the other adds DINOv3 image classification support. No duplicate/canonical pair is evident.", - "confidence": 0.99, + "summary": "These two pull requests are unrelated: one adds covariate-based forecasting support to TimesFM, while the other adds DINOv3 image classification support. The only overlap is generic model/docs/tests work, which is not enough to treat them as duplicates.", + "confidence": 0.98, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs address different models and different code paths, so there is no single underlying change to collapse them into.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "Neither PR is a duplicate of the other; both are independent feature additions in separate subsystems, so neither is a better canonical representative for this cluster.", "soft_edge_verdicts": [ { "left": "pull_request:40755", "right": "pull_request:41224", "accept": false, - "reason": "Different models and code paths (TimesFM time-series covariates vs DINOv3 image classification); they do not fix the same underlying change and would not be merged as one PR." + "reason": "Different models and different functionality: TimesFM covariate forecasting vs. DINOv3 image classification support. Shared docs/tests patterns do not indicate the same underlying bug or change." } ] }, @@ -14800,13 +15005,13 @@ { "label": "pull_request", "component_index": 4, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-41356-2", "nodes": [ @@ -14866,7 +15071,7 @@ { "label": "pull_request", "component_index": 5, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -14882,7 +15087,7 @@ "soft_pairs": [ "pull_request:41797|pull_request:45186" ], - "prepared_review_unit_hash": "73cb19002c31055ec111d538563a22c523e9281928f2d85f19cf165b9e871102", + "prepared_review_unit_hash": "6d04fc6904485368e707c98b27ea5ee42ca9f1d07b519dcfb9591d7a77958124", "component_budget": { "node_count": 2, "item_count": 2, @@ -14904,7 +15109,7 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate model-addition PRs, not duplicates: #41797 adds DeepSeekOCR and #45186 adds Isaac. They share some Transformers integration files, but they target different models and code paths.", + "summary": "These are two separate model-addition PRs with overlapping boilerplate and registry/docs files, but they implement different models and do not look mergeable as one change.", "confidence": 0.98, "canonical_issue_reason": null, "canonical_pr_reason": null, @@ -14915,7 +15120,7 @@ "left": "pull_request:41797", "right": "pull_request:45186", "accept": false, - "reason": "Different underlying changes: DeepSeekOCR vs Isaac. Shared docs/auto-registration files are expected for new model integrations and do not indicate the same fix or a PR that could be merged as one." + "reason": "Different underlying changes: one adds DeepSeek OCR, the other adds Isaac. Shared auto/docs plumbing is expected for new-model PRs, not evidence of the same fix." } ] }, @@ -14929,7 +15134,7 @@ { "label": "pull_request", "component_index": 6, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -14945,12 +15150,12 @@ "soft_pairs": [ "pull_request:42166|pull_request:45426" ], - "prepared_review_unit_hash": "c44458b8a8a9be4fe66685cd762a1528157fa896c200c396e7af829c061a9514", + "prepared_review_unit_hash": "94e4688c750ab2029f402f9934224319a87ddbf6b98c0638d34a504db00e35b1", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5407, + "serialized_chars": 5408, "estimated_input_tokens": 1352, "estimated_eval_tokens": 2960 }, @@ -14958,7 +15163,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5407, + "serialized_chars": 5408, "estimated_input_tokens": 1352, "estimated_eval_tokens": 2960 }, @@ -14967,7 +15172,7 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two unrelated model-addition PRs were clustered because they both touch Transformer auto-registration files, but they implement different models (InternVLFlash vs AXK1) and do not appear to be the same underlying change.", + "summary": "The cluster contains two unrelated feature PRs: one adds InternVLFlash, the other adds AXK1. The overlap is limited to shared auto-registration plumbing, which is not enough to treat them as the same change.", "confidence": 0.97, "canonical_issue_reason": null, "canonical_pr_reason": null, @@ -14978,7 +15183,7 @@ "left": "pull_request:42166", "right": "pull_request:45426", "accept": false, - "reason": "Reject: the PRs add different model families with different codepaths and file sets. Shared edits to __init__/auto registries are boilerplate for model integration, not evidence of the same fix or a mergeable combined PR." + "reason": "Different model integrations with different filenames, configs, and tests; shared auto-registry edits are generic boilerplate, not the same bug or change." } ] }, @@ -14992,7 +15197,7 @@ { "label": "pull_request", "component_index": 7, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15015,12 +15220,12 @@ "pull_request:43291|pull_request:44815", "pull_request:43665|pull_request:44815" ], - "prepared_review_unit_hash": "af869772927016408959bba30c7230e0e7560e2048c770e9a1f3ae5f753d376b", + "prepared_review_unit_hash": "60886efd49f4cdce90a921d77f190fb8abb077efb094f314bb3ddf20601769ca", "component_budget": { "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 10167, + "serialized_chars": 10168, "estimated_input_tokens": 2542, "estimated_eval_tokens": 5340 }, @@ -15028,7 +15233,7 @@ "node_count": 4, "item_count": 4, "soft_pair_count": 6, - "serialized_chars": 10167, + "serialized_chars": 10168, "estimated_input_tokens": 2542, "estimated_eval_tokens": 5340 }, @@ -15037,48 +15242,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are not true duplicates overall. PR #42227 is a concrete Whisper pipeline bug fix, while #43291 is a separate Whisper tokenizer/test change, #43665 is test-only noise, and #44815 is an unrelated FP8/dequantization fix.", - "confidence": 0.92, + "summary": "These PRs are not true duplicates; they span unrelated Whisper pipeline/tokenizer work, model test tweaks, and an FP8 dequant/loading fix. The only plausible overlap is broad \u201cWhisper\u201d terminology, but the concrete code paths differ enough to reject merges.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "#42227 is the most concrete and self-contained fix: it has an explicit issue target, a focused code-path change in ASR Whisper language return handling, and supporting tests.", + "canonical_pr_reason": "No single canonical PR for duplicate-triage: the cluster is heterogeneous, but PR #42227 is the strongest standalone representative because it has a clear bug fix, explicit issue linkage, and a focused code-path change.", "best_issue_reason": null, - "best_pr_reason": "#42227 is the best representative PR for the cluster because it addresses a real underlying bug rather than test churn or unrelated model-loading/quantization work.", + "best_pr_reason": "PR #42227 is the best representative artifact overall: it addresses a concrete Whisper ASR bug, includes a targeted implementation change, and is tied to an explicit tracking issue.", "soft_edge_verdicts": [ { "left": "pull_request:42227", "right": "pull_request:43291", "accept": false, - "reason": "Both touch Whisper, but they fix different paths: #42227 forwards return_language through the ASR pipeline, while #43291 changes Whisper tokenizer decode/timestamp behavior and tests. Not the same underlying bug/change." + "reason": "Both mention Whisper, but #42227 fixes ASR pipeline return_language forwarding while #43291 changes Whisper tokenizer/tests; different bugs and not plausibly one PR." }, { "left": "pull_request:42227", "right": "pull_request:43665", "accept": false, - "reason": "#42227 is a Whisper runtime fix; #43665 only edits CLIP/ViT tests with no shared code-path or bug. Not duplicates." + "reason": "Unrelated model test edits (CLIP/ViT) versus a Whisper pipeline fix; no shared concrete code-path problem." }, { "left": "pull_request:42227", "right": "pull_request:44815", "accept": false, - "reason": "Whisper ASR language-return handling is unrelated to model-loading and FP8 dequantization changes. Different subsystems and different bugs." + "reason": "Whisper ASR language-return fix versus FP8 dequant/model-loading changes; completely different subsystems and failure modes." }, { "left": "pull_request:43291", "right": "pull_request:43665", "accept": false, - "reason": "#43291 is Whisper tokenizer logic; #43665 is unrelated test-file edits in CLIP/ViT. No common underlying change." + "reason": "Whisper tokenizer/test cleanup and CLIP/ViT test edits are unrelated; same broad category of tests does not imply the same bug." }, { "left": "pull_request:43291", "right": "pull_request:44815", "accept": false, - "reason": "Whisper decode/timestamp behavior and FP8 dequantization are unrelated code paths and bugs." + "reason": "Tokenizer/test changes for Whisper are unrelated to FP8 quantization/dequantization and loading logic." }, { "left": "pull_request:43665", "right": "pull_request:44815", "accept": false, - "reason": "The CLIP/ViT test-only edits in #43665 do not match the FP8/model-loading dequantization fix in #44815." + "reason": "CLIP/ViT test modifications do not overlap with FP8 dequant/model-loading changes in any concrete fixable way." } ] }, @@ -15092,7 +15297,7 @@ { "label": "pull_request", "component_index": 8, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15108,7 +15313,7 @@ "soft_pairs": [ "pull_request:42975|pull_request:42976" ], - "prepared_review_unit_hash": "60617b591df0c371a7ec5114a7742ef002d411e8dd0d437e1219354434e3ae26", + "prepared_review_unit_hash": "d3213d288c938d92f5c2515d30871f5451ff2f2c63e2b501c780e76ca8d8012d", "component_budget": { "node_count": 2, "item_count": 2, @@ -15130,18 +15335,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two maintenance PRs that both upgrade GitHub Actions, but they update different action families and address different compatibility/version bumps, so they should not be deduplicated.", - "confidence": 0.95, + "summary": "These are two separate workflow dependency upgrade PRs: one for GitHub Actions runtime compatibility, one for docker/build-push-action version bumps. They are related in theme but not the same change.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #42975 is the better cluster representative because it is the broader workflow-maintenance change and explicitly targets Node 24 compatibility across many actions.", + "canonical_pr_reason": "PR 42975 is the broader and more self-contained workflow action runtime upgrade, spanning many files and explicitly targeting Node 24 compatibility; PR 42976 is a narrower docker action version refresh.", "best_issue_reason": null, - "best_pr_reason": "PR #42975 is the best overall representative of the cluster topic: it is the broader and more explicit GitHub Actions compatibility upgrade, while #42976 is a narrower Docker action version bump.", + "best_pr_reason": "PR 42975 is the best standalone representative of the cluster because it has the clearer release goal and wider workflow impact, while still being a single coherent upgrade.", "soft_edge_verdicts": [ { "left": "pull_request:42975", "right": "pull_request:42976", "accept": false, - "reason": "Both are workflow dependency upgrades, but they update different actions (checkout/cache/upload-artifact vs docker/build-push-action) and do not fix the same concrete change." + "reason": "Reject: they update different actions for different purposes (Node 24-compatible checkout/cache/upload-artifact vs docker/build-push-action v6). Similar maintenance theme, but not the same underlying bug or mergeable into one PR without conflating separate dependency bumps." } ] }, @@ -15155,7 +15360,7 @@ { "label": "pull_request", "component_index": 9, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15178,7 +15383,7 @@ "pull_request:43448|pull_request:45490", "pull_request:43451|pull_request:45490" ], - "prepared_review_unit_hash": "a9cb13311be3ecb2735116917f14c051e4d842f81e9f9df85f0ae86d04a68992", + "prepared_review_unit_hash": "2adba3260fe6ddb17db1c74a72dca8dfb0816ba0d9d68d54ef8c0e6bd06cd2ad", "component_budget": { "node_count": 4, "item_count": 4, @@ -15200,48 +15405,48 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These PRs are all separate model-addition changes (ViT NEPA, Molmo, Molmo2, CTSM). They only overlap on shared framework files like docs/toctree and auto mappings, so they are not duplicates of the same underlying bug or change.", - "confidence": 0.98, + "summary": "These PRs are all model-addition work with overlapping framework scaffolding, but they target different models and different code paths: ViT NEPA, Molmo, Molmo2, and CTSM. Shared changes in docs/auto mappings are expected and do not indicate duplication.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: each one introduces a different model integration with its own files and codepath.", + "canonical_pr_reason": "No canonical PR here: the four pull requests are for distinct model implementations, not the same bug or change.", "best_issue_reason": null, - "best_pr_reason": "PR 43451 (Molmo2) is the strongest representative of this cluster because it has the broadest implementation surface and highest review activity, but it is still a distinct change from the others.", + "best_pr_reason": "PR 43451 (Molmo2) is the broadest and most developed of the set, with the largest surface area and most review activity, so it is the best representative PR for the cluster if one must be chosen.", "soft_edge_verdicts": [ { "left": "pull_request:42978", "right": "pull_request:43448", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo. Shared docs/auto-registration files are superficial overlap, not the same fix." + "reason": "Different model additions: ViT NEPA vs Molmo. Similar auto/doc scaffolding, but not the same concrete feature or fix." }, { "left": "pull_request:42978", "right": "pull_request:43451", "accept": false, - "reason": "Different model additions: ViT NEPA vs Molmo2, with separate model files and implementations." + "reason": "Different model additions: ViT NEPA vs Molmo2. Shared Transformers boilerplate does not make them duplicates." }, { "left": "pull_request:42978", "right": "pull_request:45490", "accept": false, - "reason": "Different model additions: ViT NEPA vs CTSM. Same framework plumbing, but no shared codepath problem." + "reason": "Different model additions: ViT NEPA vs CTSM. They touch the same framework areas but implement unrelated models." }, { "left": "pull_request:43448", "right": "pull_request:43451", "accept": false, - "reason": "Molmo and Molmo2 are distinct model ports with different architectures and files; not mergeable as one bug/change." + "reason": "Molmo and Molmo2 are separate model families with separate implementations, configs, and docs; not one mergeable change." }, { "left": "pull_request:43448", "right": "pull_request:45490", "accept": false, - "reason": "Molmo vs CTSM are unrelated model integrations; overlap is only in common registry/docs updates." + "reason": "Molmo vs CTSM are unrelated model additions. Overlap is only in shared registration/documentation plumbing." }, { "left": "pull_request:43451", "right": "pull_request:45490", "accept": false, - "reason": "Molmo2 and CTSM are separate model implementations, not the same underlying change." + "reason": "Molmo2 and CTSM are distinct model implementations, so they are not the same underlying change." } ] }, @@ -15255,7 +15460,7 @@ { "label": "pull_request", "component_index": 10, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15274,52 +15479,52 @@ "pull_request:43488|pull_request:45476", "pull_request:43532|pull_request:45476" ], - "prepared_review_unit_hash": "598280c219cb888b3aa248bab5b45f510020632848474e2671034e1feaf3725f", + "prepared_review_unit_hash": "06fe7f1e2581430c87d02dcbb4fd86cffcbd1606e113c7cf246a675904577eea", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 7469, - "estimated_input_tokens": 1868, - "estimated_eval_tokens": 3992 + "serialized_chars": 7468, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 7469, - "estimated_input_tokens": 1868, - "estimated_eval_tokens": 3992 + "serialized_chars": 7468, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three unrelated '[don't merge]' PRs that only overlap on incidental test-file edits; the shared vit test path does not indicate the same underlying fix or change.", - "confidence": 0.96, + "summary": "All three PRs are unrelated test/synthetic changes: one tweaks packaging/versioning, one rewires comment-driven CI workflows, and one adds a caller workflow plus unrelated tokenization/test edits. They only overlap superficially on a shared test file, not on the same bug or change.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR stands out as a duplicate representative; 45476 is the most self-contained/substantive of the three, but it is still unrelated to the others.", + "canonical_pr_reason": "No single canonical PR stands out because the three PRs target different code paths and are not duplicates of one another; if forced, 43532 is the most substantial CI-related change, but it is still not representative of the others.", "best_issue_reason": null, - "best_pr_reason": "45476 is the best representative only in the sense that it is the most coherent standalone PR, but it does not match the other PRs as the same change.", + "best_pr_reason": "43532 is the most coherent standalone change set, centered on CI/comment workflow behavior, but it should not be treated as a cluster canonical because it does not match the other PRs' scope.", "soft_edge_verdicts": [ { "left": "pull_request:43488", "right": "pull_request:43532", "accept": false, - "reason": "43488 is a setup/dependency/version check PR, while 43532 is a workflow/testing-utils PR; the overlap in vit test touchpoints is incidental." + "reason": "Shared vit test filename is incidental; 43488 is setup/version/dependency churn, while 43532 is workflow and notification plumbing. Different code paths and different intent." }, { "left": "pull_request:43488", "right": "pull_request:45476", "accept": false, - "reason": "43488 changes packaging/dependency metadata, whereas 45476 adds a CI workflow and unrelated CLIP/Chinese-CLIP test edits; not the same code-path fix." + "reason": "43488 changes packaging/version/deps, while 45476 adds a PR CI workflow and touches CLIP/chinese-CLIP test scaffolding. No same underlying bug or mergeable combined fix." }, { "left": "pull_request:43532", "right": "pull_request:45476", "accept": false, - "reason": "Both touch tests/CI, but they modify different workflows and different model areas; the shared vit test file is not enough to treat them as the same change." + "reason": "Both mention CI/test infrastructure and a vit test file, but 43532 is about self-comment/self-scheduled workflow logic, whereas 45476 adds a new PR-CI caller plus unrelated tokenizer/test edits. Not the same concrete change." } ] }, @@ -15333,7 +15538,7 @@ { "label": "pull_request", "component_index": 11, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15349,7 +15554,7 @@ "soft_pairs": [ "pull_request:43612|pull_request:43613" ], - "prepared_review_unit_hash": "4a48926e4e2e05016c6e89b018607273646819f9d25b7f41bd77599a14767b8e", + "prepared_review_unit_hash": "6bdb893c182097484f72bc42915f4cff7c3528777ec9df7dad238d7b1c891b13", "component_budget": { "node_count": 2, "item_count": 2, @@ -15371,18 +15576,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related but distinct feature PRs: one adds a Promptable Concept Segmentation pipeline, the other adds a Promptable Visual Segmentation pipeline. They share some pipeline/doc scaffolding, but they target different models and code paths, so they should not be clustered as duplicates.", - "confidence": 0.97, + "summary": "These are related but distinct feature PRs: one adds Promptable Concept Segmentation, the other Promptable Visual Segmentation. They share docs/pipeline registration scaffolding, but they do not fix the same code-path or implement the same change.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": null, + "canonical_pr_reason": "PR #43613 is the better representative because it is the broader/more integrated segmentation pipeline addition, touching more model backends and pipeline plumbing.", "best_issue_reason": null, - "best_pr_reason": null, + "best_pr_reason": "PR #43613 is the stronger standalone PR for this cluster because it has wider implementation coverage and is the more complete segmentation feature addition.", "soft_edge_verdicts": [ { "left": "pull_request:43612", "right": "pull_request:43613", "accept": false, - "reason": "Different features and different concrete integrations: 43612 centers on concept segmentation with SAM3, while 43613 targets visual segmentation and touches SAM/SAM2/EdgeTAM/SAM3 tracker. Shared documentation and pipeline registration are broad scaffolding, not the same underlying change." + "reason": "Different features: concept segmentation vs visual segmentation. Similar docs and shared pipeline scaffolding are not enough to treat them as the same underlying change." } ] }, @@ -15396,7 +15601,7 @@ { "label": "pull_request", "component_index": 12, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15412,12 +15617,12 @@ "soft_pairs": [ "pull_request:43651|pull_request:43663" ], - "prepared_review_unit_hash": "0085a96816bcdd342f6b2ba4f2084924e960065c7e6d22dc441574e0b5ca7b57", + "prepared_review_unit_hash": "d834eba410e4169fbe9ffb39bf75ac9b94756881af61c350df06e0363c190f6c", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4292, + "serialized_chars": 4291, "estimated_input_tokens": 1073, "estimated_eval_tokens": 2402 }, @@ -15425,7 +15630,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4292, + "serialized_chars": 4291, "estimated_input_tokens": 1073, "estimated_eval_tokens": 2402 }, @@ -15434,18 +15639,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both items touch `src/transformers/trainer.py`, but they address different trainer extension points: one changes gradient-accumulation loss scaling, the other refactors signature-based column filtering. They are not duplicate PRs.", + "summary": "Two unrelated Trainer extensibility PRs were clustered together only because both touch `src/transformers/trainer.py`: one adds `_loss_is_scaled_for_ga` to control gradient-accumulation loss scaling, the other factors out signature-column detection into `_get_signature_columns` for column filtering. They address different code paths and are not duplicates.", "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR here; the two PRs target different code paths and different customization hooks, so neither can represent the other as a duplicate.", + "canonical_pr_reason": "PR 43651 is the better representative of the cluster because it makes a concrete behavior change in the training step and is the more substantial, user-facing Trainer fix.", "best_issue_reason": null, - "best_pr_reason": "Neither PR is a good global duplicate representative because they solve unrelated trainer concerns. If one must be picked, 43651 is the more concrete bug/behavior change, but it still does not subsume 43663.", + "best_pr_reason": "PR 43651 is the strongest single PR to keep as the cluster representative; it has the clearest end-user impact and a more self-contained behavioral change than the refactoring-only helper added in 43663.", "soft_edge_verdicts": [ { "left": "pull_request:43651", "right": "pull_request:43663", "accept": false, - "reason": "Different changes in the same file: 43651 alters loss normalization during training, while 43663 adds an overridable signature-column helper for dataset column filtering. They do not fix the same underlying bug and would not plausibly merge into one PR." + "reason": "Reject: these modify different Trainer mechanisms. 43651 changes gradient accumulation loss scaling in `training_step`, while 43663 introduces a helper for model signature/column filtering. Shared file alone is not enough to treat them as the same change." } ] }, @@ -15459,7 +15664,7 @@ { "label": "pull_request", "component_index": 13, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15477,52 +15682,52 @@ "pull_request:43656|pull_request:43842", "pull_request:43836|pull_request:43842" ], - "prepared_review_unit_hash": "1ea3ba19090d629f59021f2199adc50692738eaa6c65ef7f4687326c0123a4a2", + "prepared_review_unit_hash": "f8042d67d07e88b89a8f2b2110cde9bd6f0ec3c6023b8d75a2bd591059b7bca8", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 5992, - "estimated_input_tokens": 1498, - "estimated_eval_tokens": 3252 + "serialized_chars": 5993, + "estimated_input_tokens": 1499, + "estimated_eval_tokens": 3254 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 2, - "serialized_chars": 5992, - "estimated_input_tokens": 1498, - "estimated_eval_tokens": 3252 + "serialized_chars": 5993, + "estimated_input_tokens": 1499, + "estimated_eval_tokens": 3254 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "All three PRs are trying to address TypeAdapter-related NameErrors in `src/transformers/cli/serve.py`. The two PRs targeting issue #43824 are the closest duplicates; the earlier #43656 is related but broader and less directly aligned with the TypeAdapter import/annotation failure.", + "summary": "All three pull requests target the same TypeAdapter import/type-annotation NameError in `src/transformers/cli/serve.py`, but they vary in specificity. The most targeted fix is PR #43842; the others are broader or earlier variants of the same underlying problem.", "confidence": 0.88, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #43842 is the most focused and direct fix: it imports `TypeAdapter` only under `TYPE_CHECKING` and stringizes the runtime annotation, which cleanly resolves the NameError when pydantic is absent. It also explicitly targets the reported issue (#43824).", + "canonical_pr_reason": "PR #43842 is the cleanest and most focused fix: it guards the `TypeAdapter` import under `TYPE_CHECKING` and quotes the runtime annotation, directly addressing the import-time NameError with minimal unrelated change.", "best_issue_reason": null, - "best_pr_reason": "PR #43842 is the best representative because it addresses the concrete failure mode with the smallest, clearest code change and the most specific issue linkage.", + "best_pr_reason": "PR #43842 is the best representative because it precisely fixes the concrete failure mode (`TypeAdapter` NameError when pydantic is absent) with the smallest, most self-contained diff.", "soft_edge_verdicts": [ { "left": "pull_request:43656", "right": "pull_request:43842", "accept": true, - "reason": "Both PRs are fixing the same CLI TypeAdapter annotation NameError in the same file; #43656 and #43842 are plausibly mergeable into one solution, even though #43656 is broader and touches additional annotation cleanup." + "reason": "Both address the same `TypeAdapter` NameError in the CLI serve path and appear to be alternative fixes for the same import-time annotation problem in the same file." }, { "left": "pull_request:43836", "right": "pull_request:43842", "accept": true, - "reason": "These are near-duplicate fixes for the same TypeAdapter NameError in `serve.py`; #43836\u2019s string-literal workaround and #43842\u2019s TYPE_CHECKING import/stringized annotation are alternative implementations of the same bug fix." + "reason": "Both are fixes for the same `TypeAdapter` NameError in `serve.py`; despite #43836 being broader, the core change is the same concrete annotation/import issue and could plausibly be merged into one PR." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary is grounded in the packet: 43836 and 43842 both target issue #43824 in the same file, and 43656 is clearly related but broader. The soft-edge acceptances are conservative enough given the shared filename and TypeAdapter/annotation focus, and the mergeability claim is plausible rather than overstated." + "feedback": "The summary is grounded in the packet: all three PRs are in `src/transformers/cli/serve.py` and each is clearly aimed at the same TypeAdapter-related import/type-annotation NameError, with PR #43842 being the most focused fix. The pairwise accept calls are also conservative enough given the shared file and bug pattern." }, "evaluator_used": true, "retried": false, @@ -15533,7 +15738,7 @@ { "label": "pull_request", "component_index": 14, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15552,7 +15757,7 @@ "pull_request:43757|pull_request:45500", "pull_request:43757|pull_request:45506" ], - "prepared_review_unit_hash": "01c33bc963e7f0ba36ed7a427db39f572861f21fabed6710902203b682053d7b", + "prepared_review_unit_hash": "4447f433e0b642137eee441f92a2d823a184fb435e4ee91eb8c6146e666e83ef", "component_budget": { "node_count": 3, "item_count": 3, @@ -15574,30 +15779,30 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Two near-identical GPT-OSS GGUF support PRs are present, plus an earlier partial workaround. The latest PR (45506) is the strongest representative; the earlier PR (43757) is a limited fallback only and should not be merged as a duplicate of the full-support PRs.", + "summary": "Two PRs are effectively the same GPT-OSS GGUF-loading fix, with 45506 the later, non-draft version; 43757 is only an earlier partial fallback and not the same full change.", "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45506 is the latest and most complete implementation for GPT-OSS GGUF loading, explicitly superseding 43757 and covering docs, integration code, and tests.", + "canonical_pr_reason": "PR #45506 is the clearest canonical PR: it is the later finalized version of the GPT-OSS GGUF-loading work, includes the full support across loader/integration/tests/docs, and explicitly supersedes #43757.", "best_issue_reason": null, - "best_pr_reason": "PR 45506 is the best cluster representative because it fully addresses the underlying GPT-OSS GGUF loading task and is a cleaner canonical duplicate target than the narrower fallback-only PR.", + "best_pr_reason": "PR #45506 is the best representative of the cluster because it is the complete, non-draft implementation of GPT-OSS GGUF support and directly subsumes the earlier draft PR.", "soft_edge_verdicts": [ { "left": "pull_request:45500", "right": "pull_request:45506", "accept": true, - "reason": "Same underlying GPT-OSS GGUF loading feature, same issue target, and overlapping files/diff; 45506 is essentially the later, fuller version of 45500." + "reason": "Same concrete GPT-OSS GGUF loading change set: same target issue, same files, and the later PR is just the finalized version of the earlier draft." }, { "left": "pull_request:43757", "right": "pull_request:45500", "accept": false, - "reason": "43757 only adds a temporary gpt-neox fallback for one architecture path, while 45500 is a broader GGUF loading feature; they do not appear to be the same concrete change." + "reason": "43757 only adds a fallback for the gpt-oss architecture in one loader path, while 45500 implements broader full GGUF loading support across loader, integration, docs, and tests." }, { "left": "pull_request:43757", "right": "pull_request:45506", "accept": false, - "reason": "43757 is a narrow workaround, whereas 45506 implements full GGUF support with docs, integration, and tests; same issue target alone is not enough to treat them as duplicates." + "reason": "Same as above: 43757 is a partial workaround, not the same underlying full GGUF-loading fix as 45506." } ] }, @@ -15611,13 +15816,13 @@ { "label": "pull_request", "component_index": 15, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", "reason": null, - "source": "llm", - "cache_hit": false, + "source": "cache", + "cache_hit": true, "model": "gpt-5.4-mini?service_tier=flex", "cluster_id": "cluster-43779-2", "nodes": [ @@ -15677,7 +15882,7 @@ { "label": "pull_request", "component_index": 16, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -15695,7 +15900,7 @@ "pull_request:43838|pull_request:45283", "pull_request:43973|pull_request:45283" ], - "prepared_review_unit_hash": "be2c93a177e9dfc00b51bc0ba9b57f20502e1545c380122387f4dc914c3770ba", + "prepared_review_unit_hash": "7105e8e9fca39b9892a98470abd8d783453f3812684244a1327f48556aa95486", "component_budget": { "node_count": 3, "item_count": 3, @@ -15717,24 +15922,24 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate pull requests for different model-support tasks: Qwen3-ASR integration, Lfm2Audio integration, and Qwen3.5 GGUF loading support. They share only broad audio/model-integration similarity, not the same underlying change.", - "confidence": 0.97, + "summary": "These are three unrelated model-support PRs (Qwen3-ASR, Lfm2.5 Audio, and Qwen3.5 GGUF loading). They share a broad audio/LLM-adjacent ecosystem but not the same underlying bug or code-path change, so the soft links should be rejected.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 43838 is the most substantial and central model-support PR in the cluster: it has the largest scope, explicit issue linkage, and core implementation files for a new Qwen3-ASR integration. The others are unrelated model additions/loading support.", + "canonical_pr_reason": "PR 43838 is the strongest representative: it adds a full Qwen3-ASR integration end-to-end (model/config/processor/autos/docs/tests) and explicitly targets issue 43837.", "best_issue_reason": null, - "best_pr_reason": "PR 43838 is the best representative PR because it is the main end-to-end feature implementation rather than a narrower loader-mapping change, and it is clearly tied to a tracked issue.", + "best_pr_reason": "PR 43838 is the best single representative of the cluster because it is the most complete and clearly scoped model-support implementation, with substantial code, docs, and test coverage.", "soft_edge_verdicts": [ { "left": "pull_request:43838", "right": "pull_request:45283", "accept": false, - "reason": "Qwen3-ASR support vs Qwen3.5 GGUF loading support are different model families and different code paths; they are not the same change and would not plausibly merge into one PR." + "reason": "Different changes: Qwen3-ASR model integration vs Qwen3.5 GGUF loading support. They do not fix the same concrete code-path problem." }, { "left": "pull_request:43973", "right": "pull_request:45283", "accept": false, - "reason": "Lfm2Audio model integration and Qwen3.5 GGUF loading support address unrelated models/features, with no shared concrete bug or implementation target." + "reason": "Different model additions and different mechanisms: Lfm2.5 Audio support vs Qwen3.5 GGUF mapping/loading. Same broad area only, not the same change." } ] }, @@ -15748,7 +15953,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 2, "status": "reviewed", @@ -15766,7 +15971,6 @@ "pull_request:44017", "pull_request:44018", "pull_request:44019", - "pull_request:44024", "pull_request:44044", "pull_request:44056", "pull_request:44059", @@ -15796,8 +16000,8 @@ "pull_request:44066|pull_request:44072", "pull_request:44013|pull_request:44044", "pull_request:44066|pull_request:44086", - "pull_request:44066|pull_request:44071", "pull_request:44018|pull_request:44068", + "pull_request:44066|pull_request:44071", "pull_request:44066|pull_request:44068", "pull_request:44018|pull_request:44066", "pull_request:44019|pull_request:44722", @@ -15807,15 +16011,15 @@ "pull_request:44019|pull_request:44071", "pull_request:44068|pull_request:44116", "pull_request:44068|pull_request:44722", - "pull_request:44071|pull_request:44722", "pull_request:44018|pull_request:44116", + "pull_request:44071|pull_request:44722", "pull_request:44018|pull_request:44086", "pull_request:44018|pull_request:44085", - "pull_request:44071|pull_request:44085", "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", "pull_request:44059|pull_request:44086", - "pull_request:44059|pull_request:44722", "pull_request:44085|pull_request:44086", + "pull_request:44059|pull_request:44722", "pull_request:44098|pull_request:44101", "pull_request:44018|pull_request:44071", "pull_request:44017|pull_request:44722", @@ -15824,22 +16028,23 @@ "pull_request:44068|pull_request:44071", "pull_request:44071|pull_request:44086", "pull_request:44059|pull_request:44071", - "pull_request:43997|pull_request:44073", "pull_request:44017|pull_request:44071", + "pull_request:43997|pull_request:44073", "pull_request:44059|pull_request:44068", "pull_request:44073|pull_request:44101", "pull_request:44018|pull_request:44059", "pull_request:44056|pull_request:44101", "pull_request:44073|pull_request:44161", "pull_request:44101|pull_request:44161", - "pull_request:43997|pull_request:44017", + "pull_request:44017|pull_request:44018", "pull_request:43997|pull_request:44056", "pull_request:44071|pull_request:44076", "pull_request:44074|pull_request:44101", - "pull_request:44056|pull_request:44161", "pull_request:43997|pull_request:44018", + "pull_request: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:44010|pull_request:44073", @@ -15858,373 +16063,336 @@ "pull_request:43996|pull_request:44007", "pull_request:43996|pull_request:44129", "pull_request:44044|pull_request:44129", - "pull_request:44044|pull_request:44154", - "pull_request:44024|pull_request:44076" + "pull_request:44044|pull_request:44154" ], - "prepared_review_unit_hash": "b3da15004de96406115cef73bf0790a7bfa7e5f58e62d3797df419156456a05d", + "prepared_review_unit_hash": "8f094394d71060f2893abafe4167f03e441f178cf5ad94cf3565d8d91bcd2805", "component_budget": { "node_count": 28, "item_count": 28, - "soft_pair_count": 77, - "serialized_chars": 84507, - "estimated_input_tokens": 21127, - "estimated_eval_tokens": 42510 + "soft_pair_count": 78, + "serialized_chars": 84896, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 }, "budget": { - "node_count": 28, - "item_count": 28, + "node_count": 27, + "item_count": 27, "soft_pair_count": 72, - "serialized_chars": 76939, - "estimated_input_tokens": 19235, - "estimated_eval_tokens": 38726 + "serialized_chars": 75094, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "This cluster is mostly a collection of model-specific output-tracing refactors, not one shared duplicate change. The only strong duplicate sets are within GPT-J/CodeGen and GPT-Neo; the rest touch different model code paths and should stay separate.", - "confidence": 0.86, - "canonical_issue_reason": null, - "canonical_pr_reason": "PR 44722 is the best representative for the overlapping GPT-J/CodeGen output-tracing work: it is the latest, broadest version in that subcluster and explicitly targets the umbrella tracing refactor.", - "best_issue_reason": null, - "best_pr_reason": "PR 44722 is the strongest global PR candidate because it most cleanly subsumes the GPT-J/CodeGen overlap and is the most complete representative of the recurring standardized output-tracing change in this cluster.", + "summary": "This cluster is mostly a family of model-specific refactors for standardized output tracing, all loosely tied to issue #43979, but the PRs are not all duplicates of each other. The clearest duplicate subgroup is the GPT-J/CodeGen line (#44066, #44085, #44722), while the GPT-Neo line (#44018, #44068) is separate.", + "confidence": 0.67, + "canonical_issue_reason": "All items are pull requests; the only issue reference is the external tracker #43979, so there is no issue artifact in this cluster to choose as canonical.", + "canonical_pr_reason": "PR #44722 is the strongest representative of the GPT-J/CodeGen subgroup: it is the later, broader revision, explicitly targets #43979, and overlaps the same GPT-J output-tracing code path as #44066 and #44085.", + "best_issue_reason": "No issue node is present here; #43979 is only referenced externally, so there is no in-cluster issue candidate to rank as best.", + "best_pr_reason": "#44722 is the best global representative among the accepted duplicates because it captures the same GPT-J tracing refactor as #44066/#44085 while being the most complete revision in that subgroup. The rest of the cluster is split across separate model-specific refactors.", "soft_edge_verdicts": [ { "left": "pull_request:43996", "right": "pull_request:44085", "accept": false, - "reason": "Same umbrella issue target, but different model families and different concrete code paths." + "reason": "Both are output-tracing refactors, but they touch different model families (CVT/FNet vs GPT-J) and different code paths, so they are not the same concrete change." }, { "left": "pull_request:43996", "right": "pull_request:44044", "accept": false, - "reason": "CVT/FNet refactor vs DeBERTa refactor; these are separate model implementations." + "reason": "Different models and implementations; this is a thematic similarity only, not the same underlying bug or patch." }, { "left": "pull_request:44066", "right": "pull_request:44085", "accept": true, - "reason": "Both refactor GPT-J output tracing on the same code path; 44066 just also includes CodeGen." + "reason": "Both target GPT-J output tracing in the same file and same code path; #44066 also includes CodeGen, but the GPT-J refactor overlaps enough to plausibly merge as one change." }, { "left": "pull_request:44007", "right": "pull_request:44072", "accept": false, - "reason": "RegNet and EfficientNet are distinct model code paths, so this is not the same change." + "reason": "These are separate image-model refactors (RegNet/ResNet/RT-DETR vs EfficientNet) and do not patch the same concrete implementation." }, { "left": "pull_request:44072", "right": "pull_request:44722", "accept": false, - "reason": "EfficientNet output tracing vs GPT-J/CodeGen output tracing are unrelated model-specific changes." + "reason": "Different subsystems and files; they share the generic output-tracing theme but not the same code-path problem." }, { "left": "pull_request:44066", "right": "pull_request:44072", "accept": false, - "reason": "Different model families and unrelated refactors." + "reason": "GPT-J/CodeGen vs EfficientNet are unrelated model paths; same refactor style is not enough to treat them as duplicates." }, { "left": "pull_request:44013", "right": "pull_request:44044", "accept": false, - "reason": "MobileNetV2 and DeBERTa are separate code paths with no concrete overlap." + "reason": "MobileNetV2 and DeBERTa v2 are different models with different output-handling code; not mergeable as one fix." }, { "left": "pull_request:44066", "right": "pull_request:44086", "accept": false, - "reason": "GPT-J/CodeGen refactor vs MGP-STR refactor; same theme, different implementation." - }, - { - "left": "pull_request:44066", - "right": "pull_request:44071", - "accept": false, - "reason": "GPT-J/CodeGen output tracing is unrelated to MPT's output tracing change." + "reason": "GPT-J/CodeGen output tracing and MGP-STR output tracing are separate implementations in different model stacks." }, { "left": "pull_request:44018", "right": "pull_request:44068", "accept": true, - "reason": "Same GPT-Neo output-tracing refactor on the same model/file; this looks like a direct duplicate or later revision." + "reason": "Both are GPT-Neo output-tracing refactors in the same model/file family and appear to be the same underlying change with slightly different revisions." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are different model code paths; similarity is only at the architectural refactor level." }, { "left": "pull_request:44066", "right": "pull_request:44068", "accept": false, - "reason": "GPT-J/CodeGen and GPT-Neo are different models and code paths." + "reason": "Different models (GPT-J vs GPT-Neo), so despite similar wording they are not the same concrete patch." }, { "left": "pull_request:44018", "right": "pull_request:44066", "accept": false, - "reason": "GPT-Neo refactor vs GPT-J/CodeGen refactor; not the same underlying change." + "reason": "GPT-Neo and GPT-J/CodeGen are distinct model implementations; the overlap is only the shared tracing pattern." }, { "left": "pull_request:44019", "right": "pull_request:44722", "accept": false, - "reason": "ResNet and GPT-J/CodeGen are unrelated model implementations." + "reason": "ResNet output tracing and GPT-J/CodeGen output tracing are unrelated concrete changes." }, { "left": "pull_request:44086", "right": "pull_request:44722", "accept": false, - "reason": "MGP-STR and GPT-J/CodeGen are different concrete changes." + "reason": "MGP-STR and GPT-J/CodeGen are separate model-specific changes; not the same underlying bug or patch." }, { "left": "pull_request:44019", "right": "pull_request:44085", "accept": false, - "reason": "ResNet vs GPT-J; same general pattern, different code paths." + "reason": "Different model families and code paths; they do not fix the same concrete issue." }, { "left": "pull_request:44018", "right": "pull_request:44019", "accept": false, - "reason": "GPT-Neo and ResNet are unrelated model refactors." + "reason": "GPT-Neo vs ResNet are separate model implementations, so these are not duplicates." }, { "left": "pull_request:44019", "right": "pull_request:44071", "accept": false, - "reason": "ResNet and MPT are separate code paths." + "reason": "ResNet and MPT output tracing are unrelated patches in different files." }, { "left": "pull_request:44068", "right": "pull_request:44116", "accept": false, - "reason": "GPT-Neo and Flaubert/XLM are different model families." + "reason": "GPT-Neo and Flaubert/XLM are different model stacks; same theme, different code paths." }, { - "left": "pull_request:44071", + "left": "pull_request:44068", "right": "pull_request:44722", "accept": false, - "reason": "MPT output tracing vs GPT-J/CodeGen output tracing are distinct changes." + "reason": "GPT-Neo output tracing is a separate subgroup from GPT-J/CodeGen output tracing." }, { - "left": "pull_request:44056", - "right": "pull_request:44086", + "left": "pull_request:44018", + "right": "pull_request:44116", "accept": false, - "reason": "MPNet and MGP-STR are different model implementations." + "reason": "GPT-Neo vs Flaubert/XLM are different concrete implementations, not one shared fix." }, { - "left": "pull_request:44068", - "right": "pull_request:44722", + "left": "pull_request:44085", + "right": "pull_request:44086", "accept": false, - "reason": "GPT-Neo refactor vs GPT-J/CodeGen refactor; not the same underlying bug/change." + "reason": "GPT-J and MGP-STR are unrelated model code paths." }, { - "left": "pull_request:44071", + "left": "pull_request:44059", "right": "pull_request:44086", "accept": false, - "reason": "MPT and MGP-STR are separate code paths." + "reason": "GPT-2 and MGP-STR are different models with different tracing logic; not the same change." }, { - "left": "pull_request:44059", + "left": "pull_request:44085", "right": "pull_request:44086", "accept": false, - "reason": "GPT2 and MGP-STR are unrelated refactors." + "reason": "The overlap is only broad output-tracing refactoring; the affected code paths are different model families." }, { "left": "pull_request:44059", "right": "pull_request:44722", "accept": false, - "reason": "GPT2 output tracing does not duplicate GPT-J/CodeGen output tracing." - }, - { - "left": "pull_request:44085", - "right": "pull_request:44086", - "accept": false, - "reason": "GPT-J and MGP-STR are different model code paths." + "reason": "GPT-2 output tracing is a separate implementation from GPT-J/CodeGen." }, { "left": "pull_request:44098", "right": "pull_request:44101", "accept": false, - "reason": "ViLT vs Flaubert/XLM; same broad theme, different concrete changes." + "reason": "ViLT and Flaubert/XLM are different model families and not the same concrete patch." }, { "left": "pull_request:44018", "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and MPT are distinct model implementations." + "reason": "GPT-Neo and MPT are distinct model implementations; this is not a duplicate fix." }, { "left": "pull_request:44017", "right": "pull_request:44722", "accept": false, - "reason": "Segformer output tracing is separate from GPT-J/CodeGen output tracing." + "reason": "SegFormer and GPT-J/CodeGen patches are unrelated beyond the shared decorator-style refactor theme." }, { "left": "pull_request:44017", "right": "pull_request:44161", "accept": false, - "reason": "Segformer and LongT5 are unrelated model code paths." + "reason": "SegFormer and LongT5 touch different model code paths and cannot be treated as one change." }, { "left": "pull_request:44017", "right": "pull_request:44073", "accept": false, - "reason": "Segformer and VisualBert are different model refactors." + "reason": "SegFormer and VisualBert are separate models; same general pattern does not make them duplicates." }, { "left": "pull_request:44068", "right": "pull_request:44071", "accept": false, - "reason": "GPT-Neo and MPT are not the same underlying change." + "reason": "GPT-Neo and MPT are different model families with different output-capture implementations." }, { "left": "pull_request:44071", "right": "pull_request:44086", "accept": false, - "reason": "MPT and MGP-STR are different model implementations." + "reason": "MPT and MGP-STR are unrelated concrete patches." }, { "left": "pull_request:44059", "right": "pull_request:44071", "accept": false, - "reason": "GPT2 and MPT output tracing are separate code paths." + "reason": "GPT-2 and MPT are separate model implementations and do not target the same code path." }, { - "left": "pull_request:43997", - "right": "pull_request:44073", + "left": "pull_request:44017", + "right": "pull_request:44071", "accept": false, - "reason": "RegNet and VisualBert are unrelated model refactors." + "reason": "SegFormer and MPT are different model families; not mergeable as one PR." }, { - "left": "pull_request:44074", - "right": "pull_request:44101", + "left": "pull_request:43997", + "right": "pull_request:44073", "accept": false, - "reason": "TextNet and Flaubert/XLM are different concrete changes." + "reason": "RegNet and VisualBert are different model code paths, so this is only a thematic similarity." }, { - "left": "pull_request:44017", - "right": "pull_request:44018", + "left": "pull_request:44059", + "right": "pull_request:44068", "accept": false, - "reason": "Segformer and GPT-Neo are different model-specific refactors." + "reason": "GPT-2 and GPT-Neo are distinct models; same refactor style does not imply duplicate content." }, { "left": "pull_request:44056", - "right": "pull_request:44116", - "accept": false, - "reason": "MPNet and Flaubert/XLM are separate code paths." - }, - { - "left": "pull_request:44073", "right": "pull_request:44101", "accept": false, - "reason": "VisualBert and Flaubert/XLM are not the same change." - }, - { - "left": "pull_request:43997", - "right": "pull_request:44056", - "accept": false, - "reason": "RegNet and MPNet are different model implementations." + "reason": "MPNet and Flaubert/XLM are separate implementations with different output handling." }, { "left": "pull_request:44073", "right": "pull_request:44161", "accept": false, - "reason": "VisualBert and LongT5 are unrelated output-tracing refactors." + "reason": "VisualBert and LongT5 are unrelated model-specific changes." }, { - "left": "pull_request:44074", + "left": "pull_request:44101", "right": "pull_request:44161", "accept": false, - "reason": "TextNet and LongT5 are different model code paths." + "reason": "Flaubert/XLM and LongT5 touch different code paths and are not the same fix." }, { - "left": "pull_request:43997", + "left": "pull_request:44017", "right": "pull_request:44018", "accept": false, - "reason": "RegNet and GPT-Neo are unrelated model changes." + "reason": "SegFormer and GPT-Neo are different models; shared refactor language is insufficient." }, { - "left": "pull_request:44071", - "right": "pull_request:44076", + "left": "pull_request:43997", + "right": "pull_request:44056", "accept": false, - "reason": "MPT and ImageGPT are different concrete refactors." + "reason": "RegNet and MPNet are separate model implementations with no shared concrete patch." }, { "left": "pull_request:44010", "right": "pull_request:44071", "accept": false, - "reason": "SqueezeBert and MPT are separate code paths." + "reason": "SqueezeBert and MPT are unrelated model code paths." }, { "left": "pull_request:44010", "right": "pull_request:44073", "accept": false, - "reason": "SqueezeBert and VisualBert are different model-specific changes." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44018", - "accept": false, - "reason": "SqueezeBert and GPT-Neo are unrelated model refactors." - }, - { - "left": "pull_request:44010", - "right": "pull_request:44044", - "accept": false, - "reason": "SqueezeBert and DeBERTa are different code paths." + "reason": "SqueezeBert and VisualBert are different models, so they are not duplicates." }, { "left": "pull_request:44066", "right": "pull_request:44722", "accept": true, - "reason": "Same GPT-J/CodeGen output-tracing change; 44722 is the later superset and overlaps the GPT-J code path." + "reason": "Both touch the same GPT-J output-tracing code path, and #44722 is a broader later revision that overlaps the same concrete change." }, { "left": "pull_request:44085", "right": "pull_request:44722", "accept": true, - "reason": "Both are GPT-J output-tracing refactors on the same underlying code path." + "reason": "Both are GPT-J tracing refactors in the same file; #44722 is the more complete version of the same underlying change." }, { "left": "pull_request:44085", "right": "pull_request:44129", "accept": false, - "reason": "GPT-J and SpeechT5 are unrelated model changes." + "reason": "GPT-J and SpeechT5 are different model families with different output-tracing logic." }, { "left": "pull_request:43996", "right": "pull_request:44007", "accept": false, - "reason": "CVT/FNet refactor vs RegNet refactor; different models." + "reason": "CVT/FNet and RegNet/ResNet/RT-DETR are different model code paths; not the same underlying fix." }, { "left": "pull_request:43996", "right": "pull_request:44129", "accept": false, - "reason": "CVT/FNet and SpeechT5 are separate model code paths." + "reason": "CVT/FNet and SpeechT5 are unrelated implementations beyond the generic tracing refactor theme." }, { "left": "pull_request:44044", "right": "pull_request:44129", "accept": false, - "reason": "DeBERTa and SpeechT5 are different concrete refactors." + "reason": "DeBERTa v2 and SpeechT5 patch different model code paths." }, { "left": "pull_request:44044", "right": "pull_request:44154", "accept": false, - "reason": "DeBERTa and Vits are unrelated model implementations." - }, - { - "left": "pull_request:44024", - "right": "pull_request:44076", - "accept": false, - "reason": "FocalNet and ImageGPT are different model-specific output-tracing changes." + "reason": "DeBERTa v2 and VITS are separate model implementations and not the same concrete change." } ] }, "evaluator_result": { "accept": true, - "feedback": "The summary stays conservative and matches the packet: most PRs are model-specific output-tracing refactors, while the only clearly mergeable overlap clusters are GPT-J/CodeGen and GPT-Neo. The accepted soft-edge pairs are all within the same underlying model code path or a later superset, so they look mergeable into a single bugfix PR. I don\u2019t see overstatement beyond the evidence provided." + "feedback": "The summary is grounded in the packet: it correctly treats this as a mostly model-specific output-tracing refactor cluster tied to #43979, keeps the GPT-J/CodeGen and GPT-Neo subgroups separate, and does not overmerge unrelated model families. The accepted soft edges also look mergeable as single PRs for the same bugfix within the same model/file family." }, "evaluator_used": true, "retried": false, @@ -16235,7 +16403,7 @@ { "label": "pull_request", "component_index": 17, - "component_count": 24, + "component_count": 25, "review_unit_index": 2, "review_unit_count": 2, "status": "reviewed", @@ -16247,76 +16415,85 @@ "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": "a19487638c7a863b25e5be7e1d25c8f267f7fe226ced74e554da54289f33f9a2", + "prepared_review_unit_hash": "4a3c93a7dc00b9269dfc12f1a55dc5b2b82b9c738f3431b16e68fb1b00d6d730", "component_budget": { "node_count": 28, "item_count": 28, - "soft_pair_count": 77, - "serialized_chars": 84507, - "estimated_input_tokens": 21127, - "estimated_eval_tokens": 42510 + "soft_pair_count": 78, + "serialized_chars": 84896, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 }, "budget": { - "node_count": 6, - "item_count": 6, - "soft_pair_count": 5, - "serialized_chars": 13068, - "estimated_input_tokens": 3267, - "estimated_eval_tokens": 6790 + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17123, + "estimated_input_tokens": 4281, + "estimated_eval_tokens": 8818 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": true, "analyst_result": { - "summary": "These PRs are all the same broad refactor pattern\u2014migrating model code to standardized output tracing\u2014but they target different model implementations, so they are not duplicates of one another. No issue is present; the best representative PR is the one with the clearest, most general refactor framing.", - "confidence": 0.91, + "summary": "These PRs all belong to the same broad housekeeping theme\u2014migrating individual model implementations to standardized output capture\u2014but they affect different models and code paths, so they are not duplicates of one another. None of the soft edges should be merged under strict triage.", + "confidence": 0.94, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #44073 is the cleanest representative of the cluster: it has an explicit standardized output-tracing refactor title, a straightforward single-model implementation, and the diff clearly shows the new capture_outputs/can_return_tuple setup.", + "canonical_pr_reason": "No PR is a duplicate of the others; if one must be chosen as the cluster representative, 44073 is the clearest exemplar because it explicitly wires in both capture_outputs and can_return_tuple plus model-specific output hooks.", "best_issue_reason": null, - "best_pr_reason": "PR #44073 is the best overall representative because it most directly expresses the shared refactor theme without extra model-specific complexity, while still showing the core output-tracing migration pattern.", + "best_pr_reason": "44073 is the best single representative of the theme because it shows the standardized output-tracing refactor in a concrete, self-contained model implementation. It is more representative than the smaller or more specialized variants, but it is still not a duplicate of them.", "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Different models (FocalNet vs ImageGPT) and different code paths; same refactor theme is not enough to treat them as the same change." + }, { "left": "pull_request:44073", "right": "pull_request:44074", "accept": false, - "reason": "Same refactor family, but different models (VisualBert vs TextNet) with separate code paths; they are not the same underlying change and would not be merged as one PR." + "reason": "VisualBert and TextNet are unrelated model implementations; these are parallel standardized-output refactors, not one concrete bug fix or mergeable duplicate PR." }, { "left": "pull_request:44056", "right": "pull_request:44073", "accept": false, - "reason": "Both are standardized output-tracing refactors, but they modify different model stacks (MPNet vs VisualBert) and different layer hooks, so they are not duplicates." + "reason": "MPNet and VisualBert touch different architectures and output-capture wiring; they are not the same underlying change." }, { "left": "pull_request:44056", "right": "pull_request:44074", "accept": false, - "reason": "These are model-specific refactors for MPNet and TextNet; similar mechanics do not make them the same bug or change." + "reason": "MPNet vs TextNet are separate model refactors, not a single fix to one code path." }, { "left": "pull_request:43997", "right": "pull_request:44010", "accept": false, - "reason": "Both touch output handling, but RegNet and SqueezeBert are unrelated implementations with distinct forward paths and no evidence of a shared concrete fix." + "reason": "RegNet and SqueezeBert are different models with different implementation details; this is only a thematic similarity." }, { "left": "pull_request:44010", "right": "pull_request:44129", "accept": false, - "reason": "SqueezeBert and SpeechT5 are different model families and the diffs show different output-tracing changes, so they are not the same underlying PR." + "reason": "SqueezeBert and SpeechT5 modify different model stacks and output semantics; not a duplicate or mergeable single PR." } ] }, @@ -16330,7 +16507,7 @@ { "label": "pull_request", "component_index": 18, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16367,16 +16544,13 @@ "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:44000|pull_request:44026", "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", "pull_request:44001|pull_request:44025", - "pull_request:44001|pull_request:44026", "pull_request:44002|pull_request:44029", - "pull_request:44002|pull_request:44025", - "pull_request:44001|pull_request:44003", "pull_request:44002|pull_request:44003", - "pull_request:44000|pull_request:44003", "pull_request:44003|pull_request:44004", "pull_request:43998|pull_request:44004", "pull_request:44001|pull_request:44004", @@ -16398,280 +16572,268 @@ "pull_request:44028|pull_request:44030", "pull_request:44029|pull_request:44030" ], - "prepared_review_unit_hash": "330f0ff17b99f83ce727ca052c729b8869874e3239f3c3196ffc7dc435f40071", + "prepared_review_unit_hash": "9a231e6af7b9802507e4562d09b573c6f1466a88f7f2b426ee9f2efba0b5cdc8", "component_budget": { "node_count": 13, "item_count": 13, - "soft_pair_count": 42, - "serialized_chars": 40936, - "estimated_input_tokens": 10234, - "estimated_eval_tokens": 20724 + "soft_pair_count": 39, + "serialized_chars": 39781, + "estimated_input_tokens": 9946, + "estimated_eval_tokens": 20148 }, "budget": { "node_count": 13, "item_count": 13, - "soft_pair_count": 42, - "serialized_chars": 40936, - "estimated_input_tokens": 10234, - "estimated_eval_tokens": 20724 + "soft_pair_count": 39, + "serialized_chars": 39781, + "estimated_input_tokens": 9946, + "estimated_eval_tokens": 20148 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are many similarly named but model-specific output-tracing refactors. They share a theme, but they do not describe the same underlying bug or change, so this is not a duplicate cluster.", - "confidence": 0.95, + "summary": "All items are separate model-specific output-tracing refactors. They share the same cleanup pattern, but they do not fix the same concrete code path and should not be deduplicated into one PR.", + "confidence": 0.97, "canonical_issue_reason": null, - "canonical_pr_reason": "No single canonical PR: each item targets a different model/code path, so they are not duplicates of one another even though the refactor theme is the same.", + "canonical_pr_reason": "No single PR is a true duplicate canonical here; the cluster is a set of parallel refactors across different model implementations.", "best_issue_reason": null, - "best_pr_reason": "No best PR as a duplicate representative; the safest representative is effectively none because the changes are separate, model-scoped refactors.", + "best_pr_reason": "PR 44025 is the best representative of the cluster theme because it is a focused multi-file output-tracing refactor, but it is still distinct from the others.", "soft_edge_verdicts": [ { "left": "pull_request:43998", "right": "pull_request:44002", "accept": false, - "reason": "Different models (`timm_backbone` vs `upernet`) and separate forward/output-handling code paths; same refactor theme only." + "reason": "Both are output-tracing refactors, but they touch different model code paths (`timm_backbone` vs `upernet`) and are not mergeable as one fix." }, { "left": "pull_request:43998", "right": "pull_request:44000", "accept": false, - "reason": "Different model families (`timm_backbone` vs `vision_text_dual_encoder`); not the same concrete fix." + "reason": "Same refactor theme, but different models and forward paths (`timm_backbone` vs `vision_text_dual_encoder`)." }, { "left": "pull_request:43998", "right": "pull_request:44001", "accept": false, - "reason": "`timm_backbone` and `univnet` are unrelated implementations; only the output-tracing refactor pattern overlaps." + "reason": "Shared cleanup pattern only; `timm_backbone` and `univnet` are unrelated model implementations." }, { "left": "pull_request:43998", "right": "pull_request:44003", "accept": false, - "reason": "`timm_backbone` is a single-model refactor, while `44003` spans `falcon_mamba` and `mamba`; not one underlying change." + "reason": "Both are refactors, but they affect different architectures (`timm_backbone` vs `falcon_mamba`/`mamba`) and different concrete code paths." }, { "left": "pull_request:43998", "right": "pull_request:44029", "accept": false, - "reason": "Unrelated model code paths (`timm_backbone` vs `rwkv`); same broad refactor style is insufficient." + "reason": "These are separate model-specific output-tracing changes (`timm_backbone` vs `rwkv`), not the same underlying bug or change." }, { "left": "pull_request:43999", "right": "pull_request:44001", "accept": false, - "reason": "`mobilenet_v1` and `univnet` are distinct models with different output semantics; not the same bug/change." + "reason": "Both modernize output handling, but `mobilenet_v1` and `univnet` are distinct implementations and cannot be merged into one PR." }, { "left": "pull_request:43999", "right": "pull_request:44002", "accept": false, - "reason": "Different model implementations (`mobilenet_v1` vs `upernet`); same tracing cleanup theme only." + "reason": "Same high-level pattern, but `mobilenet_v1` and `upernet` are different code paths and model families." }, { "left": "pull_request:43999", "right": "pull_request:44025", "accept": false, - "reason": "`mobilenet_v1` and `depth_anything` are separate model-specific refactors, not mergeable as one duplicate fix." + "reason": "Both are output-tracing refactors, but they target different models (`mobilenet_v1` vs `depth_anything`) and different forward signatures." }, { "left": "pull_request:43999", "right": "pull_request:44026", "accept": false, - "reason": "Different encoder/decoder wrapper vs mobile vision model; not the same concrete code-path problem." + "reason": "These are separate model refactors (`mobilenet_v1` vs `vision_encoder_decoder`), not one concrete fix." }, { "left": "pull_request:43999", "right": "pull_request:44027", "accept": false, - "reason": "`mobilenet_v1` and `speech_encoder_decoder` are unrelated code paths; no duplicate evidence beyond the shared refactor pattern." + "reason": "Shared refactor style only; `mobilenet_v1` and `speech_encoder_decoder` are unrelated code paths." }, { "left": "pull_request:44000", "right": "pull_request:44001", "accept": false, - "reason": "Different models (`vision_text_dual_encoder` vs `univnet`); the changes are not the same underlying fix." + "reason": "Both are output-tracing refactors, but they change different model stacks (`vision_text_dual_encoder` vs `univnet`)." }, { "left": "pull_request:44000", "right": "pull_request:44002", "accept": false, - "reason": "`vision_text_dual_encoder` and `upernet` have distinct forward signatures and output behavior; not a duplicate." + "reason": "Different models and different forward implementations (`vision_text_dual_encoder` vs `upernet`)." }, { "left": "pull_request:44000", - "right": "pull_request:44025", + "right": "pull_request:44003", "accept": false, - "reason": "Different model families and different output-tracing implementations; too broad to treat as one change." + "reason": "Although both are tracing refactors, they apply to unrelated model code paths and cannot plausibly be merged into one PR." }, { "left": "pull_request:44000", - "right": "pull_request:44026", + "right": "pull_request:44025", "accept": false, - "reason": "`vision_text_dual_encoder` vs `vision_encoder_decoder` is only superficially similar; the code paths are separate." + "reason": "Same theme, but `vision_text_dual_encoder` and `depth_anything` are different model families with separate fixes." }, { "left": "pull_request:44001", "right": "pull_request:44002", "accept": false, - "reason": "`univnet` and `upernet` are unrelated architectures; same refactor theme does not imply duplicate change." + "reason": "Both are output-tracing cleanup, but they affect different models (`univnet` vs `upernet`) and separate forward contracts." }, { "left": "pull_request:44001", - "right": "pull_request:44025", + "right": "pull_request:44003", "accept": false, - "reason": "Different model-specific output cleanup; no shared concrete bug or code path." + "reason": "These are not the same bug/change; they target different model implementations (`univnet` vs `mamba`/`falcon_mamba`)." }, { "left": "pull_request:44001", - "right": "pull_request:44026", + "right": "pull_request:44025", "accept": false, - "reason": "`univnet` and `vision_encoder_decoder` target different subsystems and are not plausibly one merged PR." + "reason": "Different model families and different code paths; same refactor pattern is not enough for a duplicate." }, { "left": "pull_request:44002", "right": "pull_request:44029", "accept": false, - "reason": "`upernet` and `rwkv` are unrelated model implementations; not the same underlying fix." - }, - { - "left": "pull_request:44002", - "right": "pull_request:44025", - "accept": false, - "reason": "Separate model-specific refactors with no shared code path." - }, - { - "left": "pull_request:44001", - "right": "pull_request:44003", - "accept": false, - "reason": "`univnet` and `falcon_mamba/mamba` touch different architectures and output mechanisms." + "reason": "`upernet` and `rwkv` are unrelated model implementations; this is only a thematic similarity." }, { "left": "pull_request:44002", "right": "pull_request:44003", "accept": false, - "reason": "`upernet` and `falcon_mamba/mamba` are different model families; only the refactor style matches." - }, - { - "left": "pull_request:44000", - "right": "pull_request:44003", - "accept": false, - "reason": "Different model code paths; not the same concrete bug or change." + "reason": "Shared output-tracing cleanup, but the underlying forward logic lives in different models and is not mergeable as one PR." }, { "left": "pull_request:44003", "right": "pull_request:44004", "accept": false, - "reason": "`mamba`/`falcon_mamba` output-tracing refactor is unrelated to `codegen` cache/output changes." + "reason": "Different subsystems entirely (`mamba`/`falcon_mamba` vs `codegen`), so these are not the same concrete change." }, { "left": "pull_request:43998", "right": "pull_request:44004", "accept": false, - "reason": "`timm_backbone` and `codegen` affect different subsystems and call paths; not a duplicate." + "reason": "Both touch output handling, but `timm_backbone` and `codegen` are unrelated code paths." }, { "left": "pull_request:44001", "right": "pull_request:44004", "accept": false, - "reason": "`univnet` and `codegen` are separate model implementations with different output contracts." + "reason": "These are separate refactors in different model families (`univnet` vs `codegen`)." }, { "left": "pull_request:44002", "right": "pull_request:44004", "accept": false, - "reason": "`upernet` and `codegen` are unrelated; same broad refactor label is not enough." + "reason": "`upernet` and `codegen` do not share the same underlying fix; only the refactor motif matches." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Different models and different forward contracts; not a duplicate PR pair." }, { "left": "pull_request:44025", "right": "pull_request:44026", "accept": false, - "reason": "`depth_anything` and `vision_encoder_decoder` are separate model wrappers; not the same fix." + "reason": "Both are output-tracing refactors, but they target different wrappers (`depth_anything` vs `vision_encoder_decoder`) and distinct forward paths." }, { "left": "pull_request:44025", "right": "pull_request:44027", "accept": false, - "reason": "`depth_anything` and `speech_encoder_decoder` are unrelated model-specific refactors." + "reason": "`depth_anything` and `speech_encoder_decoder` are separate model-specific changes, not one shared fix." }, { "left": "pull_request:44025", "right": "pull_request:44028", "accept": false, - "reason": "`depth_anything` and `superpoint` use different model classes and output code paths." + "reason": "Different model implementations (`depth_anything` vs `superpoint`), so this is not the same underlying bug or change." }, { "left": "pull_request:44025", "right": "pull_request:44029", "accept": false, - "reason": "`depth_anything` and `rwkv` are distinct architectures; same tracing refactor theme only." + "reason": "These are parallel output-tracing refactors in different models (`depth_anything` vs `rwkv`)." }, { "left": "pull_request:44025", "right": "pull_request:44030", "accept": false, - "reason": "`depth_anything` and `dpr` target different models and output contracts; not duplicates." + "reason": "Different code paths and model families (`depth_anything` vs `dpr`); same pattern, not same PR." }, { "left": "pull_request:44026", "right": "pull_request:44029", "accept": false, - "reason": "`vision_encoder_decoder` and `rwkv` are unrelated code paths; not one underlying issue." + "reason": "Both are refactors around model outputs, but `vision_encoder_decoder` and `rwkv` are unrelated implementations." }, { "left": "pull_request:44026", "right": "pull_request:44027", "accept": false, - "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are different wrappers, but still separate changes." + "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are similar wrapper types, but they are distinct code paths and separate changes." }, { "left": "pull_request:44026", "right": "pull_request:44028", "accept": false, - "reason": "`vision_encoder_decoder` and `superpoint` are different model families and not mergeable as one PR." + "reason": "Different model families (`vision_encoder_decoder` vs `superpoint`); not one concrete bugfix." }, { "left": "pull_request:44026", "right": "pull_request:44030", "accept": false, - "reason": "`vision_encoder_decoder` and `dpr` are separate model-specific output changes." + "reason": "Separate model-specific refactors (`vision_encoder_decoder` vs `dpr`), so they should not be merged as duplicates." }, { "left": "pull_request:44027", "right": "pull_request:44028", "accept": false, - "reason": "`speech_encoder_decoder` and `superpoint` do not share a concrete code path or bug." + "reason": "These touch different models (`speech_encoder_decoder` vs `superpoint`) and do not represent the same underlying change." }, { "left": "pull_request:44027", "right": "pull_request:44029", "accept": false, - "reason": "`speech_encoder_decoder` and `rwkv` are unrelated model implementations." + "reason": "Unrelated model code paths (`speech_encoder_decoder` vs `rwkv`), despite sharing the same refactor theme." }, { "left": "pull_request:44027", "right": "pull_request:44030", "accept": false, - "reason": "`speech_encoder_decoder` and `dpr` are separate model wrappers with different output handling." + "reason": "Different model-specific output handling changes; not a duplicate pair." }, { "left": "pull_request:44028", "right": "pull_request:44029", "accept": false, - "reason": "`superpoint` and `rwkv` are different architectures; only the refactor theme overlaps." + "reason": "`superpoint` and `rwkv` are unrelated code paths, so this is only superficial similarity." }, { "left": "pull_request:44028", "right": "pull_request:44030", "accept": false, - "reason": "`superpoint` and `dpr` are separate model-specific changes, not a single duplicate fix." + "reason": "Different models and different forward implementations (`superpoint` vs `dpr`)." }, { "left": "pull_request:44029", "right": "pull_request:44030", "accept": false, - "reason": "`rwkv` and `dpr` touch different model code paths and output behavior; not the same change." + "reason": "Both are model-specific refactors, but they affect different architectures (`rwkv` vs `dpr`) and are not mergeable into one PR." } ] }, @@ -16685,7 +16847,7 @@ { "label": "pull_request", "component_index": 19, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16704,52 +16866,52 @@ "pull_request:44662|pull_request:45075", "pull_request:45075|pull_request:45401" ], - "prepared_review_unit_hash": "ab317a3058e004cc23c01a74311c78198d533eebb7988efc3933cd0e8336c847", + "prepared_review_unit_hash": "1ae9d8cacd76a707d18b86d11c11f686fb0d37db393af09f8a8ed476def6f407", "component_budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 11073, - "estimated_input_tokens": 2769, - "estimated_eval_tokens": 5794 + "serialized_chars": 11071, + "estimated_input_tokens": 2768, + "estimated_eval_tokens": 5792 }, "budget": { "node_count": 3, "item_count": 3, "soft_pair_count": 3, - "serialized_chars": 11073, - "estimated_input_tokens": 2769, - "estimated_eval_tokens": 5794 + "serialized_chars": 11071, + "estimated_input_tokens": 2768, + "estimated_eval_tokens": 5792 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are three separate model-addition PRs for different architectures (PenguinVL, DeepSeek-OCR-2, Voxtral TTS). They share the usual Transformers integration scaffold, but they do not look like the same underlying change.", - "confidence": 0.97, + "summary": "These are three separate model-integration pull requests for different models (PenguinVL, DeepSeek-OCR-2, Voxtral TTS). They share common Transformers plumbing files, but the implementations, model classes, and targets are distinct, so they are not duplicates.", + "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "No canonical duplicate PR emerges here; the PRs are distinct model implementations rather than overlapping fixes.", + "canonical_pr_reason": "PR #45075 is the most mature-looking standalone integration: it is not marked draft, has substantial review activity, and includes a full model package plus tests. It is not a duplicate of the others, but it is the best single representative if one must be chosen from this cluster.", "best_issue_reason": null, - "best_pr_reason": "PR 44662 is the strongest standalone representative in the group: non-draft, broad implementation/test coverage, and high review activity. It is still not a duplicate of the other PRs.", + "best_pr_reason": "PR #45075 is the strongest representative for the cluster because it is a complete, non-draft model addition with substantial review activity. The other PRs are also standalone model additions, but they are separate changes for different architectures and should not be merged into one.", "soft_edge_verdicts": [ { "left": "pull_request:44662", "right": "pull_request:45401", "accept": false, - "reason": "PenguinVL vs Voxtral TTS are different model integrations. Shared auto-registry and docs plumbing are generic scaffold, not the same fix/change." + "reason": "Different models and code paths: PenguinVL vs Voxtral TTS. Shared auto/docs registration files are expected for independent model additions and do not make these the same change." }, { "left": "pull_request:44662", "right": "pull_request:45075", "accept": false, - "reason": "PenguinVL and DeepSeek-OCR-2 are unrelated model additions. Similar filenames reflect standard model onboarding, not one concrete shared code-path change." + "reason": "Different model integrations: PenguinVL and DeepSeek-OCR-2. Overlap is limited to shared Transformers registry/docs boilerplate, not the same underlying bug or feature." }, { "left": "pull_request:45075", "right": "pull_request:45401", "accept": false, - "reason": "DeepSeek-OCR-2 and Voxtral TTS are distinct model implementations with different model-specific files and behavior; this is not mergeable as one duplicate PR." + "reason": "DeepSeek-OCR-2 and Voxtral TTS are unrelated model additions with different implementations and targets. The shared files are generic registration updates, not evidence of one concrete mergeable change." } ] }, @@ -16763,7 +16925,7 @@ { "label": "pull_request", "component_index": 20, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16779,12 +16941,12 @@ "soft_pairs": [ "pull_request:44979|pull_request:45363" ], - "prepared_review_unit_hash": "247a8b461d08abde754a61f6383060cbf593f427f418fd093c4a7f6b5ac240a0", + "prepared_review_unit_hash": "3d300441ed7501b66471d4b3bfff21be52eb6f034b5b703b375894d6cb13dffd", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4407, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -16792,7 +16954,7 @@ "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 4408, + "serialized_chars": 4407, "estimated_input_tokens": 1102, "estimated_eval_tokens": 2460 }, @@ -16801,18 +16963,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These two PRs are related feature work, but they are not duplicates: PR #44979 introduces a standalone Module Fusion API and tests, while PR #45363 wires KernelConfig-based n-to-1 kernel fusion into model loading/integration code. They overlap conceptually, not as the same concrete change.", - "confidence": 0.91, + "summary": "These PRs are related to model/kernel fusion, but they address different layers of the stack: one introduces a generic module fusion API, while the other adds a specific n-to-1 kernel-fusion path via KernelConfig and integration changes. They are not duplicates.", + "confidence": 0.95, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #44979 is the more canonical representative because it defines the core module-fusion API surface and tests, which the other PR can build on or integrate with.", + "canonical_pr_reason": null, "best_issue_reason": null, - "best_pr_reason": "PR #44979 best represents the cluster as the foundational implementation of module fusion primitives rather than a specific integration example.", + "best_pr_reason": "PR 45363 is the more concrete end-user feature: it wires fusion into existing loading/integration code paths and includes an example, making it the better standalone representative of the cluster.", "soft_edge_verdicts": [ { "left": "pull_request:44979", "right": "pull_request:45363", "accept": false, - "reason": "Different change sets and code paths: #44979 adds a new module_fusion API, while #45363 adds KernelConfig-driven n-to-1 fusion integration and example code. They are related features, not the same bug/fix or a mergeable single PR." + "reason": "Different concrete changes: 44979 adds a generic Module Fusion API and tests, while 45363 implements n-to-1 kernel fusion through KernelConfig and integration plumbing. Related topic, but not the same code-path change." } ] }, @@ -16826,7 +16988,7 @@ { "label": "pull_request", "component_index": 21, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16842,7 +17004,7 @@ "soft_pairs": [ "pull_request:45086|pull_request:45317" ], - "prepared_review_unit_hash": "7d7c419fbad4f8ceeeae1af97c511ade8eb19d28975c021389c0640e201e0456", + "prepared_review_unit_hash": "11ee7051d443ee9ae38db57d12f6b21717af7160b6450dfca79d45f7071e2e98", "component_budget": { "node_count": 2, "item_count": 2, @@ -16864,18 +17026,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "Both PRs fix the same Mistral regex patch AttributeError in `src/transformers/tokenization_utils_tokenizers.py`; 45317 is the stronger canonical PR because it includes the same code change plus tests.", - "confidence": 0.96, + "summary": "Both pull requests target the same Mistral regex patch in the same file, with essentially identical code changes; PR 45317 is the more complete version because it also adds tests.", + "confidence": 0.99, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45317 is the more complete version of the same fix: same code-path change as 45086, plus test coverage in `tests/models/auto/test_tokenization_auto.py`.", + "canonical_pr_reason": "PR 45317 is the better canonical PR because it makes the same fix as 45086 and includes accompanying test coverage, so it is the more complete representative of the duplicate pair.", "best_issue_reason": null, - "best_pr_reason": "45317 is the best representative because it addresses the same concrete bug as 45086 and adds validation, making it the most complete merge candidate.", + "best_pr_reason": "PR 45317 is the best representative of the cluster: same underlying fix as 45086, plus tests, making it the more merge-ready and informative duplicate target.", "soft_edge_verdicts": [ { "left": "pull_request:45086", "right": "pull_request:45317", "accept": true, - "reason": "Same underlying AttributeError fix in the same function and file, with identical core diff; 45317 simply adds tests." + "reason": "Same concrete bug fix in the same code path (_patch_mistral_regex / tokenization_utils_tokenizers.py); the diffs are effectively the same, and the second PR just adds tests, so they plausibly could be merged as one change." } ] }, @@ -16889,7 +17051,7 @@ { "label": "pull_request", "component_index": 22, - "component_count": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16905,7 +17067,7 @@ "soft_pairs": [ "pull_request:45332|pull_request:45333" ], - "prepared_review_unit_hash": "3e4e003158cf9f75f31d454f0830428b3e67187aa75e017ffc7d3f0ae0bc9724", + "prepared_review_unit_hash": "8ffe091d15de12d586690691501506c4eab49839b4bd71825ec8ffce82e4bbeb", "component_budget": { "node_count": 2, "item_count": 2, @@ -16927,18 +17089,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are related feature PRs for heterogeneous model support, but they are not the same change: one adds per-layer heterogeneous config plumbing, the other builds heterogeneous modeling/cache behavior on top of it. I would not merge them as duplicates.", - "confidence": 0.88, + "summary": "These are closely related heterogeneous-feature PRs, but they are not the same change: one adds per-layer config plumbing, the other adds broader heterogeneous model/runtime support. They overlap in config utilities and tests, yet they target different parts of the codepath and could reasonably land as separate PRs.", + "confidence": 0.91, "canonical_issue_reason": null, - "canonical_pr_reason": "PR 45332 is the broader umbrella change: it adds heterogeneous model support and also touches the config/cache/modeling integration, so it better represents the full feature area than the config-only PR.", + "canonical_pr_reason": "PR #45332 is the broader umbrella change, covering heterogeneous modeling/runtime support in addition to config plumbing, while #45333 is the narrower config-only foundation.", "best_issue_reason": null, - "best_pr_reason": "PR 45332 is the best representative artifact because it is the more complete end-to-end implementation and depends on the config-side groundwork reflected in PR 45333.", + "best_pr_reason": "PR #45332 is the best representative of the cluster because it contains the larger end-to-end heterogeneous support work and subsumes the config changes used by #45333.", "soft_edge_verdicts": [ { "left": "pull_request:45332", "right": "pull_request:45333", "accept": false, - "reason": "They share heterogeneity infrastructure, but they address different layers of the feature: 45333 adds per-layer config support, while 45332 adds modeling/cache behavior for heterogeneous models. They are related, but not the same concrete change and would not plausibly collapse into one duplicate PR." + "reason": "Related feature work, but not the same concrete change: #45333 adds per-layer config support, while #45332 adds heterogeneous modeling/cache/runtime support. They overlap in shared helpers, yet they are separate slices of the feature and not safely mergeable as duplicates." } ] }, @@ -16952,7 +17114,88 @@ { "label": "pull_request", "component_index": 23, - "component_count": 24, + "component_count": 25, + "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-45351-3", + "nodes": [ + "pull_request:45351", + "pull_request:45509", + "pull_request:45515" + ], + "soft_pairs": [ + "pull_request:45351|pull_request:45509", + "pull_request:45351|pull_request:45515", + "pull_request:45509|pull_request:45515" + ], + "prepared_review_unit_hash": "12efe7296ae86d20a9691bb3beafb427279ef36b4754af0e6721218cdd6f9c1f", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5821, + "estimated_input_tokens": 1456, + "estimated_eval_tokens": 3168 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5821, + "estimated_input_tokens": 1456, + "estimated_eval_tokens": 3168 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three pull requests target the same crash in `get_device_properties()` when CUDA/ROCm is detected but `torch.cuda.is_available()` is false. They are close variants of the same fix, with PR 45351 the most complete/canonical version.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45351 is the strongest canonical PR: it explicitly links the tracking issue, guards the CUDA/ROCm capability call with `torch.cuda.is_available()`, and keeps the XPU path intact in the same function.", + "best_issue_reason": null, + "best_pr_reason": "PR 45351 is the best representative of the cluster because it is the most complete and coherent fix for the exact crash path, while the other two are narrower variants of the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45351", + "right": "pull_request:45509", + "accept": true, + "reason": "Same underlying bug in `get_device_properties()`: avoid calling CUDA capability APIs when CUDA is present but unavailable. Both PRs patch the same code path and could plausibly be consolidated." + }, + { + "left": "pull_request:45351", + "right": "pull_request:45515", + "accept": true, + "reason": "Same crash fix in the same function, with nearly identical logic around `torch.cuda.is_available()`. These are alternate implementations of the same change." + }, + { + "left": "pull_request:45509", + "right": "pull_request:45515", + "accept": true, + "reason": "Both target the same `get_device_properties()` availability guard for CUDA/ROCm. They address the same concrete failure mode and are mergeable into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: all three PRs touch the same `get_device_properties()` path in `src/transformers/testing_utils.py` and address the CUDA/ROCm availability crash. Calling PR 45351 the most canonical/comprehensive version is reasonable because it has the explicit issue link and preserves the XPU branch in the same refactor. The pairwise soft-edge accepts are also defensible as the PRs look like close variants of the same fix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 24, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -16968,40 +17211,40 @@ "soft_pairs": [ "pull_request:45415|pull_request:45425" ], - "prepared_review_unit_hash": "6e4909aaaea00bafd711cf7883af51cd69ffd6dcfad919d7d6544a45101d8fa7", + "prepared_review_unit_hash": "09cbc2edb152fd0953da59dc909841a68059cdd04df707cfd3f0c1fa39f448ec", "component_budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5304, - "estimated_input_tokens": 1326, - "estimated_eval_tokens": 2908 + "serialized_chars": 5306, + "estimated_input_tokens": 1327, + "estimated_eval_tokens": 2910 }, "budget": { "node_count": 2, "item_count": 2, "soft_pair_count": 1, - "serialized_chars": 5304, - "estimated_input_tokens": 1326, - "estimated_eval_tokens": 2908 + "serialized_chars": 5306, + "estimated_input_tokens": 1327, + "estimated_eval_tokens": 2910 }, "overflow_policy": "truncate_then_skip", "trimmed": false, "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are two separate typing-focused PRs, not duplicates: one is a broad type-checking sweep across many transformers modules, while the other adds specific typing helpers for modeling_utils/PEFT integration.", - "confidence": 0.9, + "summary": "These PRs are both typing-related, but they address different changes: 45415 is a broad repo-wide type-checking sweep, while 45425 adds specific typing protocols/helpers around modeling_utils/PEFT. They do not appear to fix the same concrete code-path problem, so I would not cluster them as duplicates.", + "confidence": 0.93, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45415 is the broader, more representative typing sweep and best captures the cluster\u2019s general theme.", + "canonical_pr_reason": "PR 45425 is the more focused and concrete typing change; it better represents the cluster than the broad type-checking sweep in 45415.", "best_issue_reason": null, - "best_pr_reason": "PR #45415 is the best representative of the cluster because it covers the wider cross-file type-checking change set.", + "best_pr_reason": "PR 45425 is the best representative because it makes a specific typing/API adjustment in a narrower area, whereas 45415 is a large mechanical type-checking pass.", "soft_edge_verdicts": [ { "left": "pull_request:45415", "right": "pull_request:45425", "accept": false, - "reason": "Both are typing-related, but they address different scopes and code paths; #45415 is a broad lint/type-checking cleanup, while #45425 adds targeted typing protocols/helpers for specific modeling_utils/PEFT uses. They would not plausibly be merged as one concrete fix." + "reason": "Shared typing-related files are not enough here: 45415 is a broad annotation/type-checking sweep across many modules, while 45425 adds targeted typing protocols and modeling_utils integration. They are not the same underlying change and would not plausibly merge into one PR." } ] }, @@ -17014,8 +17257,8 @@ }, { "label": "pull_request", - "component_index": 24, - "component_count": 24, + "component_index": 25, + "component_count": 25, "review_unit_index": 1, "review_unit_count": 1, "status": "reviewed", @@ -17031,7 +17274,7 @@ "soft_pairs": [ "pull_request:45470|pull_request:45487" ], - "prepared_review_unit_hash": "143e61e685e46b05b7f94a899a20fadbe95580e2e71f5c82773eae430c6827eb", + "prepared_review_unit_hash": "1c2d791880bc098c4a5b373854e202acece21f45d93cc94dea699fc37380d767", "component_budget": { "node_count": 2, "item_count": 2, @@ -17053,18 +17296,18 @@ "aggressively_trimmed": false, "split": false, "analyst_result": { - "summary": "These are not duplicates: one PR is a broad model-parallel/device-placement fix across multiple model implementations, while the other only skips a SAM3 Lite text test due to a test logic mismatch. They do not look like the same underlying change or bug.", + "summary": "These are not duplicates: one PR is a targeted test skip for Sam3LiteText flash-attn dispatch behavior, while the other is a broad model-parallel fix across multiple model implementations. They do not appear to fix the same concrete bug or code path.", "confidence": 0.96, "canonical_issue_reason": null, - "canonical_pr_reason": "PR #45487 is the substantive code fix: it addresses a concrete model-parallel/device issue in AltCLIP/ChineseCLIP and related model files, with implementation changes across model code. PR #45470 is only a targeted test skip and is not the main fix.", + "canonical_pr_reason": "PR 45487 is the stronger canonical PR because it makes an actual code-path fix across model implementations, whereas 45470 only skips a test in one model suite.", "best_issue_reason": null, - "best_pr_reason": "PR #45487 is the best representative because it fixes an actual code-path problem, spans the relevant model implementations, and is the more durable change compared with a one-off test skip in PR #45470.", + "best_pr_reason": "PR 45487 is the best representative of the cluster since it addresses a substantive runtime/model-parallel issue; 45470 is just a test adjustment and is not the same change.", "soft_edge_verdicts": [ { "left": "pull_request:45470", "right": "pull_request:45487", "accept": false, - "reason": "Different scope and intent: #45470 only skips a SAM3 Lite text flash-attn test, while #45487 changes AltCLIP/ChineseCLIP and related model code to fix a model-parallel/device issue. They do not appear to address the same underlying bug or be mergeable into one PR." + "reason": "Reject: 45470 only skips a flash-attn dispatch test for Sam3LiteText, while 45487 changes model-parallel handling in several model files. Different underlying problems and not plausibly one merged PR." } ] }, diff --git a/analysis/current/manifest.json b/analysis/current/manifest.json index 5d3ec9d45ee19d8c108480479cedfdf477721bc1..d079445b02a81b1d8b33bf964ac272b96a79efa3 100644 --- a/analysis/current/manifest.json +++ b/analysis/current/manifest.json @@ -1,8 +1,8 @@ { - "analysis_id": "hybrid-gpt54mini-20260418t170534z", + "analysis_id": "hybrid-model-20260420t094431z", "archived_artifacts": { - "hybrid": "snapshots/20260418T170534Z/analysis-runs/hybrid-gpt54mini-20260418t170534z/analysis-report-hybrid.json", - "hybrid_reviews": "snapshots/20260418T170534Z/analysis-runs/hybrid-gpt54mini-20260418t170534z/analysis-report-hybrid.llm-reviews.json" + "hybrid": "snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.llm-reviews.json" }, "artifacts": { "hybrid": "analysis/current/analysis-report-hybrid.json", @@ -10,9 +10,9 @@ }, "channel": "canonical", "model": null, - "published_at": "2026-04-18T19:17:20Z", + "published_at": "2026-04-20T09:57:25Z", "repo": "huggingface/transformers", "schema_version": 1, - "snapshot_id": "20260418T170534Z", + "snapshot_id": "20260420T094431Z", "variant": "hybrid" } diff --git a/snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.json b/snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.json new file mode 100644 index 0000000000000000000000000000000000000000..aeeb3aaa5ccc6b2c9c61bbcca6f8cce9e91703a3 --- /dev/null +++ b/snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.json @@ -0,0 +1,2240 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260420T094431Z", + "generated_at": "2026-04-20T09:57:23Z", + "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": 0.75, + "canonical_issue_number": 41211, + "canonical_pr_number": 44339, + "issue_numbers": [ + 41211 + ], + "pr_numbers": [ + 41356, + 44339 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 41356, + "right_pr_number": 44339, + "code_similarity": 0.164, + "size_similarity": 0.096, + "file_overlap": 0.3, + "area_overlap": 0.078, + "patch_similarity": 0.182, + "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, + 1753 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43656-4", + "summary": "Cluster of 1 issues and 3 PRs centered on issue #43824.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43824, + "canonical_pr_number": 43656, + "issue_numbers": [ + 43824 + ], + "pr_numbers": [ + 43656, + 43836, + 43842 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43656, + "right_pr_number": 43836, + "code_similarity": 0.36, + "size_similarity": 0.176, + "file_overlap": 1.0, + "area_overlap": 0.051, + "patch_similarity": 0.048, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 30, + 36 + ], + [ + 315, + 323 + ], + [ + 665, + 671 + ], + [ + 931, + 937 + ], + [ + 1843, + 1849 + ], + [ + 1868, + 1874 + ] + ], + "right_ranges": [ + [ + 11, + 18 + ], + [ + 359, + 429 + ], + [ + 584, + 590 + ], + [ + 1892, + 1910 + ], + [ + 1917, + 1923 + ] + ] + } + ] + }, + { + "left_pr_number": 43656, + "right_pr_number": 43842, + "code_similarity": 0.405, + "size_similarity": 0.5, + "file_overlap": 1.0, + "area_overlap": 0.0, + "patch_similarity": 0.036, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 30, + 36 + ], + [ + 315, + 323 + ], + [ + 665, + 671 + ], + [ + 931, + 937 + ], + [ + 1843, + 1849 + ], + [ + 1868, + 1874 + ] + ], + "right_ranges": [ + [ + 54, + 61 + ], + [ + 587, + 593 + ] + ] + } + ] + }, + { + "left_pr_number": 43836, + "right_pr_number": 43842, + "code_similarity": 0.332, + "size_similarity": 0.088, + "file_overlap": 1.0, + "area_overlap": 0.033, + "patch_similarity": 0.017, + "shared_filenames": [ + "src/transformers/cli/serve.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/cli/serve.py", + "left_ranges": [ + [ + 11, + 18 + ], + [ + 359, + 429 + ], + [ + 584, + 590 + ], + [ + 1892, + 1910 + ], + [ + 1917, + 1923 + ] + ], + "right_ranges": [ + [ + 54, + 61 + ], + [ + 587, + 593 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43366-4", + "summary": "Cluster of 1 issues and 3 PRs centered on issue #43366.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 43366, + "canonical_pr_number": 45506, + "issue_numbers": [ + 43366 + ], + "pr_numbers": [ + 43757, + 45500, + 45506 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 43757, + "right_pr_number": 45500, + "code_similarity": 0.096, + "size_similarity": 0.149, + "file_overlap": 0.2, + "area_overlap": 0.005, + "patch_similarity": 0.031, + "shared_filenames": [ + "src/transformers/modeling_gguf_pytorch_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/modeling_gguf_pytorch_utils.py", + "left_ranges": [ + [ + 462, + 474 + ] + ], + "right_ranges": [ + [ + 171, + 277 + ], + [ + 456, + 462 + ], + [ + 518, + 525 + ], + [ + 620, + 627 + ], + [ + 709, + 753 + ] + ] + } + ] + }, + { + "left_pr_number": 43757, + "right_pr_number": 45506, + "code_similarity": 0.096, + "size_similarity": 0.15, + "file_overlap": 0.2, + "area_overlap": 0.005, + "patch_similarity": 0.031, + "shared_filenames": [ + "src/transformers/modeling_gguf_pytorch_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/modeling_gguf_pytorch_utils.py", + "left_ranges": [ + [ + 462, + 474 + ] + ], + "right_ranges": [ + [ + 171, + 277 + ], + [ + 456, + 462 + ], + [ + 518, + 525 + ], + [ + 620, + 627 + ], + [ + 709, + 753 + ] + ] + } + ] + }, + { + "left_pr_number": 45500, + "right_pr_number": 45506, + "code_similarity": 0.997, + "size_similarity": 0.996, + "file_overlap": 1.0, + "area_overlap": 0.996, + "patch_similarity": 0.996, + "shared_filenames": [ + "docs/source/en/model_doc/gpt_oss.md", + "src/transformers/integrations/ggml.py", + "src/transformers/modeling_gguf_pytorch_utils.py", + "tests/quantization/ggml/test_ggml.py" + ], + "shared_file_areas": [ + { + "filename": "docs/source/en/model_doc/gpt_oss.md", + "left_ranges": [ + [ + 70, + 93 + ] + ], + "right_ranges": [ + [ + 70, + 93 + ] + ] + }, + { + "filename": "src/transformers/integrations/ggml.py", + "left_ranges": [ + [ + 89, + 109 + ] + ], + "right_ranges": [ + [ + 89, + 109 + ] + ] + }, + { + "filename": "src/transformers/modeling_gguf_pytorch_utils.py", + "left_ranges": [ + [ + 171, + 277 + ], + [ + 456, + 462 + ], + [ + 518, + 525 + ], + [ + 620, + 627 + ], + [ + 709, + 753 + ] + ], + "right_ranges": [ + [ + 171, + 277 + ], + [ + 456, + 462 + ], + [ + 518, + 525 + ], + [ + 620, + 627 + ], + [ + 709, + 753 + ] + ] + }, + { + "filename": "tests/quantization/ggml/test_ggml.py", + "left_ranges": [ + [ + 351, + 358 + ], + [ + 386, + 406 + ] + ], + "right_ranges": [ + [ + 351, + 358 + ], + [ + 386, + 405 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-45341-4", + "summary": "Cluster of 1 issues and 3 PRs centered on issue #45341.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45341, + "canonical_pr_number": 45351, + "issue_numbers": [ + 45341 + ], + "pr_numbers": [ + 45351, + 45509, + 45515 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45351, + "right_pr_number": 45509, + "code_similarity": 0.427, + "size_similarity": 0.308, + "file_overlap": 1.0, + "area_overlap": 0.148, + "patch_similarity": 0.093, + "shared_filenames": [ + "src/transformers/testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3207, + 3230 + ] + ], + "right_ranges": [ + [ + 3204, + 3210 + ] + ] + } + ] + }, + { + "left_pr_number": 45351, + "right_pr_number": 45515, + "code_similarity": 0.437, + "size_similarity": 0.359, + "file_overlap": 1.0, + "area_overlap": 0.148, + "patch_similarity": 0.091, + "shared_filenames": [ + "src/transformers/testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3207, + 3230 + ] + ], + "right_ranges": [ + [ + 3204, + 3210 + ] + ] + } + ] + }, + { + "left_pr_number": 45509, + "right_pr_number": 45515, + "code_similarity": 0.946, + "size_similarity": 0.857, + "file_overlap": 1.0, + "area_overlap": 1.0, + "patch_similarity": 0.833, + "shared_filenames": [ + "src/transformers/testing_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/testing_utils.py", + "left_ranges": [ + [ + 3204, + 3210 + ] + ], + "right_ranges": [ + [ + 3204, + 3210 + ] + ] + } + ] + } + ] + }, + { + "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-45081-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #45081.", + "status": "open", + "confidence": 0.8, + "canonical_issue_number": 45081, + "canonical_pr_number": 45317, + "issue_numbers": [ + 45081 + ], + "pr_numbers": [ + 45086, + 45317 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target", + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 45086, + "right_pr_number": 45317, + "code_similarity": 0.589, + "size_similarity": 0.34, + "file_overlap": 0.5, + "area_overlap": 1.0, + "patch_similarity": 0.136, + "shared_filenames": [ + "src/transformers/tokenization_utils_tokenizers.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/tokenization_utils_tokenizers.py", + "left_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ], + "right_ranges": [ + [ + 1360, + 1370 + ], + [ + 1374, + 1380 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-43698-3", + "summary": "Cluster of 1 issues and 2 PRs centered on issue #43698.", + "status": "open", + "confidence": 0.75, + "canonical_issue_number": 43698, + "canonical_pr_number": 43779, + "issue_numbers": [ + 43698 + ], + "pr_numbers": [ + 43779, + 43816 + ], + "evidence_types": [ + "closing_reference", + "shared_issue_target" + ], + "pr_comparisons": [ + { + "left_pr_number": 43779, + "right_pr_number": 43816, + "code_similarity": 0.418, + "size_similarity": 0.538, + "file_overlap": 1.0, + "area_overlap": 0.02, + "patch_similarity": 0.02, + "shared_filenames": [ + "src/transformers/integrations/integration_utils.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/integrations/integration_utils.py", + "left_ranges": [ + [ + 2227, + 2241 + ], + [ + 2303, + 2309 + ] + ], + "right_ranges": [ + [ + 2278, + 2291 + ], + [ + 2309, + 2322 + ] + ] + } + ] + } + ] + }, + { + "cluster_id": "cluster-44018-2", + "summary": "Cluster of 2 related pull requests linked by soft_similarity.", + "status": "open", + "confidence": 0.5, + "canonical_issue_number": null, + "canonical_pr_number": 44068, + "issue_numbers": [], + "pr_numbers": [ + 44018, + 44068 + ], + "evidence_types": [ + "soft_similarity" + ], + "pr_comparisons": [ + { + "left_pr_number": 44018, + "right_pr_number": 44068, + "code_similarity": 0.766, + "size_similarity": 0.939, + "file_overlap": 1.0, + "area_overlap": 0.425, + "patch_similarity": 0.866, + "shared_filenames": [ + "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "tests/models/gpt_neo/test_modeling_gpt_neo.py" + ], + "shared_file_areas": [ + { + "filename": "src/transformers/models/gpt_neo/modeling_gpt_neo.py", + "left_ranges": [ + [ + 26, + 46 + ], + [ + 138, + 143 + ], + [ + 182, + 187 + ], + [ + 283, + 295 + ], + [ + 327, + 341 + ], + [ + 348, + 354 + ], + [ + 360, + 369 + ], + [ + 399, + 406 + ], + [ + 411, + 419 + ], + [ + 428, + 433 + ], + [ + 472, + 492 + ], + [ + 507, + 513 + ], + [ + 519, + 528 + ], + [ + 541, + 559 + ], + [ + 562, + 567 + ], + [ + 595, + 601 + ], + [ + 607, + 614 + ], + [ + 627, + 643 + ], + [ + 685, + 690 + ], + [ + 708, + 714 + ], + [ + 720, + 727 + ], + [ + 740, + 757 + ], + [ + 761, + 766 + ], + [ + 780, + 786 + ], + [ + 791, + 798 + ], + [ + 807, + 822 + ], + [ + 840, + 845 + ] + ], + "right_ranges": [ + [ + 26, + 31 + ], + [ + 34, + 44 + ], + [ + 136, + 141 + ], + [ + 180, + 185 + ], + [ + 281, + 293 + ], + [ + 325, + 339 + ], + [ + 346, + 352 + ], + [ + 358, + 367 + ], + [ + 397, + 404 + ], + [ + 409, + 417 + ], + [ + 426, + 434 + ], + [ + 463, + 483 + ], + [ + 498, + 504 + ], + [ + 510, + 519 + ], + [ + 532, + 550 + ], + [ + 553, + 564 + ], + [ + 586, + 592 + ], + [ + 598, + 605 + ], + [ + 618, + 634 + ], + [ + 676, + 688 + ], + [ + 699, + 705 + ], + [ + 711, + 718 + ], + [ + 731, + 748 + ], + [ + 752, + 762 + ], + [ + 771, + 777 + ], + [ + 782, + 789 + ], + [ + 798, + 813 + ], + [ + 831, + 836 + ] + ] + }, + { + "filename": "tests/models/gpt_neo/test_modeling_gpt_neo.py", + "left_ranges": [ + [ + 458, + 464 + ] + ], + "right_ranges": [ + [ + 458, + 464 + ] + ] + } + ] + } + ] + } + ], + "duplicate_issues": [], + "duplicate_prs": [ + { + "cluster_id": "cluster-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, shared_issue_target 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-43366-4", + "canonical_pr_number": 45506, + "duplicate_pr_numbers": [ + 43757, + 45500 + ], + "target_issue_number": 43366, + "reason": "PRs in cluster-43366-4 are treated as duplicates because they converge on issue #43366 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-43656-4", + "canonical_pr_number": 43656, + "duplicate_pr_numbers": [ + 43836, + 43842 + ], + "target_issue_number": 43824, + "reason": "PRs in cluster-43656-4 are treated as duplicates because they converge on issue #43824 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-43698-3", + "canonical_pr_number": 43779, + "duplicate_pr_numbers": [ + 43816 + ], + "target_issue_number": 43698, + "reason": "PRs in cluster-43698-3 are treated as duplicates because they converge on issue #43698 with closing_reference, shared_issue_target evidence." + }, + { + "cluster_id": "cluster-43979-11", + "canonical_pr_number": 44007, + "duplicate_pr_numbers": [ + 43996, + 44013, + 44044, + 44066, + 44072, + 44085, + 44129, + 44154, + 44722 + ], + "target_issue_number": 43979, + "reason": "PRs in cluster-43979-11 are treated as duplicates because they converge on issue #43979 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-44018-2", + "canonical_pr_number": 44068, + "duplicate_pr_numbers": [ + 44018 + ], + "target_issue_number": null, + "reason": "PRs in cluster-44018-2 are treated as duplicates because they share soft_similarity evidence." + }, + { + "cluster_id": "cluster-45081-3", + "canonical_pr_number": 45317, + "duplicate_pr_numbers": [ + 45086 + ], + "target_issue_number": 45081, + "reason": "PRs in cluster-45081-3 are treated as duplicates because they converge on issue #45081 with closing_reference, shared_issue_target, soft_similarity evidence." + }, + { + "cluster_id": "cluster-45341-4", + "canonical_pr_number": 45351, + "duplicate_pr_numbers": [ + 45509, + 45515 + ], + "target_issue_number": 45341, + "reason": "PRs in cluster-45341-4 are treated as duplicates because they converge on issue #45341 with closing_reference, shared_issue_target, soft_similarity evidence." + } + ], + "best_issue": { + "cluster_id": "cluster-43366-4", + "issue_number": 43366, + "reason": "Issue #43366 is the strongest global issue candidate because it is open, belongs to a cluster with 4 artifacts, and carries 5 discussion comments plus 7 inbound references.", + "score": 157.0 + }, + "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/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.llm-reviews.json b/snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.llm-reviews.json new file mode 100644 index 0000000000000000000000000000000000000000..10e71a151296dc202e65a8dbaf38c5ac3971b560 --- /dev/null +++ b/snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.llm-reviews.json @@ -0,0 +1,17322 @@ +{ + "schema_version": "1.0", + "repo": "huggingface/transformers", + "snapshot_id": "20260420T094431Z", + "generated_at": "2026-04-20T09:57:23Z", + "model": "gpt-5.4-mini?service_tier=flex", + "reviews": [ + { + "label": "issue", + "component_index": 1, + "component_count": 13, + "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": "def90bc16496d6f679e8ab2e49a31cde60469c2df424e68eb6615a9ca14b9d26", + "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": "No duplicate cluster here: one issue is about Tapas tokenization, one about private-model auth/download messaging, and one about a generic model-loading failure. The only soft similarities are at the subsystem level, not the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": "issue #16618 is the broadest model-loading report in the set and is the closest representative of the only related pair; the other issues describe different symptoms and causes.", + "canonical_pr_reason": null, + "best_issue_reason": "#16618 is the most general and user-facing of the issues, so it is the best single representative if one must be chosen.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:13244", + "right": "issue:16618", + "accept": false, + "reason": "Tapas tokenization mismatch vs. a BERT model-loading failure are different bugs in different code paths; shared ML infrastructure is too broad to treat as duplicate." + }, + { + "left": "issue:15980", + "right": "issue:16618", + "accept": false, + "reason": "Both involve model download/load behavior, but #15980 is specifically an unauthorized private-model download error while #16618 is a generic 'can't load model' report for bert-base-uncased; 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": 2, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:24643", + "issue:30277", + "issue:34634", + "issue:34928", + "issue:35545", + "issue:36010", + "issue:36331", + "issue:39290", + "issue:41093", + "issue:41720", + "issue:41762", + "issue:42915", + "issue:43039", + "issue:43296", + "issue:43366", + "issue:43404", + "issue:43425", + "issue:43531", + "issue:43541", + "issue:43596", + "issue:43638", + "issue:43716", + "issue:43792", + "issue:43827", + "issue:43828", + "issue:43854", + "issue:43866", + "issue:43901", + "issue:43975", + "issue:44292", + "issue:44322", + "issue:44457", + "issue:44509", + "issue:44512", + "issue:44560", + "issue:44661", + "issue:44805", + "issue:44841", + "issue:44863", + "issue:44918", + "issue:45070", + "issue:45081", + "issue:45084", + "issue:45161", + "issue:45237", + "issue:45362", + "issue:45464", + "issue:45507" + ], + "soft_pairs": [ + "issue:44322|issue:45464", + "issue:43716|issue:45237", + "issue:43828|issue:45237", + "issue:45237|issue:45507", + "issue:44292|issue:45237", + "issue:30277|issue:35545", + "issue:35545|issue:42915", + "issue:35545|issue:41720", + "issue:44509|issue:44512", + "issue:43638|issue:44805", + "issue:35545|issue:36010", + "issue:43596|issue:44805", + "issue:44805|issue:45161", + "issue:44805|issue:44918", + "issue:44661|issue:44805", + "issue:30277|issue:43638", + "issue:34634|issue:35545", + "issue:43596|issue:43638", + "issue:43975|issue:44457", + "issue:43039|issue:45161", + "issue:44292|issue:44457", + "issue:34928|issue:44805", + "issue:24643|issue:30277", + "issue:43975|issue:45237", + "issue:43638|issue:45161", + "issue:34928|issue:36331", + "issue:43596|issue:45161", + "issue:43716|issue:45507", + "issue:39290|issue:45081", + "issue:43828|issue:45161", + "issue:43425|issue:44292", + "issue:34928|issue:41720", + "issue:44918|issue:45161", + "issue:43541|issue:43716", + "issue:43828|issue:44292", + "issue:43716|issue:44292", + "issue:44292|issue:44918", + "issue:43039|issue:44918", + "issue:41762|issue:43638", + "issue:43827|issue:44512", + "issue:43716|issue:44560", + "issue:43716|issue:44918", + "issue:43039|issue:43828", + "issue:30277|issue:36331", + "issue:43039|issue:44805", + "issue:34928|issue:36010", + "issue:41762|issue:43596", + "issue:43901|issue:44512", + "issue:39290|issue:41720", + "issue:43039|issue:44560", + "issue:43866|issue:44863", + "issue:43854|issue:43866", + "issue:43425|issue:45070", + "issue:39290|issue:43531", + "issue:41093|issue:43638", + "issue:43404|issue:43866", + "issue:45362|issue:45464", + "issue:43792|issue:45070", + "issue:43541|issue:43828", + "issue:36010|issue:39290", + "issue:44841|issue:45084", + "issue:41093|issue:41762", + "issue:43296|issue:43366" + ], + "prepared_review_unit_hash": "94e914bca7ac44e87e08ddaf6a3ada324e9127ca80320efe843f8e204003e97c", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 63, + "serialized_chars": 46472, + "estimated_input_tokens": 11618, + "estimated_eval_tokens": 23492 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is mostly a grab bag of unrelated model/runtime/docs issues. The only clear duplicate family is the DeepSpeed ZeRO-3 zero-size/index-out-of-bounds loading bug around Bert/Gemma3; the rest are distinct failures across different models, APIs, and doc cleanups.", + "confidence": 0.9, + "canonical_issue_reason": "issue:43596 is the best representative of the one real duplicate family here: it states the concrete ZeRO-3 zero-size index-out-of-bounds failure clearly, and it has a near-exact duplicate (43638) plus a closely related follow-up (41762) on the same loading path.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43596 is the strongest global issue representative because it captures the shared low-level DeepSpeed ZeRO-3 loading bug in a model-agnostic way while still being specific enough to anchor the small duplicate cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44322", + "right": "issue:45464", + "accept": false, + "reason": "Different failures: Qwen3.5 config attribute access vs chat/completions streaming behavior on a different path." + }, + { + "left": "issue:43716", + "right": "issue:45237", + "accept": false, + "reason": "Both involve model/runtime problems, but one is an autocast dtype mismatch and the other is an AMD GPU execution failure for GPT-OSS." + }, + { + "left": "issue:43828", + "right": "issue:45237", + "accept": false, + "reason": "Different concrete bugs: Phi-tiny-MoE autocast dtype mismatch vs GPT-OSS AMD GPU failure." + }, + { + "left": "issue:45237", + "right": "issue:45507", + "accept": false, + "reason": "Different model/code paths: GPT-OSS on AMD GPUs vs GraniteMoEHybrid calling an invalid method." + }, + { + "left": "issue:44292", + "right": "issue:45237", + "accept": false, + "reason": "Different models and symptoms: Qwen-3-8B-NVFP4 runtime error vs GPT-OSS AMD GPU failure." + }, + { + "left": "issue:30277", + "right": "issue:35545", + "accept": false, + "reason": "DeepSpeed collective mismatch on Jamba-v01 is unrelated to ModernBERT ONNX export." + }, + { + "left": "issue:35545", + "right": "issue:42915", + "accept": false, + "reason": "Different bugs in different models: ONNX export failure vs FineGrainedFP8Config load failure." + }, + { + "left": "issue:35545", + "right": "issue:41720", + "accept": false, + "reason": "ModernBERT ONNX export and Qwen3 auto-device-map cuda assert are unrelated." + }, + { + "left": "issue:44509", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs cleanup around v5 removals, but they concern different removed topics and are not the same concrete change." + }, + { + "left": "issue:43638", + "right": "issue:44805", + "accept": false, + "reason": "Both mention index/shape problems, but one is a ZeRO-3 Bert loading failure and the other is a different mask/tensor mismatch symptom." + }, + { + "left": "issue:35545", + "right": "issue:36010", + "accept": false, + "reason": "ONNX export failure and GenerationMixin import error are unrelated." + }, + { + "left": "issue:43596", + "right": "issue:44805", + "accept": false, + "reason": "Similar index/shape language, but the specific error and code path differ; not enough evidence for the same bug." + }, + { + "left": "issue:44805", + "right": "issue:45161", + "accept": false, + "reason": "Mask/tensor shape mismatch is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:44661", + "right": "issue:44805", + "accept": false, + "reason": "Tokenizer-mapping issue and mask/tensor mismatch are different problems." + }, + { + "left": "issue:30277", + "right": "issue:43638", + "accept": false, + "reason": "Both touch DeepSpeed, but one is a collective mismatch and the other is a zero-size index error; different bugs." + }, + { + "left": "issue:34634", + "right": "issue:35545", + "accept": false, + "reason": "BarkProcessor voice preset regression is unrelated to ModernBERT export." + }, + { + "left": "issue:43596", + "right": "issue:43638", + "accept": true, + "reason": "Same concrete ZeRO-3 zero-size index-out-of-bounds loading bug, with nearly identical symptom and code path." + }, + { + "left": "issue:43975", + "right": "issue:44457", + "accept": false, + "reason": "Different issues: incorrect detokenization vs LoRA merge/save/reload output mismatch." + }, + { + "left": "issue:43039", + "right": "issue:45161", + "accept": false, + "reason": "Liger Kernel cross_entropy invocation and GPT-OSS tensor-parallel failure are different code paths." + }, + { + "left": "issue:44292", + "right": "issue:44457", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and LoRA merge/reload mismatch are unrelated." + }, + { + "left": "issue:34928", + "right": "issue:44805", + "accept": false, + "reason": "Both are tensor-shape failures, but the contexts and symptoms differ: activation checkpointing/FSDP vs mask indexing." + }, + { + "left": "issue:24643", + "right": "issue:30277", + "accept": false, + "reason": "Different DeepSpeed-related failures: weight dimensionality check vs collective mismatch." + }, + { + "left": "issue:43975", + "right": "issue:45237", + "accept": false, + "reason": "Detokenization bug is unrelated to GPT-OSS AMD GPU loading/runtime failure." + }, + { + "left": "issue:43638", + "right": "issue:45161", + "accept": false, + "reason": "ZeRO-3 Bert loading failure is unrelated to GPT-OSS tensor-parallel support." + }, + { + "left": "issue:34928", + "right": "issue:36331", + "accept": false, + "reason": "Activation checkpointing/FSDP tensor mismatch and CustomTrainer kwarg mismatch are unrelated." + }, + { + "left": "issue:43716", + "right": "issue:45507", + "accept": false, + "reason": "Different failures: image preprocessor dtype mismatch vs GraniteMoEHybrid invalid method call." + }, + { + "left": "issue:39290", + "right": "issue:45081", + "accept": false, + "reason": "Gemma3 config attribute missing in vLLM and Mistral regex patch backend_tokenizer error are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:45161", + "accept": false, + "reason": "Autocast dtype mismatch on Phi-tiny-MoE is unrelated to GPT-OSS tensor-parallel failure." + }, + { + "left": "issue:44292", + "right": "issue:44918", + "accept": false, + "reason": "Both mention Qwen3-family models, but one is an NVFP4 runtime error and the other is a TRL SFT trainer embedding-unpacking failure." + }, + { + "left": "issue:43039", + "right": "issue:43828", + "accept": false, + "reason": "Different execution bugs: Liger Kernel cross_entropy path vs autocast dtype mismatch." + }, + { + "left": "issue:30277", + "right": "issue:36331", + "accept": false, + "reason": "DeepSpeed collective mismatch and CustomTrainer compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:43039", + "right": "issue:44805", + "accept": false, + "reason": "Liger Kernel cross_entropy issue is unrelated to mask/tensor shape mismatch." + }, + { + "left": "issue:34928", + "right": "issue:36010", + "accept": false, + "reason": "FSDP activation checkpointing mismatch and GenerationMixin import error are unrelated." + }, + { + "left": "issue:41762", + "right": "issue:43596", + "accept": true, + "reason": "Same underlying DeepSpeed ZeRO-3 zero-size index-out-of-bounds loading failure, just reported on a different model." + }, + { + "left": "issue:43827", + "right": "issue:44512", + "accept": false, + "reason": "Both are docs issues caused by v5 removals, but they reference different removed features and should not be merged as the same change." + }, + { + "left": "issue:43901", + "right": "issue:44512", + "accept": false, + "reason": "TextClassificationPipeline docs mismatch and transformers run-command docs removal are separate documentation fixes." + }, + { + "left": "issue:39290", + "right": "issue:41720", + "accept": false, + "reason": "Different bugs: Gemma3 sliding-window config error vs Qwen3 auto device mapping cuda assert." + }, + { + "left": "issue:43039", + "right": "issue:44560", + "accept": false, + "reason": "Liger Kernel cross_entropy regression is unrelated to Qwen3-vl video StopIteration." + }, + { + "left": "issue:43866", + "right": "issue:44863", + "accept": false, + "reason": "Different model-loading failures: corrupted Ovis2 checkpoint vs NemotronH implementation unable to load checkpoints." + }, + { + "left": "issue:43854", + "right": "issue:43866", + "accept": false, + "reason": "Different model/test failures: GLM-4.7-Flash unit test loading issue vs Ovis2 checkpoint corruption." + }, + { + "left": "issue:45362", + "right": "issue:45464", + "accept": false, + "reason": "Both involve Qwen3.5 chat use, but one is a transformers chat crash and the other is a streaming chat/completions API failure; not the same bug." + }, + { + "left": "issue:43792", + "right": "issue:45070", + "accept": false, + "reason": "Whisper model runtime failure is unrelated to the pydantic PretrainedConfig field breakage." + }, + { + "left": "issue:43541", + "right": "issue:43828", + "accept": false, + "reason": "Grouped_mm torch-dynamo tracing failure and autocast dtype mismatch are different execution problems." + }, + { + "left": "issue:36010", + "right": "issue:39290", + "accept": false, + "reason": "GenerationMixin import error and Gemma3 sliding_window_pattern error are unrelated." + }, + { + "left": "issue:44841", + "right": "issue:45084", + "accept": false, + "reason": "Processor failure for Voxtral-Mini-3B-2507 is unrelated to the non-template-node compile error." + }, + { + "left": "issue:41093", + "right": "issue:41762", + "accept": false, + "reason": "Mask/index shape mismatch during one path is not the same as the ZeRO-3 zero-size index-out-of-bounds loading error." + }, + { + "left": "issue:43404", + "right": "issue:43866", + "accept": false, + "reason": "Mistral3 lm_head tying bug is unrelated to the Ovis2 checkpoint corruption report." + }, + { + "left": "issue:41093", + "right": "issue:43638", + "accept": false, + "reason": "Mask/tensor shape mismatch and ZeRO-3 Bert index-out-of-bounds are different bugs." + }, + { + "left": "issue:43425", + "right": "issue:45070", + "accept": false, + "reason": "Torch 2.10 incompatibility is unrelated to the pydantic PretrainedConfig regression." + }, + { + "left": "issue:39290", + "right": "issue:43531", + "accept": false, + "reason": "Different Qwen/Gemma sliding-window problems with different model families and symptoms." + }, + { + "left": "issue:43296", + "right": "issue:43366", + "accept": false, + "reason": "PaddleOCR-VL loading failure and GGUF gpt-oss support request are unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is well grounded in the packet and stays conservative on the soft-edge calls. The only clearly supported duplicate family is the DeepSpeed ZeRO-3 index-out-of-bounds loading bug cluster (43596/43638/41762), and the other pairings are appropriately rejected as distinct issues." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 2, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:36331", + "issue:39186", + "issue:39290", + "issue:39692", + "issue:40990", + "issue:41093", + "issue:41553", + "issue:41762", + "issue:42915", + "issue:43054", + "issue:43366", + "issue:43404", + "issue:43531", + "issue:43541", + "issue:43572", + "issue:43596", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43828", + "issue:43866", + "issue:43873", + "issue:43883", + "issue:43950", + "issue:43975", + "issue:43994", + "issue:44291", + "issue:44292", + "issue:44360", + "issue:44387", + "issue:44403", + "issue:44492", + "issue:44512", + "issue:44530", + "issue:44534", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44863", + "issue:44918", + "issue:44928", + "issue:44960", + "issue:45005", + "issue:45070", + "issue:45161", + "issue:45399", + "issue:45464", + "issue:45507" + ], + "soft_pairs": [ + "issue:43572|issue:45070", + "issue:44387|issue:45005", + "issue:43541|issue:45161", + "issue:43828|issue:45464", + "issue:44291|issue:45070", + "issue:43366|issue:45464", + "issue:43531|issue:45070", + "issue:44661|issue:45464", + "issue:43541|issue:44589", + "issue:39290|issue:43643", + "issue:43054|issue:43994", + "issue:41093|issue:43596", + "issue:42915|issue:43541", + "issue:40990|issue:44960", + "issue:39186|issue:44918", + "issue:43404|issue:45005", + "issue:43645|issue:45399", + "issue:44530|issue:45005", + "issue:44863|issue:45005", + "issue:43873|issue:45005", + "issue:42915|issue:45399", + "issue:44360|issue:44512", + "issue:44918|issue:45507", + "issue:43975|issue:44292", + "issue:43950|issue:44534", + "issue:44492|issue:44512", + "issue:43644|issue:43950", + "issue:44292|issue:45507", + "issue:43828|issue:44928", + "issue:39692|issue:43054", + "issue:43866|issue:44403", + "issue:41553|issue:43883", + "issue:44560|issue:45507", + "issue:36331|issue:41762" + ], + "prepared_review_unit_hash": "ea18976fa602f7739bdcee183c749da69a271efded6bc9d799f1a02715b00976", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37619, + "estimated_input_tokens": 9405, + "estimated_eval_tokens": 19066 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Two clear duplicates in this set: the non-persistent buffer regression around `from_pretrained()` / Transformers v5, and the broader duplicate wording of that same issue. The rest are same-surface-area only and not the same concrete bug.", + "confidence": 0.94, + "canonical_issue_reason": "issue:43950 is the clearest representative of the non-persistent-buffer regression: it names the failing path (`from_pretrained`) and the specific corruption of `register_buffer(persistent=False)`.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43950 best captures the underlying bug with enough detail to anchor the cluster; issue:43644 is the same regression but phrased more generally.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43572", + "right": "issue:45070", + "accept": false, + "reason": "Both mention config/field problems after v5, but one is `pad_token_idx` in StableLmConfig and the other is a pydantic `PretrainedConfig` field issue; different bugs." + }, + { + "left": "issue:44387", + "right": "issue:45005", + "accept": false, + "reason": "Both touch v5/weights/quantization themes, but one is an int4 OOM/memory regression and the other is tied weights in translation models." + }, + { + "left": "issue:43541", + "right": "issue:45161", + "accept": false, + "reason": "Both involve MoE/inference, but the concrete failures differ: grouped_mm tracing error vs tensor-parallel support not working for GPT-OSS." + }, + { + "left": "issue:43828", + "right": "issue:45464", + "accept": false, + "reason": "Different models and failure modes: dtype mismatch under autocast vs chat/completions streaming failure." + }, + { + "left": "issue:44291", + "right": "issue:45070", + "accept": false, + "reason": "`init_empty_weights` unexpected argument error is unrelated to the pydantic `PretrainedConfig` field regression." + }, + { + "left": "issue:43366", + "right": "issue:45464", + "accept": false, + "reason": "GPT-OSS GGUF support request is unrelated to Qwen3.5 streaming inference failures." + }, + { + "left": "issue:43531", + "right": "issue:45070", + "accept": false, + "reason": "Sliding-window behavior in Qwen3-MoE is a different model-specific bug from the config-field/pydantic issue." + }, + { + "left": "issue:44661", + "right": "issue:45464", + "accept": false, + "reason": "Tokenizer mapping failure when adding a model is unrelated to Qwen3.5 streaming inference." + }, + { + "left": "issue:43541", + "right": "issue:44589", + "accept": false, + "reason": "Grouped_mm tracing failure and missing Float8 storage type are different bugs." + }, + { + "left": "issue:39290", + "right": "issue:43643", + "accept": false, + "reason": "Both are config-field issues, but one is Gemma3/vLLM missing `sliding_window_pattern` and the other is `trust_remote_code` returning incomplete config objects." + }, + { + "left": "issue:43054", + "right": "issue:43994", + "accept": false, + "reason": "Both are SigLIP2-related, but one reports poor text embeddings and the other nonsensical AutoModel/pipeline results; not the same concrete defect." + }, + { + "left": "issue:41093", + "right": "issue:43596", + "accept": false, + "reason": "Both are index errors around loading, but the shapes and code paths differ (`mask` mismatch vs zero-sized index during ZeRO-3/BertModel load)." + }, + { + "left": "issue:42915", + "right": "issue:43541", + "accept": false, + "reason": "Different issues: FineGrainedFP8Config failure for Qwen3-MoE vs float32 grouped_mm tracing failure for Mixtral." + }, + { + "left": "issue:40990", + "right": "issue:44960", + "accept": false, + "reason": "Unrelated model/perplexity report versus a GLM5 issue with no concrete overlap." + }, + { + "left": "issue:39186", + "right": "issue:44918", + "accept": false, + "reason": "FSDP 2-D weight error and TRL SFT unpacking failure are distinct training/runtime bugs." + }, + { + "left": "issue:43404", + "right": "issue:45005", + "accept": false, + "reason": "Both mention tied weights, but one is a specific Mistral3 lm_head tie bug and the other is a broader translation-model v5 issue." + }, + { + "left": "issue:43645", + "right": "issue:45399", + "accept": false, + "reason": "Custom model initialization in Jupyter is unrelated to the flash-attn2 fallback/check ordering problem." + }, + { + "left": "issue:44530", + "right": "issue:45005", + "accept": false, + "reason": "PagedAttentionCache/group-type crash on Qwen3.5 is unrelated to tied-weight translation-model regressions." + }, + { + "left": "issue:44863", + "right": "issue:45005", + "accept": false, + "reason": "Checkpoint-loading failure for NemotronH is not the same bug as tied weights on translation models." + }, + { + "left": "issue:43873", + "right": "issue:45005", + "accept": false, + "reason": "Quantization/offloading behavior and tied-weight handling are different issues." + }, + { + "left": "issue:42915", + "right": "issue:45399", + "accept": false, + "reason": "Qwen3-MoE FP8 loading and flash-attn2 fallback gating are unrelated." + }, + { + "left": "issue:44360", + "right": "issue:44512", + "accept": false, + "reason": "A DSA indexer ReLU discussion is not the same as docs still mentioning the removed run command." + }, + { + "left": "issue:44918", + "right": "issue:45507", + "accept": false, + "reason": "TRL input-embedding unpacking and GraniteMoEHybrid invalid method calls are different model/integration bugs." + }, + { + "left": "issue:43975", + "right": "issue:44292", + "accept": false, + "reason": "Detokenization bug in deepseek-coder v5 is unrelated to Qwen-3-8B-NVFP4 runtime errors." + }, + { + "left": "issue:43950", + "right": "issue:44534", + "accept": true, + "reason": "Same regression: Transformers v5 corrupts non-persistent buffers; one title names the concrete `from_pretrained()` path, the other states the same junk-buffer symptom." + }, + { + "left": "issue:44492", + "right": "issue:44512", + "accept": false, + "reason": "A typo in cache strategies is unrelated to docs referencing the removed run command." + }, + { + "left": "issue:43644", + "right": "issue:43950", + "accept": true, + "reason": "Same underlying bug: v5 incorrectly fills/corrupts non-persistent buffers. `43950` is the clearer, more specific report." + }, + { + "left": "issue:44292", + "right": "issue:45507", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime failure and GraniteMoEHybrid invalid method calls are unrelated." + }, + { + "left": "issue:43828", + "right": "issue:44928", + "accept": false, + "reason": "Different problems in different models: autocast dtype mismatch vs Qwen3.5 RLHF gradient explosion/SDPA fallback." + }, + { + "left": "issue:39692", + "right": "issue:43054", + "accept": false, + "reason": "SigLIP2 docs example errors (model/processor mismatch and quantization failure) are not the same as a text-embedding quality regression." + }, + { + "left": "issue:43866", + "right": "issue:44403", + "accept": false, + "reason": "Corrupted Ovis2 checkpoint and generic loading noise are unrelated." + }, + { + "left": "issue:41553", + "right": "issue:43883", + "accept": false, + "reason": "Bad AutoTokenizer error message for Voxtral and missing `all_tied_weights_keys` on MolmoForCausalLM are different issues." + }, + { + "left": "issue:44560", + "right": "issue:45507", + "accept": false, + "reason": "Qwen3-vl video StopIteration and GraniteMoEHybrid invalid method calls are unrelated." + }, + { + "left": "issue:36331", + "right": "issue:41762", + "accept": false, + "reason": "CustomTrainer `compute_loss` kwarg mismatch and Gemma3 ZeRO-3 loading index error are distinct bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the two accepted soft-edge pairs are conservative and well-supported by the issue titles: 43950/44534 and 43644/43950 clearly describe the same non-persistent-buffer regression. The other verdicts stay cautious and do not overclaim duplicates." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 3, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:36331", + "issue:38175", + "issue:41093", + "issue:41553", + "issue:41762", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43054", + "issue:43257", + "issue:43278", + "issue:43404", + "issue:43493", + "issue:43643", + "issue:43742", + "issue:43782", + "issue:43792", + "issue:43824", + "issue:43827", + "issue:43854", + "issue:43856", + "issue:43866", + "issue:43872", + "issue:43881", + "issue:43883", + "issue:43940", + "issue:43950", + "issue:43975", + "issue:44220", + "issue:44291", + "issue:44368", + "issue:44387", + "issue:44403", + "issue:44451", + "issue:44488", + "issue:44509", + "issue:44589", + "issue:44661", + "issue:44863", + "issue:44960", + "issue:44991", + "issue:45020", + "issue:45081", + "issue:45305", + "issue:45313", + "issue:45356", + "issue:45399", + "issue:45520" + ], + "soft_pairs": [ + "issue:43872|issue:44589", + "issue:41553|issue:43881", + "issue:38175|issue:43054", + "issue:42947|issue:44387", + "issue:43054|issue:43493", + "issue:43881|issue:43883", + "issue:44403|issue:44451", + "issue:42947|issue:43856", + "issue:43975|issue:45520", + "issue:43643|issue:45020", + "issue:44403|issue:44488", + "issue:43782|issue:44960", + "issue:42491|issue:44368", + "issue:41762|issue:42915", + "issue:43975|issue:44451", + "issue:45020|issue:45356", + "issue:44661|issue:45020", + "issue:43824|issue:43872", + "issue:43054|issue:43278", + "issue:44991|issue:45399", + "issue:43792|issue:44220", + "issue:43827|issue:44509", + "issue:45081|issue:45399", + "issue:43404|issue:43950", + "issue:42915|issue:45020", + "issue:43257|issue:43866", + "issue:43742|issue:45399", + "issue:43742|issue:45020", + "issue:41553|issue:44991", + "issue:43940|issue:45313", + "issue:43872|issue:44291", + "issue:43854|issue:44863", + "issue:42915|issue:43872", + "issue:41553|issue:44488", + "issue:43257|issue:45305", + "issue:36331|issue:41093" + ], + "prepared_review_unit_hash": "19323840d6f22a77861d4ba4e10d79dcbed2906cad12b132531e2c79ec9c9edb", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38577, + "estimated_input_tokens": 9645, + "estimated_eval_tokens": 19546 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly heterogeneous issues with a few true duplicates. Clear duplicate pairs are the two `_is_hf_initialized` loading errors and the two v5 docs pipeline-removal issues; one Whisper audio-loading pair also looks like the same underlying bug. Most other soft pairs are only loosely related by model family or subsystem and should stay separate.", + "confidence": 0.81, + "canonical_issue_reason": "No single issue cleanly represents the whole set because it spans unrelated bugs. If an umbrella must be chosen, issue 45020 is the broadest current regression around remote_code-based model loading, but it does not cover most of the other items.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45020 is the best global issue candidate only as a broad remote_code-loading regression anchor; otherwise the cluster is too mixed for a true canonical issue.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43872", + "right": "issue:44589", + "accept": false, + "reason": "Both are TypeErrors during loading, but they involve different failure points: `_is_hf_initialized` vs missing `Float8_e4m3fnStorage`." + }, + { + "left": "issue:41553", + "right": "issue:43881", + "accept": false, + "reason": "Different bugs: a misleading AutoTokenizer error for Voxtral vs a concrete load failure for GLM-4v-9b." + }, + { + "left": "issue:38175", + "right": "issue:43054", + "accept": false, + "reason": "Same model family, but one is zero probabilities and the other is worse text embeddings; not the same concrete bug." + }, + { + "left": "issue:42947", + "right": "issue:44387", + "accept": false, + "reason": "Different training problems: LoRA gradient checkpointing vs int4 quantization memory growth." + }, + { + "left": "issue:43054", + "right": "issue:43493", + "accept": false, + "reason": "Both concern SigLIP2, but one is a quality regression and the other is an HF-vs-JAX implementation discrepancy." + }, + { + "left": "issue:43881", + "right": "issue:43883", + "accept": false, + "reason": "Unrelated load failures in different models and code paths." + }, + { + "left": "issue:44403", + "right": "issue:44451", + "accept": false, + "reason": "Docs/noise complaint vs a model-loading failure for ScandiBERT." + }, + { + "left": "issue:42947", + "right": "issue:43856", + "accept": false, + "reason": "Both involve training efficiency, but they are different issues: LoRA checkpointing vs Qwen3 MoE memory usage." + }, + { + "left": "issue:43975", + "right": "issue:45520", + "accept": false, + "reason": "Tokenizer detokenization regression is unrelated to the Python 3.13 `flash_attn` import error." + }, + { + "left": "issue:43643", + "right": "issue:45020", + "accept": false, + "reason": "Related to remote_code, but one is a specific missing-fields bug in AutoConfig and the other is a broader regression claim." + }, + { + "left": "issue:44403", + "right": "issue:44488", + "accept": false, + "reason": "Different symptoms and targets; one is docs/noise, the other is a model load failure." + }, + { + "left": "issue:43782", + "right": "issue:44960", + "accept": false, + "reason": "Different models and failure modes." + }, + { + "left": "issue:42491", + "right": "issue:44368", + "accept": false, + "reason": "Different problems: hf4.x/hf5.x LoRA compatibility vs a tie_word_embeddings warning." + }, + { + "left": "issue:41762", + "right": "issue:42915", + "accept": false, + "reason": "Different model families and different DeepSpeed/FP8 failure modes." + }, + { + "left": "issue:43975", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer detokenization bug vs unrelated model loading failure." + }, + { + "left": "issue:45020", + "right": "issue:45356", + "accept": false, + "reason": "Remote_code regression vs Kimi tokenizer codec/fix_mistral_regex regression; not the same bug." + }, + { + "left": "issue:44661", + "right": "issue:45020", + "accept": false, + "reason": "One is a new-model-like script edge case, the other is a broad remote_code loading regression." + }, + { + "left": "issue:43824", + "right": "issue:43872", + "accept": false, + "reason": "ImportError for a missing class vs a TypeError in bitsandbytes initialization." + }, + { + "left": "issue:43054", + "right": "issue:43278", + "accept": false, + "reason": "Both mention embeddings, but one is SigLIP2 quality and the other is a dtype change during evaluation." + }, + { + "left": "issue:44991", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer loading regression vs flash-attn fallback logic." + }, + { + "left": "issue:43792", + "right": "issue:44220", + "accept": true, + "reason": "Likely the same Whisper audio feature-extraction bug; the generic 'can't run' report matches the `_torch_extract_fbank_features()` failure." + }, + { + "left": "issue:43827", + "right": "issue:44509", + "accept": true, + "reason": "Same docs gap: both say removed v5 pipeline tasks are still referenced." + }, + { + "left": "issue:45081", + "right": "issue:45399", + "accept": false, + "reason": "Tokenizer regex crash vs flash-attn fallback being blocked are different code paths." + }, + { + "left": "issue:43404", + "right": "issue:43950", + "accept": false, + "reason": "Tied lm_head weights vs non-persistent buffer corruption are unrelated issues." + }, + { + "left": "issue:42915", + "right": "issue:45020", + "accept": false, + "reason": "Different Qwen3/DeepSpeed/FP8 issue vs broad remote_code loading regression." + }, + { + "left": "issue:43257", + "right": "issue:43866", + "accept": false, + "reason": "Different Qwen3 MoE problems: weight conversion/loading vs a corrupted Ovis2 checkpoint." + }, + { + "left": "issue:43742", + "right": "issue:45399", + "accept": false, + "reason": "MobileLLM key error is unrelated to flash-attn fallback blocking." + }, + { + "left": "issue:43742", + "right": "issue:45020", + "accept": false, + "reason": "A specific model-key error, not the broader remote_code regression." + }, + { + "left": "issue:41553", + "right": "issue:44991", + "accept": false, + "reason": "Different tokenizer-related failures in different models and code paths." + }, + { + "left": "issue:43940", + "right": "issue:45313", + "accept": false, + "reason": "Both are ZeRO-3 weight-loading failures, but for different Qwen model variants with different missing-weight patterns." + }, + { + "left": "issue:43872", + "right": "issue:44291", + "accept": true, + "reason": "Same underlying bug: loading with `init_empty_weights`/bitsandbytes trips on unexpected `_is_hf_initialized` handling." + }, + { + "left": "issue:43854", + "right": "issue:44863", + "accept": false, + "reason": "Different models and failures: GLM-4.7-Flash test loading vs NemotronH checkpoint loading." + }, + { + "left": "issue:36331", + "right": "issue:41093", + "accept": false, + "reason": "Unrelated errors: custom trainer loss signature vs a mask shape mismatch." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The clustering is grounded in the packet and the verdicts stay conservative. The accepted pairs are well-supported by the titles and issue descriptions: the two `_is_hf_initialized`/`init_empty_weights` TypeErrors, the two v5 docs pipeline-removal issues, and the Whisper audio-feature extraction reports. The remaining soft pairs are appropriately rejected as only loosely related by subsystem or model family." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 4, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:24643", + "issue:30064", + "issue:36010", + "issue:38175", + "issue:39692", + "issue:41093", + "issue:42491", + "issue:42617", + "issue:42915", + "issue:43317", + "issue:43366", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43638", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43720", + "issue:43824", + "issue:43828", + "issue:43872", + "issue:43950", + "issue:43957", + "issue:43994", + "issue:44075", + "issue:44451", + "issue:44479", + "issue:44488", + "issue:44530", + "issue:44534", + "issue:44560", + "issue:44589", + "issue:44661", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44877", + "issue:44910", + "issue:44918", + "issue:44960", + "issue:44977", + "issue:45030", + "issue:45084", + "issue:45313", + "issue:45325", + "issue:45357" + ], + "soft_pairs": [ + "issue:41093|issue:44805", + "issue:44479|issue:44560", + "issue:44530|issue:44863", + "issue:43872|issue:45084", + "issue:43366|issue:43828", + "issue:24643|issue:30064", + "issue:43366|issue:43531", + "issue:43550|issue:44910", + "issue:43645|issue:43646", + "issue:43950|issue:45325", + "issue:43366|issue:44589", + "issue:44560|issue:44918", + "issue:43824|issue:43957", + "issue:42915|issue:43957", + "issue:42617|issue:43366", + "issue:38175|issue:43994", + "issue:43646|issue:43950", + "issue:44560|issue:44805", + "issue:36010|issue:42915", + "issue:43638|issue:44661", + "issue:43596|issue:44661", + "issue:44877|issue:45030", + "issue:43317|issue:44756", + "issue:43644|issue:44534", + "issue:38175|issue:39692", + "issue:45313|issue:45357", + "issue:42491|issue:43720", + "issue:44451|issue:44488", + "issue:44960|issue:44977", + "issue:43708|issue:44075" + ], + "prepared_review_unit_hash": "54d380f1f9e6dded91bfc4ac0da3a645400ef84064f76bb1688f82de61c6b480", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36649, + "estimated_input_tokens": 9163, + "estimated_eval_tokens": 18582 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is not one coherent cluster; it breaks into a few small duplicate pairs and many unrelated issues. The clearest duplicates are around the mask-shape regression, custom model initialization in Transformers 5.0.0, SigLIP2 output corruption, and non-persistent buffer corruption.", + "confidence": 0.82, + "canonical_issue_reason": "No single canonical issue fits the whole set because the items span several unrelated bugs; the best-defined duplicate subclusters are 41093/44805, 43645/43646, 38175/43994, and 43644/44534.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue among this heterogeneous set; 43644 is a strong representative for the non-persistent-buffer regression, but it does not cover the other duplicate subclusters.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41093", + "right": "issue:44805", + "accept": true, + "reason": "Same mask-shape mismatch regression; 44805 is the same failure pattern with the same tensor-size mismatch." + }, + { + "left": "issue:44479", + "right": "issue:44560", + "accept": false, + "reason": "Both are video-related, but they describe different model families and different failure modes; not clearly the same concrete bug." + }, + { + "left": "issue:44530", + "right": "issue:44863", + "accept": false, + "reason": "Qwen3.5 cache/linear_attention crash vs NemotronH checkpoint loading are unrelated." + }, + { + "left": "issue:43872", + "right": "issue:45084", + "accept": false, + "reason": "bitsandbytes initialization incompatibility and template-node compile error are different code paths." + }, + { + "left": "issue:43366", + "right": "issue:43828", + "accept": false, + "reason": "GGUF gpt-oss support and Phi-tiny-MoE autocast dtype mismatch are unrelated." + }, + { + "left": "issue:24643", + "right": "issue:30064", + "accept": false, + "reason": "DeepSpeed weight-dimension error and void segmentation-map processing failure are unrelated." + }, + { + "left": "issue:43366", + "right": "issue:43531", + "accept": false, + "reason": "gpt-oss GGUF support and Qwen3-MoE sliding_window bug are different issues." + }, + { + "left": "issue:43550", + "right": "issue:44910", + "accept": false, + "reason": "Different models and different failure mechanisms: SDPA/torch.compile vs flash-attention position_ids handling." + }, + { + "left": "issue:43645", + "right": "issue:43646", + "accept": true, + "reason": "Both report Transformers 5.0.0 breaking custom model initialization; the notebook-specific report is a variant of the same regression." + }, + { + "left": "issue:43950", + "right": "issue:45325", + "accept": false, + "reason": "Non-persistent buffer corruption and Qwen2.5-VL rope-index scaling are unrelated." + }, + { + "left": "issue:43366", + "right": "issue:44589", + "accept": false, + "reason": "GGUF architecture support and missing Float8 storage type are unrelated." + }, + { + "left": "issue:44560", + "right": "issue:44918", + "accept": false, + "reason": "Video StopIteration in Qwen3-VL embedding and TRL SFT input-embedding unpacking are different bugs." + }, + { + "left": "issue:43824", + "right": "issue:43957", + "accept": false, + "reason": "Importing Qwen2_5_VL class and meta-device model loading failures are unrelated." + }, + { + "left": "issue:42915", + "right": "issue:43957", + "accept": false, + "reason": "Qwen3Moe FP8 config failure and meta-device loading regressions are different problems." + }, + { + "left": "issue:42617", + "right": "issue:43366", + "accept": false, + "reason": "3d_parallel.py execution failure and GGUF gpt-oss support are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:43994", + "accept": true, + "reason": "Same SigLIP2 checkpoint/output regression; the second report rephrases the same bad model behavior in AutoModel/pipeline terms." + }, + { + "left": "issue:43646", + "right": "issue:43950", + "accept": false, + "reason": "Custom model initialization regression and non-persistent buffer corruption are both Transformers 5 issues but not the same bug." + }, + { + "left": "issue:44560", + "right": "issue:44805", + "accept": false, + "reason": "Different failures: video StopIteration versus token/mask shape mismatch." + }, + { + "left": "issue:36010", + "right": "issue:42915", + "accept": false, + "reason": "GenerationMixin import error and Qwen3Moe FP8 failure are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44661", + "accept": false, + "reason": "DeepSpeed ZeRO-3 Bert index error and TOKENIZER_MAPPING_NAMES add-new-model-like failure are different code paths." + }, + { + "left": "issue:43596", + "right": "issue:44661", + "accept": false, + "reason": "These are distinct issues: ZeRO-3/Bert initialization versus model registration in tokenizer mapping." + }, + { + "left": "issue:44877", + "right": "issue:45030", + "accept": false, + "reason": "Both involve config loading/validation, but they concern different model families and different validation failures." + }, + { + "left": "issue:43317", + "right": "issue:44756", + "accept": false, + "reason": "Dequantized-model offload failure and mmap OOM workaround are not the same bug." + }, + { + "left": "issue:43644", + "right": "issue:44534", + "accept": true, + "reason": "Both describe the same Transformers v5 regression where non-persistent buffers get filled with junk." + }, + { + "left": "issue:38175", + "right": "issue:39692", + "accept": false, + "reason": "Both mention SigLIP2, but one is output probabilities while the other is a docs-example/model-processor/quantization issue." + }, + { + "left": "issue:45313", + "right": "issue:45357", + "accept": false, + "reason": "ZeRO-3 language_model loading failure and save_pretrained visual-encoder-key regression are different bugs." + }, + { + "left": "issue:42491", + "right": "issue:43720", + "accept": false, + "reason": "Qwen3_MoE LoRA compatibility and BitNet packed-weight loading are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:44488", + "accept": false, + "reason": "Both are model-loading failures, but the titles point to different checkpoints and no shared concrete failure mode." + }, + { + "left": "issue:44960", + "right": "issue:44977", + "accept": false, + "reason": "GLM5 and Qwen3.5 flash-attention generation failures are unrelated." + }, + { + "left": "issue:43708", + "right": "issue:44075", + "accept": false, + "reason": "Trainer max_steps resume bug and SGD args not being used are different training/optimizer issues." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The accepted soft edges are supported by closely matching titles and issue descriptions, and the rejected pairs are appropriately cautious where the evidence only suggests broad thematic similarity rather than the same bug. The cluster-level conclusion that this is heterogeneous with a few small duplicate subclusters is also well supported." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 5, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:36010", + "issue:39404", + "issue:39692", + "issue:41720", + "issue:42491", + "issue:42915", + "issue:42947", + "issue:43257", + "issue:43381", + "issue:43404", + "issue:43425", + "issue:43454", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43708", + "issue:43828", + "issue:43854", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44155", + "issue:44164", + "issue:44291", + "issue:44292", + "issue:44387", + "issue:44402", + "issue:44451", + "issue:44479", + "issue:44488", + "issue:44661", + "issue:44683", + "issue:44743", + "issue:44756", + "issue:44863", + "issue:44877", + "issue:44912", + "issue:44928", + "issue:44960", + "issue:44991", + "issue:45216", + "issue:45276", + "issue:45292", + "issue:45335", + "issue:45362", + "issue:45406" + ], + "soft_pairs": [ + "issue:43854|issue:43931", + "issue:43856|issue:43873", + "issue:41720|issue:42915", + "issue:45362|issue:45406", + "issue:42491|issue:45276", + "issue:39692|issue:43994", + "issue:44488|issue:44991", + "issue:43257|issue:43854", + "issue:44661|issue:44683", + "issue:43856|issue:44387", + "issue:42915|issue:44756", + "issue:43644|issue:43645", + "issue:42491|issue:43931", + "issue:42491|issue:43856", + "issue:45276|issue:45335", + "issue:43404|issue:43454", + "issue:36010|issue:39404", + "issue:44164|issue:44291", + "issue:44402|issue:44488", + "issue:41720|issue:44155", + "issue:43873|issue:44387", + "issue:39692|issue:44960", + "issue:44387|issue:44756", + "issue:43425|issue:43957", + "issue:44292|issue:44912", + "issue:43854|issue:45216", + "issue:42947|issue:43381", + "issue:44291|issue:44756", + "issue:45276|issue:45292", + "issue:44743|issue:44960", + "issue:43381|issue:44928", + "issue:43957|issue:44756", + "issue:44164|issue:44756", + "issue:43646|issue:43708", + "issue:45292|issue:45335", + "issue:44479|issue:44877", + "issue:44402|issue:44451", + "issue:43645|issue:44164", + "issue:43257|issue:44863", + "issue:43425|issue:43828", + "issue:43828|issue:43957" + ], + "prepared_review_unit_hash": "2eaaa3da7a1f901a99a53ab665d8bd4cb3e06591f523c030d412fd2e9f202b22", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 41, + "serialized_chars": 39811, + "estimated_input_tokens": 9953, + "estimated_eval_tokens": 20162 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous; only the resize_token_embeddings subgroup looks like true duplicates. Most other pairs are topical similarities rather than the same underlying bug.", + "confidence": 0.71, + "canonical_issue_reason": "No single canonical issue fits the whole cluster. For the only clear duplicate subgroup, issue 45292 is the best canonical candidate because it is the most generic description of the resize_token_embeddings regression and aligns with 45276/45335.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45292 is the broadest, most reusable representative of the strongest duplicate set and is the best anchor for triage cleanup.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43854", + "right": "issue:43931", + "accept": false, + "reason": "Different model-loading failures for different models; same broad area, not the same bug." + }, + { + "left": "issue:43856", + "right": "issue:43873", + "accept": false, + "reason": "Both concern quantization/memory behavior, but the reported failures and likely code paths differ." + }, + { + "left": "issue:41720", + "right": "issue:42915", + "accept": false, + "reason": "Qwen3 device-mapping crash vs Qwen3 MoE LoRA compatibility issue; not the same concrete problem." + }, + { + "left": "issue:45362", + "right": "issue:45406", + "accept": false, + "reason": "Different model/processor bugs: chat crash vs processor missing _tokenizer in serve." + }, + { + "left": "issue:42491", + "right": "issue:45276", + "accept": false, + "reason": "Both involve Qwen3.5/Qwen3 MoE, but one is a hf4.x/hf5.x compatibility issue and the other is token embedding resizing." + }, + { + "left": "issue:39692", + "right": "issue:43994", + "accept": false, + "reason": "Both mention SigLIP2, but one is a docs example error and the other is an inference/output regression." + }, + { + "left": "issue:44488", + "right": "issue:44991", + "accept": false, + "reason": "Tokenizer-loading failures for different models; similar symptom, not clearly the same root cause." + }, + { + "left": "issue:43257", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3 MoE weight conversion vs GLM-4.7 loading; unrelated bugs." + }, + { + "left": "issue:44661", + "right": "issue:44683", + "accept": false, + "reason": "Tokenizer-mapping failure vs torch flex_attention compilation failure; unrelated." + }, + { + "left": "issue:43856", + "right": "issue:44387", + "accept": false, + "reason": "Both are memory/quantization complaints, but the specific failure modes are different." + }, + { + "left": "issue:42915", + "right": "issue:44756", + "accept": false, + "reason": "Qwen3 MoE HF-version incompatibility vs mmap/OOM on Strix Halo; unrelated." + }, + { + "left": "issue:43644", + "right": "issue:43645", + "accept": false, + "reason": "Both are Transformers 5.0 regressions, but one is buffer initialization and the other is custom-model/Jupyter initialization." + }, + { + "left": "issue:42491", + "right": "issue:43931", + "accept": false, + "reason": "Different issues: MoE checkpoint compatibility vs Qwen3-VL shape mismatch." + }, + { + "left": "issue:42491", + "right": "issue:43856", + "accept": false, + "reason": "Different Qwen3 MoE problems; no shared concrete bug." + }, + { + "left": "issue:45276", + "right": "issue:45335", + "accept": true, + "reason": "Near-identical resize_token_embeddings regression affecting model-specific tied embeddings; same underlying bug family." + }, + { + "left": "issue:43404", + "right": "issue:43454", + "accept": false, + "reason": "Both are missing lm_head tying, but they are different model implementations and not proven to share one fixable code path." + }, + { + "left": "issue:36010", + "right": "issue:39404", + "accept": false, + "reason": "ImportError for GenerationMixin vs Whisper pipeline return_language regression; unrelated." + }, + { + "left": "issue:44164", + "right": "issue:44291", + "accept": false, + "reason": "save/from_pretrained extra_state handling vs init_empty_weights argument mismatch; separate serialization/loading bugs." + }, + { + "left": "issue:44402", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer vocab-size inconsistency vs a specific model-loading failure; not the same defect." + }, + { + "left": "issue:41720", + "right": "issue:44155", + "accept": false, + "reason": "CUDA assert in Qwen3 device mapping vs AudioFlamingo3 batched inference leakage; unrelated." + }, + { + "left": "issue:43873", + "right": "issue:44387", + "accept": false, + "reason": "General quantization/offloading complaint vs int4 reserved-memory OOM regression; not the same code-path problem." + }, + { + "left": "issue:39692", + "right": "issue:44960", + "accept": false, + "reason": "SigLIP2 example/model mismatch vs GLM5 issue; unrelated." + }, + { + "left": "issue:44387", + "right": "issue:44756", + "accept": false, + "reason": "Int4 CUDA reserved-memory regression vs Strix Halo mmap OOM; different mechanisms." + }, + { + "left": "issue:43425", + "right": "issue:43957", + "accept": false, + "reason": "Torch version incompatibility vs meta-device loading bug; not the same issue." + }, + { + "left": "issue:44292", + "right": "issue:44912", + "accept": false, + "reason": "Generic model-loading/initialization bug vs MXFP4 quantization fallback; unrelated." + }, + { + "left": "issue:43854", + "right": "issue:45216", + "accept": false, + "reason": "Different model families and different failure modes: load correctness vs save_pretrained regression." + }, + { + "left": "issue:42947", + "right": "issue:43381", + "accept": false, + "reason": "Both mention gradient checkpointing, but one is ineffective with PEFT LoRA and the other is eval-mode incompatibility." + }, + { + "left": "issue:44291", + "right": "issue:44756", + "accept": false, + "reason": "init_empty_weights TypeError vs mmap/OOM on Strix Halo; unrelated." + }, + { + "left": "issue:45276", + "right": "issue:45292", + "accept": false, + "reason": "Same resize_token_embeddings theme, but the evidence only shows similarity; not enough to treat as the same concrete bug without more detail." + }, + { + "left": "issue:44743", + "right": "issue:44960", + "accept": false, + "reason": "Qwen3.5 cache reset bug vs GLM5 model issue; unrelated." + }, + { + "left": "issue:43381", + "right": "issue:44928", + "accept": false, + "reason": "Gradient checkpointing eval-mode issue vs RLHF NaN from 3D position_ids/SDPA fallback; unrelated." + }, + { + "left": "issue:43957", + "right": "issue:44756", + "accept": false, + "reason": "Meta-device loading regression vs mmap/OOM workaround; different problems." + }, + { + "left": "issue:44164", + "right": "issue:44756", + "accept": false, + "reason": "Serialization extra_state bug vs Strix Halo mmap OOM; unrelated." + }, + { + "left": "issue:43646", + "right": "issue:43708", + "accept": false, + "reason": "Custom model initialization regression vs Trainer resume_from_checkpoint max_steps calculation; unrelated." + }, + { + "left": "issue:45292", + "right": "issue:45335", + "accept": true, + "reason": "Same resize_token_embeddings regression expressed generically and for a specific model class; clearly the same underlying bug family." + }, + { + "left": "issue:44479", + "right": "issue:44877", + "accept": false, + "reason": "Video input regression for Qwen models vs strict config preventing granite_speech loading; unrelated." + }, + { + "left": "issue:44402", + "right": "issue:44451", + "accept": false, + "reason": "Tokenizer vocab-size mismatch vs a model-specific loading failure; not the same bug." + }, + { + "left": "issue:43645", + "right": "issue:44164", + "accept": false, + "reason": "Notebook/custom-model initialization regression vs extra_state save/load failure; separate issues." + }, + { + "left": "issue:43257", + "right": "issue:44863", + "accept": false, + "reason": "Qwen3 MoE deepspeed loading bug vs NemotronH checkpoint loading failure; unrelated." + }, + { + "left": "issue:43425", + "right": "issue:43828", + "accept": false, + "reason": "Torch version incompatibility vs autocast dtype mismatch; different code paths." + }, + { + "left": "issue:43828", + "right": "issue:43957", + "accept": false, + "reason": "Autocast dtype mismatch vs meta-device model loading breakage; unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: this set is clearly heterogeneous, and the resize_token_embeddings trio is the only convincingly related subgroup. The two accepted soft edges stay conservative enough given the titles and high similarity scores, and the chosen canonical anchor (45292) is a reasonable generic representative for that subgroup." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 6, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:32090", + "issue:34567", + "issue:36032", + "issue:39404", + "issue:40444", + "issue:40990", + "issue:42947", + "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:43883", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43957", + "issue:43994", + "issue:44079", + "issue:44164", + "issue:44206", + "issue:44220", + "issue:44265", + "issue:44479", + "issue:44488", + "issue:44509", + "issue:44530", + "issue:44589", + "issue:44671", + "issue:44683", + "issue:44756", + "issue:44805", + "issue:44863", + "issue:44898", + "issue:44912", + "issue:44918", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45161" + ], + "soft_pairs": [ + "issue:44589|issue:44756", + "issue:43906|issue:44530", + "issue:43931|issue:44863", + "issue:39404|issue:44912", + "issue:43611|issue:43645", + "issue:43425|issue:43716", + "issue:44530|issue:45084", + "issue:43299|issue:43931", + "issue:44265|issue:44898", + "issue:43901|issue:44509", + "issue:44479|issue:44805", + "issue:43596|issue:44683", + "issue:43994|issue:44079", + "issue:44589|issue:45084", + "issue:43856|issue:44756", + "issue:44671|issue:44991", + "issue:39404|issue:44671", + "issue:40990|issue:42947", + "issue:43425|issue:44918", + "issue:36032|issue:43452", + "issue:44991|issue:45081", + "issue:32090|issue:39404", + "issue:43582|issue:43632", + "issue:43957|issue:44164", + "issue:44488|issue:45081", + "issue:44164|issue:44479", + "issue:40990|issue:43421", + "issue:43425|issue:45161", + "issue:43883|issue:44488", + "issue:43611|issue:43644", + "issue:44206|issue:44220", + "issue:44488|issue:44987", + "issue:42947|issue:44756", + "issue:43716|issue:45084", + "issue:34567|issue:40444", + "issue:43883|issue:44991" + ], + "prepared_review_unit_hash": "0968bb0c42ea639d854aa4e76d33c73574c94fb39c09a6d13d629b8978feb498", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 38479, + "estimated_input_tokens": 9620, + "estimated_eval_tokens": 19496 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most items are unrelated regressions or model-specific bugs, with a few loose topical similarities around Transformers v5 loading/tokenizer regressions, but no clear duplicate core.", + "confidence": 0.92, + "canonical_issue_reason": "No single canonical issue fits the whole set; the items span multiple independent bugs (trainer, tokenizer, vision, audio, MoE, loading, docs).", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43611 is the closest broad representative because it captures the largest shared theme here (Transformers v5 model-loading regressions), but it is still not a true duplicate umbrella for the rest of the cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44589", + "right": "issue:44756", + "accept": false, + "reason": "Different failures: float8 storage lookup vs mmap/OOM on Strix Halo; no shared code-path bug." + }, + { + "left": "issue:43906", + "right": "issue:44530", + "accept": false, + "reason": "Unrelated: isolated reproduction of a prior issue vs Qwen3.5 PagedAttentionCache linear_attention crash." + }, + { + "left": "issue:43931", + "right": "issue:44863", + "accept": false, + "reason": "Both are model-loading problems, but for different architectures and failure modes (shape mismatch vs checkpoint loading support)." + }, + { + "left": "issue:39404", + "right": "issue:44912", + "accept": false, + "reason": "Whisper pipeline return_language regression is unrelated to GPT-OSS MXFP4 quantization fallback." + }, + { + "left": "issue:43611", + "right": "issue:43645", + "accept": false, + "reason": "Both are v5 regressions, but one is base_model_prefix loading and the other is custom-model initialization in notebooks; different code paths." + }, + { + "left": "issue:43425", + "right": "issue:43716", + "accept": false, + "reason": "Torch version incompatibility is not the same bug as a Mistral image preprocessor/model dtype mismatch." + }, + { + "left": "issue:44530", + "right": "issue:45084", + "accept": false, + "reason": "Different subsystems and errors: Qwen3.5 cache group-type crash vs template compilation failure." + }, + { + "left": "issue:43299", + "right": "issue:43931", + "accept": false, + "reason": "Both involve Qwen3-VL loading, but one is a v5.dev0 MoE loading break and the other is a checkpoint shape mismatch; not clearly the same bug." + }, + { + "left": "issue:44265", + "right": "issue:44898", + "accept": false, + "reason": "torch.export/torch_compilable_check is unrelated to Perceiver positional encoding interpolation failure." + }, + { + "left": "issue:43901", + "right": "issue:44509", + "accept": false, + "reason": "Both are docs issues, but they concern different pipeline/task documentation changes." + }, + { + "left": "issue:44479", + "right": "issue:44805", + "accept": false, + "reason": "Video-input regression for Qwen VL models is unrelated to a generic mask shape IndexError." + }, + { + "left": "issue:43596", + "right": "issue:44683", + "accept": false, + "reason": "Deepspeed ZeRO-3/BertModel index error is unrelated to compiled flex_attention failing on newer torch." + }, + { + "left": "issue:43994", + "right": "issue:44079", + "accept": false, + "reason": "SigLIP2 bad outputs and ModelOutput key assignment bugs are different problems." + }, + { + "left": "issue:44589", + "right": "issue:45084", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to non-template-node compilation errors." + }, + { + "left": "issue:43856", + "right": "issue:44756", + "accept": false, + "reason": "MoE memory usage regression is unrelated to Strix Halo mmap/OOM behavior." + }, + { + "left": "issue:44671", + "right": "issue:44991", + "accept": false, + "reason": "CamemBERT prediction quality regression and tokenizer loading failure are distinct issues." + }, + { + "left": "issue:39404", + "right": "issue:44671", + "accept": false, + "reason": "Whisper pipeline return_language regression is unrelated to CamemBERT masked-LM output issues." + }, + { + "left": "issue:40990", + "right": "issue:42947", + "accept": false, + "reason": "High perplexity on GPT-OSS-20B and ineffective gradient checkpointing with LoRA are different training/eval bugs." + }, + { + "left": "issue:43425", + "right": "issue:44918", + "accept": false, + "reason": "Torch version compatibility is not the same as Qwen3.5 SFT embedding unpacking failure." + }, + { + "left": "issue:36032", + "right": "issue:43452", + "accept": false, + "reason": "T5Tokenizer add_special_tokens conflict is different from gguf_file breaking from_pretrained for tokenizer/model loading." + }, + { + "left": "issue:44991", + "right": "issue:45081", + "accept": false, + "reason": "Both are tokenizer-load regressions, but for different models and different failure mechanisms." + }, + { + "left": "issue:32090", + "right": "issue:39404", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error is unrelated to Whisper pipeline return_language behavior." + }, + { + "left": "issue:43582", + "right": "issue:43632", + "accept": false, + "reason": "Apple Silicon caching_allocator_warmup TypeError is unrelated to the _is_hf_initialized flag regression." + }, + { + "left": "issue:43957", + "right": "issue:44164", + "accept": false, + "reason": "Meta-device loading issues and save/from_pretrained extra_state handling are separate serialization/load bugs." + }, + { + "left": "issue:44488", + "right": "issue:45081", + "accept": false, + "reason": "Loading cjvt/sleng-bert and Mistral regex patch crashes are different tokenizer-load failures." + }, + { + "left": "issue:44164", + "right": "issue:44479", + "accept": false, + "reason": "extra_state save/load failure is unrelated to Qwen VL video-input regression." + }, + { + "left": "issue:40990", + "right": "issue:43421", + "accept": false, + "reason": "Perplexity regression on a model is unrelated to tokenizer post-processor update behavior." + }, + { + "left": "issue:43425", + "right": "issue:45161", + "accept": false, + "reason": "Torch 2.10 compatibility and GPT-OSS MoE tensor-parallel loading are different issues." + }, + { + "left": "issue:43883", + "right": "issue:44488", + "accept": false, + "reason": "Missing all_tied_weights_keys on MolmoForCausalLM is unrelated to sleng-bert tokenizer loading." + }, + { + "left": "issue:43611", + "right": "issue:43644", + "accept": false, + "reason": "Both are v5 loading regressions, but one is base_model_prefix handling and the other is non-persistent buffer initialization; different defects." + }, + { + "left": "issue:44206", + "right": "issue:44220", + "accept": false, + "reason": "Audio feature extraction issues are related in area, but one is an unsupported center arg regression and the other is a generic _torch_extract_fbank_features problem; not clearly the same bug." + }, + { + "left": "issue:44488", + "right": "issue:44987", + "accept": false, + "reason": "Sleng-bert tokenizer loading and physical-intelligence/fast model loading are different failures." + }, + { + "left": "issue:42947", + "right": "issue:44756", + "accept": false, + "reason": "LoRA gradient checkpointing inefficiency is unrelated to Strix Halo mmap/OOM mitigation." + }, + { + "left": "issue:43883", + "right": "issue:44991", + "accept": false, + "reason": "Molmo tied-weights attribute error and est-roberta tokenizer loading 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": 2, + "component_count": 13, + "review_unit_index": 7, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:36032", + "issue:39404", + "issue:41553", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43452", + "issue:43482", + "issue:43531", + "issue:43540", + "issue:43541", + "issue:43582", + "issue:43632", + "issue:43643", + "issue:43644", + "issue:43645", + "issue:43646", + "issue:43819", + "issue:43825", + "issue:43994", + "issue:44062", + "issue:44220", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44451", + "issue:44466", + "issue:44479", + "issue:44530", + "issue:44534", + "issue:44610", + "issue:44683", + "issue:44756", + "issue:44849", + "issue:44987", + "issue:44991", + "issue:45081", + "issue:45084", + "issue:45127", + "issue:45216", + "issue:45325", + "issue:45357", + "issue:45507" + ], + "soft_pairs": [ + "issue:43644|issue:43646", + "issue:43643|issue:44530", + "issue:44451|issue:44987", + "issue:43541|issue:45084", + "issue:43452|issue:43482", + "issue:43317|issue:43482", + "issue:36032|issue:42915", + "issue:43262|issue:43377", + "issue:39404|issue:43825", + "issue:42915|issue:44530", + "issue:44849|issue:45325", + "issue:36032|issue:43645", + "issue:43404|issue:44466", + "issue:45084|issue:45507", + "issue:44987|issue:44991", + "issue:43646|issue:44534", + "issue:43994|issue:44610", + "issue:43377|issue:43819", + "issue:43645|issue:44534", + "issue:44403|issue:45081", + "issue:43322|issue:43540", + "issue:45216|issue:45357", + "issue:44451|issue:45081", + "issue:44220|issue:44683", + "issue:44265|issue:44610", + "issue:43632|issue:44756", + "issue:42915|issue:45084", + "issue:41553|issue:43582", + "issue:43404|issue:45127", + "issue:44987|issue:45081", + "issue:43296|issue:43531", + "issue:44062|issue:44987", + "issue:44479|issue:45325", + "issue:42915|issue:44291" + ], + "prepared_review_unit_hash": "9f12d1f54967670c7702848b5669903a2c7df7511503a1e6a7440a1844b99c9a", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 37380, + "estimated_input_tokens": 9345, + "estimated_eval_tokens": 18946 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly heterogeneous; most soft links are false positives across unrelated model-loading, tokenizer, audio, and vision regressions. The only clear duplicate-looking pair is the Qwen3.5 `save_pretrained` checkpoint serialization regression.", + "confidence": 0.91, + "canonical_issue_reason": "issue:45216 is the broadest and clearest representative of the only strong duplicate pair in this set; it describes the underlying Qwen3.5 `save_pretrained` checkpoint corruption regression, while issue:45357 is a narrower follow-up about the same visual-encoder key problem.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45216 is the best global issue because it captures the core bug at a higher level and matches the accepted duplicate link more directly than the follow-up.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43644", + "right": "issue:43646", + "accept": false, + "reason": "Both mention Transformers v5 regressions, but one is about non-persistent buffers and the other about custom model initialization; different bugs." + }, + { + "left": "issue:43643", + "right": "issue:44530", + "accept": false, + "reason": "`trust_remote_code` config field loss is unrelated to Qwen3.5 PagedAttentionCache/group-type failures." + }, + { + "left": "issue:44451", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading failures, but they concern different models and likely different root causes." + }, + { + "left": "issue:43541", + "right": "issue:45084", + "accept": false, + "reason": "Torch Dynamo tracing on Mixtral and template compilation errors are unrelated code paths." + }, + { + "left": "issue:43452", + "right": "issue:43482", + "accept": false, + "reason": "Both involve loading, but one is generic gguf_file behavior and the other is a specific Qwen2.5-GGUF/v5 failure; not clearly the same bug." + }, + { + "left": "issue:43317", + "right": "issue:43482", + "accept": false, + "reason": "Device_map offload/dequantization and GGUF tokenizer/model loading are different failure modes." + }, + { + "left": "issue:36032", + "right": "issue:42915", + "accept": false, + "reason": "Tokenizer method-name conflict and Qwen3Moe FP8 config failure are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:43377", + "accept": false, + "reason": "Audio chat-template sample-rate defaulting and MIMI batching/padding-mask mismatch are separate audio bugs." + }, + { + "left": "issue:39404", + "right": "issue:43825", + "accept": false, + "reason": "Whisper pipeline language-return behavior and a generic pipeline translation error message are not the same issue." + }, + { + "left": "issue:42915", + "right": "issue:44530", + "accept": false, + "reason": "FineGrainedFP8Config failure and linear_attention group-type crash affect different subsystems." + }, + { + "left": "issue:44849", + "right": "issue:45325", + "accept": false, + "reason": "Qwen3.5 hidden-states output bug and Qwen2.5-VL rope-index scaling bug are distinct model-specific regressions." + }, + { + "left": "issue:36032", + "right": "issue:43645", + "accept": false, + "reason": "Tokenizer special-token conflict and custom model initialization in notebooks are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:44466", + "accept": false, + "reason": "Missing tied lm_head weights and inconsistent serialization of tied weights are related thematically but not the same concrete bug." + }, + { + "left": "issue:45084", + "right": "issue:45507", + "accept": false, + "reason": "Template-node compilation failure and GraniteMoEHybrid invalid-method call are different problems." + }, + { + "left": "issue:44987", + "right": "issue:44991", + "accept": false, + "reason": "Both are loading failures on v5, but they involve different models and likely different tokenizer/config issues." + }, + { + "left": "issue:43646", + "right": "issue:44534", + "accept": false, + "reason": "Custom model initialization and non-persistent buffer junk are unrelated Transformers v5 regressions." + }, + { + "left": "issue:43994", + "right": "issue:44610", + "accept": false, + "reason": "SigLIP2 output quality and OmDet-Turbo input-size mismatch are not the same bug." + }, + { + "left": "issue:43377", + "right": "issue:43819", + "accept": false, + "reason": "MIMI padding-mask mismatch and DAC STE mismatch are different audio-model correctness issues." + }, + { + "left": "issue:43645", + "right": "issue:44534", + "accept": false, + "reason": "Notebook custom-model initialization does not match the non-persistent-buffer serialization bug." + }, + { + "left": "issue:44403", + "right": "issue:45081", + "accept": false, + "reason": "Generic loading noise and Mistral regex patch crash are unrelated." + }, + { + "left": "issue:43322", + "right": "issue:43540", + "accept": false, + "reason": "Llava Next segfault and Qwen3OmniMoe video-processing ValueError are distinct multimodal failures." + }, + { + "left": "issue:45216", + "right": "issue:45357", + "accept": true, + "reason": "Both describe the same Qwen3.5 `save_pretrained` checkpoint corruption / incorrect visual-encoder key serialization regression." + }, + { + "left": "issue:44451", + "right": "issue:45081", + "accept": false, + "reason": "A model-loading failure for one checkpoint is not the same as a tokenizer regex-patching crash." + }, + { + "left": "issue:44220", + "right": "issue:44683", + "accept": false, + "reason": "FBank feature extraction and compiled flex_attention on torch>=2.9 are unrelated." + }, + { + "left": "issue:44265", + "right": "issue:44610", + "accept": false, + "reason": "torch.export with `torch_compilable_check` and OmDet-Turbo preprocessing are separate issues." + }, + { + "left": "issue:43632", + "right": "issue:44756", + "accept": false, + "reason": "`_is_hf_initialized` flag behavior and Strix Halo mmap OOM are different failures." + }, + { + "left": "issue:42915", + "right": "issue:45084", + "accept": false, + "reason": "Qwen3Moe FP8 loading and template compilation errors have no shared root cause." + }, + { + "left": "issue:41553", + "right": "issue:43582", + "accept": false, + "reason": "Bad AutoTokenizer error message for Voxtral and AppleSilicon warmup TypeError are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45127", + "accept": false, + "reason": "Tied lm_head weight bug in Mistral3 is not the same as LoRA merge collapse after vocab extension." + }, + { + "left": "issue:44987", + "right": "issue:45081", + "accept": false, + "reason": "Different model families and different failure points; no clear common bug." + }, + { + "left": "issue:43296", + "right": "issue:43531", + "accept": false, + "reason": "PaddleOCR-VL loading and Qwen3-MoE sliding_window behavior are unrelated." + }, + { + "left": "issue:44062", + "right": "issue:44987", + "accept": false, + "reason": "AddedToken special-argument conflict is a tokenizer construction bug, not the same as model loading failures for physical-intelligence/fast." + }, + { + "left": "issue:44479", + "right": "issue:45325", + "accept": false, + "reason": "Both involve Qwen VL temporal/video behavior, but one is a broad video-input regression and the other is a still-image rope-index scaling bug; too different to merge." + }, + { + "left": "issue:42915", + "right": "issue:44291", + "accept": false, + "reason": "Qwen3Moe FP8 loading and `_is_hf_initialized` init_empty_weights TypeError are separate regressions." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative overall. The only accepted soft pair, issue:45216 vs issue:45357, is well supported by the titles as the same Qwen3.5 save_pretrained serialization regression. The other pair verdicts generally avoid overclaiming and are phrased cautiously enough from the available evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 8, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:33666", + "issue:35707", + "issue:36032", + "issue:40444", + "issue:40990", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43296", + "issue:43317", + "issue:43377", + "issue:43421", + "issue:43425", + "issue:43482", + "issue:43525", + "issue:43531", + "issue:43611", + "issue:43632", + "issue:43644", + "issue:43646", + "issue:43742", + "issue:43749", + "issue:43792", + "issue:43818", + "issue:43856", + "issue:43881", + "issue:44079", + "issue:44265", + "issue:44291", + "issue:44403", + "issue:44488", + "issue:44534", + "issue:44568", + "issue:44589", + "issue:44683", + "issue:44811", + "issue:44829", + "issue:44877", + "issue:44991", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45357", + "issue:45381", + "issue:45479", + "issue:45491" + ], + "soft_pairs": [ + "issue:40990|issue:44811", + "issue:43611|issue:44534", + "issue:43317|issue:43856", + "issue:36032|issue:43742", + "issue:40444|issue:45325", + "issue:42915|issue:44589", + "issue:43632|issue:44534", + "issue:43531|issue:44877", + "issue:44683|issue:45020", + "issue:43818|issue:45357", + "issue:43421|issue:44568", + "issue:36032|issue:41720", + "issue:44829|issue:45479", + "issue:43425|issue:44403", + "issue:43792|issue:44589", + "issue:44403|issue:44991", + "issue:33666|issue:40444", + "issue:44079|issue:44534", + "issue:43377|issue:45491", + "issue:35707|issue:40444", + "issue:43482|issue:43525", + "issue:44291|issue:44589", + "issue:43749|issue:44079", + "issue:43644|issue:44079", + "issue:44079|issue:45072", + "issue:44265|issue:45072", + "issue:43632|issue:45072", + "issue:40444|issue:45381", + "issue:43531|issue:44291", + "issue:42175|issue:43296", + "issue:41720|issue:45081", + "issue:45325|issue:45381", + "issue:43611|issue:43646", + "issue:43881|issue:44488" + ], + "prepared_review_unit_hash": "3b8671bbcc9ed5b1cbb502bacd4bd19c2e7e75c952b0ede05ebf656a10a0683d", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 37276, + "estimated_input_tokens": 9319, + "estimated_eval_tokens": 18894 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab-bag of distinct issues, with a dominant theme of Transformers v5 loading/regression reports. Only the two Qwen2.5-VL position-id reports look like the same underlying bug; the rest are too model- or subsystem-specific to merge as duplicates.", + "confidence": 0.78, + "canonical_issue_reason": "Issue 43611 is the best representative of the cluster\u2019s main theme: broad Transformers v5 loading regressions. It is more central and reusable as an umbrella than the highly specific model-failure reports, even though this cluster is not a clean duplicate set.", + "canonical_pr_reason": null, + "best_issue_reason": "43611 is the strongest global issue candidate because it captures the repeated v5 model-loading breakage pattern seen across several nearby reports, without being tied to one narrow model or symptom.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:40990", + "right": "issue:44811", + "accept": false, + "reason": "Different failures: language-model perplexity on GPT-OSS vs Whisper batch_decode special-token handling." + }, + { + "left": "issue:43611", + "right": "issue:44534", + "accept": false, + "reason": "Both are Transformers v5 regressions, but they describe different code paths: base_model_prefix loading vs non-persistent buffer initialization." + }, + { + "left": "issue:43317", + "right": "issue:43856", + "accept": false, + "reason": "Unrelated: device_map auto/offload loading failure vs Qwen3 MoE training memory usage." + }, + { + "left": "issue:36032", + "right": "issue:43742", + "accept": false, + "reason": "Different models and symptoms: T5 tokenizer class conflict vs MobileLLM loading KeyError." + }, + { + "left": "issue:40444", + "right": "issue:45325", + "accept": false, + "reason": "Same model family, but different bugs: multi-image iterable dataset failure vs still-image temporal position scaling." + }, + { + "left": "issue:42915", + "right": "issue:44589", + "accept": false, + "reason": "Both touch float8/fp8 territory, but the reported failures are different: FineGrainedFP8Config vs missing Float8 storage object." + }, + { + "left": "issue:43632", + "right": "issue:44534", + "accept": false, + "reason": "Both are v5 regressions, but one is about the `_is_hf_initialized` flag and the other about non-persistent buffers." + }, + { + "left": "issue:43531", + "right": "issue:44877", + "accept": false, + "reason": "Different issues: Qwen3-MoE sliding window behavior vs strict config blocking granite_speech loading." + }, + { + "left": "issue:44683", + "right": "issue:45020", + "accept": false, + "reason": "Different subsystems: compiled flex_attention on torch>=2.9 vs remote_code model loading regressions." + }, + { + "left": "issue:43818", + "right": "issue:45357", + "accept": false, + "reason": "Different multimodal problems: Video-LLaVA tower/weight sharing vs Qwen3.5 save_pretrained key names." + }, + { + "left": "issue:43421", + "right": "issue:44568", + "accept": false, + "reason": "Related tokenization area, but not the same bug: runtime special-token updates vs mDeBERTa add_special_tokens output behavior." + }, + { + "left": "issue:36032", + "right": "issue:41720", + "accept": false, + "reason": "Completely unrelated: tokenizer attribute conflict vs Qwen3 device-map cudaErrorAssert." + }, + { + "left": "issue:44829", + "right": "issue:45479", + "accept": false, + "reason": "Both are training pathologies, but the concrete failures differ: flash_attention_3 degeneration vs zero-loss single-label classification." + }, + { + "left": "issue:43425", + "right": "issue:44403", + "accept": false, + "reason": "Unrelated: torch 2.10 compatibility issue vs noisy model-loading output." + }, + { + "left": "issue:43792", + "right": "issue:44589", + "accept": false, + "reason": "Different failures and code paths: Whisper runtime issue vs Float8 storage lookup error." + }, + { + "left": "issue:44403", + "right": "issue:44991", + "accept": false, + "reason": "Different symptoms: logging/noise during loading vs tokenizer loading failure for a specific model." + }, + { + "left": "issue:33666", + "right": "issue:40444", + "accept": false, + "reason": "Different Qwen2-VL family problems: multi-GPU training vs iterable-dataset multi-image finetuning." + }, + { + "left": "issue:44079", + "right": "issue:44534", + "accept": false, + "reason": "Unrelated bugs: ModelOutput key assignment vs non-persistent buffer initialization." + }, + { + "left": "issue:43377", + "right": "issue:45491", + "accept": false, + "reason": "Both involve batching/padding, but they affect different models and likely different implementations." + }, + { + "left": "issue:35707", + "right": "issue:40444", + "accept": false, + "reason": "Different generation/mechanism bug vs multimodal finetuning failure." + }, + { + "left": "issue:43482", + "right": "issue:43525", + "accept": false, + "reason": "Different loading failures: Qwen2.5-GGUF/Transformers v5 vs missing pad_token_id on Llama4Config." + }, + { + "left": "issue:44291", + "right": "issue:44589", + "accept": false, + "reason": "Different initialization failures: init_empty_weights/_is_hf_initialized vs Float8 storage object lookup." + }, + { + "left": "issue:43749", + "right": "issue:44079", + "accept": false, + "reason": "Different areas: FSDP CPU RAM efficient loading vs ModelOutput key assignment." + }, + { + "left": "issue:43644", + "right": "issue:44079", + "accept": false, + "reason": "Different bug classes: non-persistent buffers filled with junk vs ModelOutput key assignment." + }, + { + "left": "issue:44079", + "right": "issue:45072", + "accept": false, + "reason": "Unrelated: ModelOutput key handling vs bfloat16 dtype mismatch in CI." + }, + { + "left": "issue:44265", + "right": "issue:45072", + "accept": false, + "reason": "Different problems: torch.export with torch_compilable_check vs dtype mismatch during inference." + }, + { + "left": "issue:43632", + "right": "issue:45072", + "accept": false, + "reason": "Different Transformers v5 regressions: `_is_hf_initialized` handling vs bfloat16 dtype mismatch." + }, + { + "left": "issue:40444", + "right": "issue:45381", + "accept": false, + "reason": "Same model family, but one is multi-image finetuning failure and the other is video position-id mismatch." + }, + { + "left": "issue:43531", + "right": "issue:44291", + "accept": false, + "reason": "Different issues: Qwen3-MoE sliding window behavior vs unexpected `_is_hf_initialized` arg in init_empty_weights." + }, + { + "left": "issue:41720", + "right": "issue:45081", + "accept": false, + "reason": "Unrelated: Qwen3 device-map CUDA assert vs Mistral tokenizer regex patch crash." + }, + { + "left": "issue:45325", + "right": "issue:45381", + "accept": true, + "reason": "Both describe the same Qwen2.5-VL position-id / rope-index bug in transformers 5.3.0, just from still-image vs video inputs." + }, + { + "left": "issue:43611", + "right": "issue:43646", + "accept": false, + "reason": "Related only at a high level; the concrete failures differ between base_model_prefix loading and custom model initialization." + }, + { + "left": "issue:43881", + "right": "issue:44488", + "accept": false, + "reason": "Different model-loading failures for different models and causes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The cluster read is grounded in the packet: it is mostly a mix of distinct issues, with a real recurring Transformers v5 loading/regression theme. The soft-edge verdicts are conservative overall, and the single accepted Qwen2.5-VL pair is narrowly tied to the same position-id/rope-index path rather than an obviously broader leap." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 9, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:34634", + "issue:35707", + "issue:39692", + "issue:40444", + "issue:41720", + "issue:42915", + "issue:43262", + "issue:43296", + "issue:43317", + "issue:43322", + "issue:43377", + "issue:43404", + "issue:43454", + "issue:43482", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43632", + "issue:43701", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43792", + "issue:43819", + "issue:43828", + "issue:43873", + "issue:43881", + "issue:43883", + "issue:43957", + "issue:43994", + "issue:44164", + "issue:44186", + "issue:44265", + "issue:44291", + "issue:44387", + "issue:44466", + "issue:44534", + "issue:44610", + "issue:44617", + "issue:44683", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:44991", + "issue:45072", + "issue:45325", + "issue:45357" + ], + "soft_pairs": [ + "issue:43994|issue:44898", + "issue:43828|issue:44683", + "issue:43454|issue:44466", + "issue:44387|issue:44617", + "issue:42915|issue:43482", + "issue:43792|issue:43957", + "issue:43526|issue:43883", + "issue:43701|issue:45325", + "issue:43262|issue:43746", + "issue:43317|issue:43873", + "issue:34634|issue:35707", + "issue:44291|issue:44877", + "issue:43881|issue:44991", + "issue:43606|issue:44265", + "issue:44186|issue:44265", + "issue:43606|issue:43819", + "issue:43632|issue:44164", + "issue:35707|issue:41720", + "issue:43377|issue:44898", + "issue:44610|issue:45325", + "issue:43377|issue:44610", + "issue:43454|issue:43550", + "issue:43454|issue:45072", + "issue:43749|issue:44534", + "issue:39692|issue:40444", + "issue:43296|issue:44991", + "issue:44466|issue:45357", + "issue:43296|issue:44877", + "issue:43792|issue:44683", + "issue:43454|issue:43720", + "issue:41720|issue:43322", + "issue:41720|issue:44977", + "issue:43404|issue:45357" + ], + "prepared_review_unit_hash": "904bfbb2d4032d88f1578f794503f2bc0f4e837c1ab9b099fb6d219b842516e1", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36982, + "estimated_input_tokens": 9246, + "estimated_eval_tokens": 18748 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is heterogeneous; the soft links mostly connect superficially similar bug themes (loading, offloading, weight tying, generation) rather than the same concrete defect. I found no safe duplicate merges among the candidate pairs.", + "confidence": 0.66, + "canonical_issue_reason": "No single true canonical issue emerges from these pairs. If one representative must be chosen, issue 43749 is the strongest standalone bug report with the most discussion and duplicate-triage value around loading/offload regressions.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43873 is the broadest open report and the closest umbrella for ongoing offloading/quantization failures, though it still does not subsume the whole set cleanly.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43994", + "right": "issue:44898", + "accept": false, + "reason": "Different models and failure modes: SigLIP2 pipeline/nonsensical outputs vs Perceiver resolution/interpolation behavior." + }, + { + "left": "issue:43828", + "right": "issue:44683", + "accept": false, + "reason": "One is an autocast dtype mismatch in Phi-tiny-MoE; the other is a torch>=2.9 flex_attention compilation regression." + }, + { + "left": "issue:43454", + "right": "issue:44466", + "accept": false, + "reason": "Both touch tied weights, but one is an init/tie_weights bug causing bad generation while the other is device-dependent save_pretrained serialization." + }, + { + "left": "issue:44387", + "right": "issue:44617", + "accept": false, + "reason": "Both mention OOM, but one is int4 memory growth in Transformers 5.x and the other is a Sam3Video runtime OOM; different concrete causes." + }, + { + "left": "issue:42915", + "right": "issue:43482", + "accept": false, + "reason": "Different artifacts and failures: FineGrainedFP8Config load failure vs GGUF loading failure in transformers v5." + }, + { + "left": "issue:43792", + "right": "issue:43957", + "accept": false, + "reason": "Whisper runtime failure is unrelated to the meta-device loading regression affecting other models." + }, + { + "left": "issue:43526", + "right": "issue:43883", + "accept": false, + "reason": "BEiT image processor label reduction bug is unrelated to the Molmo missing all_tied_weights_keys AttributeError." + }, + { + "left": "issue:43701", + "right": "issue:45325", + "accept": false, + "reason": "Checkpoint resume key mismatch and Qwen2.5-VL rope-index scaling are different code paths." + }, + { + "left": "issue:43262", + "right": "issue:43746", + "accept": false, + "reason": "Audio sampling-rate defaulting in apply_chat_template is unrelated to GraniteSpeech PEFT checkpoint loading." + }, + { + "left": "issue:43317", + "right": "issue:43873", + "accept": false, + "reason": "Both are offloading/quantization themed, but one is a dequantized gpu+cpu offload load failure and the other is a broader quantization offloading complaint; not the same concrete bug." + }, + { + "left": "issue:34634", + "right": "issue:35707", + "accept": false, + "reason": "BarkProcessor voice_preset and progressive generation with inputs_embeds/past_key_values are unrelated failures." + }, + { + "left": "issue:44291", + "right": "issue:44877", + "accept": false, + "reason": "Unexpected _is_hf_initialized argument handling is distinct from strict config rejection for granite_speech." + }, + { + "left": "issue:43881", + "right": "issue:44991", + "accept": false, + "reason": "glm-4v-9b loading failure and EMBEDDIA tokenizer loading failure are separate model-specific issues." + }, + { + "left": "issue:43606", + "right": "issue:44265", + "accept": false, + "reason": "CPU-offload device mismatch for Bark-small is unrelated to torch.export failures involving torch_compilable_check." + }, + { + "left": "issue:44186", + "right": "issue:44265", + "accept": false, + "reason": "LayoutLMv2 tokenizer crashes on NER/padding are unrelated to torch.export/torch_compilable_check." + }, + { + "left": "issue:43606", + "right": "issue:43819", + "accept": false, + "reason": "Device mismatch in Bark offload and DAC.from_latents vs forward mismatch are different bugs." + }, + { + "left": "issue:43632", + "right": "issue:44164", + "accept": false, + "reason": "_is_hf_initialized regression and extra_state save/from_pretrained handling are related to loading internals but not the same defect." + }, + { + "left": "issue:35707", + "right": "issue:41720", + "accept": false, + "reason": "Progressive generation with inputs_embeds/past_key_values is unrelated to Qwen3 auto device-map CUDA asserts." + }, + { + "left": "issue:43377", + "right": "issue:44898", + "accept": false, + "reason": "MIMI batching/padding-mask mismatch is unrelated to Perceiver non-default-resolution failures." + }, + { + "left": "issue:44610", + "right": "issue:45325", + "accept": false, + "reason": "OmDet-Turbo processor size mismatch and Qwen2.5-VL temporal position scaling are different processor/model bugs." + }, + { + "left": "issue:43377", + "right": "issue:44610", + "accept": false, + "reason": "Batched-vs-single MIMI encoder outputs and OmDet-Turbo image-size mismatch are separate issues." + }, + { + "left": "issue:43454", + "right": "issue:43550", + "accept": false, + "reason": "lm_head tying/serialization bug is unrelated to Bamba torch.compile SDPA failure." + }, + { + "left": "issue:43454", + "right": "issue:45072", + "accept": false, + "reason": "Weight-tying bug and bfloat16 dtype mismatch are different failure modes." + }, + { + "left": "issue:43749", + "right": "issue:44534", + "accept": false, + "reason": "FSDP CPU RAM efficient loading and non-persistent buffer junk filling are both v5/loading topics but not the same code-path bug." + }, + { + "left": "issue:39692", + "right": "issue:40444", + "accept": false, + "reason": "SigLIP2 doc example errors and Qwen2.5-VL iterable-dataset finetuning failure are unrelated." + }, + { + "left": "issue:43296", + "right": "issue:44991", + "accept": false, + "reason": "PaddleOCR-VL load failure in vLLM is unrelated to EMBEDDIA tokenizer loading in transformers v5." + }, + { + "left": "issue:44466", + "right": "issue:45357", + "accept": false, + "reason": "Both involve serialization, but one is tied lm_head.weight serialization and the other is incorrect visual encoder keys in Qwen3.5 save_pretrained." + }, + { + "left": "issue:43296", + "right": "issue:44877", + "accept": false, + "reason": "PaddleOCR-VL/vLLM loading failure is unrelated to strict config handling for granite_speech." + }, + { + "left": "issue:43792", + "right": "issue:44683", + "accept": false, + "reason": "Whisper loading/runtime failure and compiled flex_attention failure are unrelated." + }, + { + "left": "issue:43454", + "right": "issue:43720", + "accept": false, + "reason": "Tied lm_head weight bug and BitNet packed-weight unpacking during accelerate loading are different defects." + }, + { + "left": "issue:41720", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3 auto device-map CUDA assert and Llava Next segmentation fault are separate model-loading failures." + }, + { + "left": "issue:41720", + "right": "issue:44977", + "accept": false, + "reason": "Auto device mapping CUDA assert and flash-attention generation failure for Qwen3.5 are different issues." + }, + { + "left": "issue:43404", + "right": "issue:45357", + "accept": false, + "reason": "Mistral3 lm_head tying bug is unrelated to Qwen3.5 visual-encoder save_pretrained key regression." + } + ] + }, + "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": 13, + "review_unit_index": 10, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:33357", + "issue:34567", + "issue:35707", + "issue:36010", + "issue:36331", + "issue:40990", + "issue:41720", + "issue:42617", + "issue:42915", + "issue:42994", + "issue:43316", + "issue:43317", + "issue:43322", + "issue:43388", + "issue:43482", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43708", + "issue:43720", + "issue:43782", + "issue:43873", + "issue:43881", + "issue:44186", + "issue:44292", + "issue:44361", + "issue:44423", + "issue:44479", + "issue:44514", + "issue:44545", + "issue:44610", + "issue:44671", + "issue:44743", + "issue:44756", + "issue:44877", + "issue:44898", + "issue:44913", + "issue:44918", + "issue:44991", + "issue:45059", + "issue:45072", + "issue:45161", + "issue:45406", + "issue:45464" + ], + "soft_pairs": [ + "issue:44423|issue:45406", + "issue:44186|issue:45072", + "issue:34567|issue:40990", + "issue:43873|issue:44756", + "issue:43526|issue:43881", + "issue:43577|issue:43708", + "issue:42617|issue:42915", + "issue:44877|issue:44991", + "issue:43550|issue:43720", + "issue:44610|issue:44898", + "issue:44479|issue:44918", + "issue:44186|issue:44610", + "issue:36331|issue:42915", + "issue:36331|issue:42617", + "issue:42994|issue:43482", + "issue:43316|issue:44913", + "issue:44671|issue:44743", + "issue:42915|issue:43317", + "issue:43550|issue:43606", + "issue:41720|issue:42617", + "issue:43531|issue:43572", + "issue:44292|issue:44479", + "issue:41720|issue:45161", + "issue:43388|issue:45059", + "issue:44186|issue:44361", + "issue:43540|issue:43782", + "issue:41720|issue:43873", + "issue:45406|issue:45464", + "issue:42915|issue:43322", + "issue:36010|issue:42617", + "issue:36331|issue:41720", + "issue:44514|issue:44545", + "issue:33357|issue:35707", + "issue:43606|issue:43720" + ], + "prepared_review_unit_hash": "1668ae07276aef46dcd4c82c43be2b98a6c326ccff952fe5b8939494fd96b3a8", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37685, + "estimated_input_tokens": 9422, + "estimated_eval_tokens": 19100 + }, + "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 across different models, processors, and trainer/config paths. The only clear duplicate pair is the Qwen2_5_VLProcessor.apply_chat_template batched-input padding=False crash (issues 44514 and 44545).", + "confidence": 0.96, + "canonical_issue_reason": "issue:44545 is the clearest representative of the only confirmed duplicate bug: the same Qwen2_5_VLProcessor.apply_chat_template crash on batched input with padding=False, with the cleaner title and later corroborating report.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44545 is the best canonical issue candidate because it precisely describes the exact duplicate bug and is the more specific, readable report.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44423", + "right": "issue:45406", + "accept": false, + "reason": "Both are serve/multimodal crashes, but the failing paths differ: continuous-batching Qwen3.5 AttributeError vs Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:44186", + "right": "issue:45072", + "accept": false, + "reason": "Different bugs: LayoutLMv2Tokenizer NER/padding crash versus dtype mismatch in SwitchTransformers/TimmWrapperModel inference." + }, + { + "left": "issue:34567", + "right": "issue:40990", + "accept": false, + "reason": "TrainerState token-count tracking is unrelated to GPT-OSS perplexity evaluation." + }, + { + "left": "issue:43873", + "right": "issue:44756", + "accept": false, + "reason": "Both mention offloading/memory, but one is quantization offloading behavior and the other is mmap OOM on Strix Halo; not the same bug." + }, + { + "left": "issue:43526", + "right": "issue:43881", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug is unrelated to glm-4v-9b loading failure." + }, + { + "left": "issue:43577", + "right": "issue:43708", + "accept": false, + "reason": "Different regression classes: Blip2 dtype propagation vs Trainer resume/max_steps calculation." + }, + { + "left": "issue:42617", + "right": "issue:42915", + "accept": false, + "reason": "Both are model-run failures, but one is a 3d_parallel.py runner issue and the other is Qwen3Moe FineGrainedFP8Config loading." + }, + { + "left": "issue:44877", + "right": "issue:44991", + "accept": false, + "reason": "Both concern config/tokenizer loading after v5, but one is strict model config handling and the other is a tokenizer-loading regression for a specific model." + }, + { + "left": "issue:43550", + "right": "issue:43720", + "accept": false, + "reason": "Different code paths: Bamba torch.compile with SDPA versus BitNet packed-weight unpacking during accelerate loading." + }, + { + "left": "issue:44610", + "right": "issue:44898", + "accept": false, + "reason": "Different vision-model bugs: OmDet processor size mismatch versus Perceiver non-default-resolution interpolation failure." + }, + { + "left": "issue:44479", + "right": "issue:44918", + "accept": false, + "reason": "Both involve Qwen-family regressions, but one is video-input handling and the other is trl SFT unpacking of embeddings." + }, + { + "left": "issue:44186", + "right": "issue:44610", + "accept": false, + "reason": "Different processors/models and failures: LayoutLMv2Tokenizer crash versus OmDet image-size mismatch." + }, + { + "left": "issue:36331", + "right": "issue:42915", + "accept": false, + "reason": "Trainer custom compute_loss signature mismatch is unrelated to Qwen3Moe FineGrainedFP8Config loading." + }, + { + "left": "issue:36331", + "right": "issue:42617", + "accept": false, + "reason": "CustomTrainer keyword-argument regression is unrelated to the 3d_parallel.py execution issue." + }, + { + "left": "issue:42994", + "right": "issue:43482", + "accept": false, + "reason": "Quantized model saving failure is not the same as Qwen2.5-GGUF loading failure under transformers v5." + }, + { + "left": "issue:43316", + "right": "issue:44913", + "accept": false, + "reason": "Different config serialization bugs: Gemma3TextConfig API mismatch vs GPTNeoX rotary_pct not persisting on reload." + }, + { + "left": "issue:44671", + "right": "issue:44743", + "accept": false, + "reason": "CamemBERT masked-LM prediction regression is unrelated to Qwen3.5 recurrent-state cache reset." + }, + { + "left": "issue:42915", + "right": "issue:43317", + "accept": false, + "reason": "FineGrainedFP8Config model load failure is unrelated to dequantized model loading with device_map=auto offload." + }, + { + "left": "issue:43550", + "right": "issue:43606", + "accept": false, + "reason": "Both are runtime/inference failures, but the model families and failures differ: Bamba SDPA compile issue vs bark-small CPU offload device mismatch." + }, + { + "left": "issue:41720", + "right": "issue:42617", + "accept": false, + "reason": "Qwen3 auto device mapping cuda assert is a different failure from the 3d_parallel.py runner problem." + }, + { + "left": "issue:43531", + "right": "issue:43572", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and StableLmConfig pad_token_idx missing are unrelated configuration bugs." + }, + { + "left": "issue:44292", + "right": "issue:44479", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error is unrelated to the Qwen2_5_vl / qwen3_vl video-input regression." + }, + { + "left": "issue:41720", + "right": "issue:45161", + "accept": false, + "reason": "Same broad GPT/Qwen MoE area, but the concrete failures differ: auto device mapping cuda assert versus tensor-parallel-only not working." + }, + { + "left": "issue:43388", + "right": "issue:45059", + "accept": false, + "reason": "Metric-label dropping in evaluation is unrelated to SAM3 PCS text/bounding-box behavior." + }, + { + "left": "issue:44186", + "right": "issue:44361", + "accept": false, + "reason": "Different tokenizer crashes for different models: LayoutLMv2Tokenizer NER/padding issue versus MLukeTokenizer task AttributeError." + }, + { + "left": "issue:43540", + "right": "issue:43782", + "accept": false, + "reason": "Qwen3OmniMoe video-input ValueError and Qwen3VL weight_only loading error are distinct model-loading/processing bugs." + }, + { + "left": "issue:41720", + "right": "issue:43873", + "accept": false, + "reason": "Both touch Qwen model execution/offloading, but one is a cuda assert with auto device mapping and the other is quantization offloading behavior." + }, + { + "left": "issue:45406", + "right": "issue:45464", + "accept": false, + "reason": "Both are serve/API crashes, but Gemma4Processor missing _tokenizer is unrelated to Qwen3.5 streaming inference failure." + }, + { + "left": "issue:42915", + "right": "issue:43322", + "accept": false, + "reason": "Qwen3Moe FineGrainedFP8Config failure and Llava Next segmentation fault are different models and different crash modes." + }, + { + "left": "issue:36010", + "right": "issue:42617", + "accept": false, + "reason": "GenerationMixin import error is unrelated to running 3d_parallel.py." + }, + { + "left": "issue:36331", + "right": "issue:41720", + "accept": false, + "reason": "Trainer custom-loss signature mismatch is unrelated to Qwen3 auto device mapping cuda errors." + }, + { + "left": "issue:44514", + "right": "issue:44545", + "accept": true, + "reason": "Exact same bug: Qwen2_5_VLProcessor.apply_chat_template crashes on batched input when padding=False; these are duplicate reports of the same path." + }, + { + "left": "issue:33357", + "right": "issue:35707", + "accept": false, + "reason": "MacOS bus error on CLIP model loading is unrelated to progressive generation with inputs_embeds and past_key_values." + }, + { + "left": "issue:43606", + "right": "issue:43720", + "accept": false, + "reason": "Different offload/loader bugs: bark-small CPU offload device mismatch versus BitNet packed-weights unpacking during accelerate loading." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet, and the duplicate call on 44514/44545 is well supported. The other pair verdicts are conservative and the distinctions are specific to the titles/body context provided." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 11, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:30064", + "issue:34567", + "issue:34634", + "issue:35707", + "issue:39692", + "issue:41720", + "issue:42886", + "issue:43381", + "issue:43421", + "issue:43531", + "issue:43550", + "issue:43596", + "issue:43606", + "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:44492", + "issue:44509", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44661", + "issue:44671", + "issue:44843", + "issue:44898", + "issue:44918", + "issue:44936", + "issue:44987", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45183", + "issue:45356", + "issue:45357", + "issue:45381", + "issue:45479" + ], + "soft_pairs": [ + "issue:41720|issue:43746", + "issue:30064|issue:34634", + "issue:43550|issue:43653", + "issue:44368|issue:45357", + "issue:44625|issue:45479", + "issue:44492|issue:44509", + "issue:43606|issue:43653", + "issue:43421|issue:43927", + "issue:34634|issue:41720", + "issue:44898|issue:45072", + "issue:43782|issue:45381", + "issue:44451|issue:44991", + "issue:44062|issue:44451", + "issue:34567|issue:35707", + "issue:43381|issue:43844", + "issue:42886|issue:44843", + "issue:44077|issue:44625", + "issue:43421|issue:43653", + "issue:44918|issue:44936", + "issue:43531|issue:44589", + "issue:44987|issue:45042", + "issue:43994|issue:44671", + "issue:39692|issue:44610", + "issue:43421|issue:44186", + "issue:43596|issue:44220", + "issue:43531|issue:43645", + "issue:45081|issue:45356", + "issue:43746|issue:45381", + "issue:45183|issue:45479", + "issue:39692|issue:43873", + "issue:44220|issue:44661" + ], + "prepared_review_unit_hash": "1a20c543544ef84e329bc5ca266c0f232f714acd92d0ca23402ae0569d84ea8e", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37251, + "estimated_input_tokens": 9313, + "estimated_eval_tokens": 18882 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous: most items are unrelated tokenizer, model-loading, training, vision, or docs bugs that only share broad library vocabulary. None of the soft edges looks like the same underlying issue or a mergeable PR pair.", + "confidence": 0.96, + "canonical_issue_reason": "No clear canonical issue: the items span many unrelated code paths, and the soft edges are only superficial title similarity rather than one underlying bug.", + "canonical_pr_reason": null, + "best_issue_reason": "Weak fit at best: issue 43421 is the broadest tokenizer-internals report, but it still does not represent the rest of the cluster well.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41720", + "right": "issue:43746", + "accept": false, + "reason": "Different failures: Qwen3 device-map CUDA assert vs GraniteSpeech PEFT local-checkpoint loading." + }, + { + "left": "issue:30064", + "right": "issue:34634", + "accept": false, + "reason": "Unrelated domains: image segmentation-map processing vs Bark voice preset handling." + }, + { + "left": "issue:43550", + "right": "issue:43653", + "accept": false, + "reason": "Different subsystems: torch.compile/SDPA model bug vs tokenizer special-token registration/decode issue." + }, + { + "left": "issue:44368", + "right": "issue:45357", + "accept": false, + "reason": "Both involve Qwen3.5, but one is a tied-embeddings warning and the other is a save_pretrained serialization bug; not the same code path." + }, + { + "left": "issue:44625", + "right": "issue:45479", + "accept": false, + "reason": "Both touch classification config, but one is num_labels propagation and the other is zero-loss behavior from problem_type handling." + }, + { + "left": "issue:44492", + "right": "issue:44509", + "accept": false, + "reason": "A typo in cache-strategy docs is unrelated to removed pipeline-task documentation." + }, + { + "left": "issue:43606", + "right": "issue:43653", + "accept": false, + "reason": "CPU offload device mismatch and BigBirdTokenizer special-token decode are separate bugs." + }, + { + "left": "issue:43421", + "right": "issue:43927", + "accept": false, + "reason": "Both concern tokenization/config mutation, but one is runtime post-processor refresh and the other is save/load custom token IDs." + }, + { + "left": "issue:34634", + "right": "issue:41720", + "accept": false, + "reason": "BarkProcessor voice_preset and Qwen3 auto device mapping are unrelated." + }, + { + "left": "issue:44898", + "right": "issue:45072", + "accept": false, + "reason": "Different model bugs: Perceiver image-size handling vs dtype mismatch in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:43782", + "right": "issue:45381", + "accept": false, + "reason": "Both are multimodal Qwen bugs, but one is from_pretrained/weight_only loading and the other is video vision_position_ids." + }, + { + "left": "issue:44451", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer-loading regressions, but for different models and different failure mechanisms." + }, + { + "left": "issue:44062", + "right": "issue:44451", + "accept": false, + "reason": "AddedToken special-arg duplication is not the same bug as loading ScandiBERT." + }, + { + "left": "issue:34567", + "right": "issue:35707", + "accept": false, + "reason": "Trainer token accounting and progressive generation with inputs_embeds/past_key_values are unrelated." + }, + { + "left": "issue:43381", + "right": "issue:43844", + "accept": false, + "reason": "Gradient checkpointing in eval mode and abnormal gradients under ZeRO-3 are distinct training issues." + }, + { + "left": "issue:42886", + "right": "issue:44843", + "accept": false, + "reason": "Both break HF_HUB_OFFLINE, but one is cache loading and the other is an unconditional model_info call in Mistral regex patching." + }, + { + "left": "issue:44077", + "right": "issue:44625", + "accept": false, + "reason": "PatchTSMixer post_init policy and Qwen3.5 num_labels propagation are unrelated." + }, + { + "left": "issue:43421", + "right": "issue:43653", + "accept": false, + "reason": "Both mention special tokens, but one is runtime post-processor syncing and the other is a model-specific tokenizer registration bug." + }, + { + "left": "issue:44918", + "right": "issue:44936", + "accept": false, + "reason": "Trainer bugs, but different ones: SFT inputs_embeds unpacking vs evaluate-after-train failure." + }, + { + "left": "issue:43531", + "right": "issue:44589", + "accept": false, + "reason": "Sliding-window Qwen3-MoE behavior and Float8 storage lookup failure are unrelated." + }, + { + "left": "issue:44987", + "right": "issue:45042", + "accept": false, + "reason": "Loading a specific model and PIL image processors requiring torchvision are separate issues." + }, + { + "left": "issue:43994", + "right": "issue:44671", + "accept": false, + "reason": "Different models and different regressions: SigLIP2 output quality vs CamemBERT masked-LM predictions." + }, + { + "left": "issue:39692", + "right": "issue:44610", + "accept": false, + "reason": "Both are vision-model processor mismatches, but one is a docs/quantization example and the other is OmDet-Turbo size mismatch." + }, + { + "left": "issue:43421", + "right": "issue:44186", + "accept": false, + "reason": "TokenizerBackend post-processor updates and LayoutLMv2 NER/padding crashes are separate tokenizer bugs." + }, + { + "left": "issue:43596", + "right": "issue:44220", + "accept": false, + "reason": "Deepspeed/BERT initialization and feature-extraction helper failure are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43645", + "accept": false, + "reason": "Qwen3-MoE sliding_window bug does not match the Jupyter custom-model initialization regression." + }, + { + "left": "issue:45081", + "right": "issue:45356", + "accept": false, + "reason": "Both involve Mistral regex/tokenizer handling, but one crashes on backend_tokenizer absence while the other is a Kimi-K2.5 codec regression." + }, + { + "left": "issue:43746", + "right": "issue:45381", + "accept": false, + "reason": "GraniteSpeech PEFT checkpoint loading and Qwen2.5-VL video position IDs are unrelated." + }, + { + "left": "issue:45183", + "right": "issue:45479", + "accept": false, + "reason": "Both are validation/labeling bugs, but one is XOR input validation messaging and the other is degenerate loss for num_labels=1." + }, + { + "left": "issue:39692", + "right": "issue:43873", + "accept": false, + "reason": "SigLIP2 docs/quantization issues and quantized offloading behavior are not the same bug." + }, + { + "left": "issue:44220", + "right": "issue:44661", + "accept": false, + "reason": "Audio feature extraction and add-new-model-like tokenizer mapping 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": 2, + "component_count": 13, + "review_unit_index": 12, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:30064", + "issue:36010", + "issue:40444", + "issue:43335", + "issue:43421", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43582", + "issue:43618", + "issue:43643", + "issue:43644", + "issue:43653", + "issue:43697", + "issue:43716", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43818", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43873", + "issue:43927", + "issue:43994", + "issue:44008", + "issue:44062", + "issue:44186", + "issue:44442", + "issue:44448", + "issue:44488", + "issue:44561", + "issue:44671", + "issue:44811", + "issue:44821", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44977", + "issue:45020", + "issue:45072", + "issue:45081", + "issue:45084", + "issue:45381" + ], + "soft_pairs": [ + "issue:43540|issue:44008", + "issue:43746|issue:43873", + "issue:43618|issue:43697", + "issue:40444|issue:43746", + "issue:44843|issue:45081", + "issue:43716|issue:44008", + "issue:43720|issue:43819", + "issue:43653|issue:43720", + "issue:43335|issue:43756", + "issue:44062|issue:44488", + "issue:30064|issue:36010", + "issue:44186|issue:44898", + "issue:43421|issue:44811", + "issue:43550|issue:45072", + "issue:43540|issue:45381", + "issue:43577|issue:43818", + "issue:43550|issue:43927", + "issue:43643|issue:44877", + "issue:43577|issue:43582", + "issue:43531|issue:43643", + "issue:43782|issue:44977", + "issue:44448|issue:44671", + "issue:43644|issue:43749", + "issue:44561|issue:45020", + "issue:43653|issue:43819", + "issue:43825|issue:43994", + "issue:44821|issue:45084", + "issue:43550|issue:43819", + "issue:44186|issue:44442", + "issue:43825|issue:43827" + ], + "prepared_review_unit_hash": "6e5f4e11276b36489bdc69578675c01afec2389cc2263536b3e02e692685f432", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36284, + "estimated_input_tokens": 9071, + "estimated_eval_tokens": 18398 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The soft-edge candidates are mostly near-miss similarities rather than true duplicates: they share broad subsystems (tokenizers, multimodal models, loading, v5 regressions) but not the same concrete bug or fix path.", + "confidence": 0.91, + "canonical_issue_reason": "No single canonical issue cleanly covers this cluster; the items are heterogeneous. If forced to pick a representative, issue:45020 is the broadest umbrella, but it only reflects the remote_code regression subset.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best umbrella-style representative because it summarizes the broadest recent-version regression theme, but it still does not subsume the unrelated tokenizer, multimodal, and loading bugs in the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43540", + "right": "issue:44008", + "accept": false, + "reason": "Different models and failures: Qwen3OmniMoe video-input processing vs Gemma 3n forward variable collision." + }, + { + "left": "issue:43746", + "right": "issue:43873", + "accept": false, + "reason": "Both are loading-related, but one is PEFT/local checkpoint loading and the other is quantization offloading behavior; not the same bug." + }, + { + "left": "issue:43618", + "right": "issue:43697", + "accept": false, + "reason": "CLIP attentions field regression and RTDetr output drift are unrelated code paths and fixes." + }, + { + "left": "issue:40444", + "right": "issue:43746", + "accept": false, + "reason": "IterableDataset multi-image finetuning failure is unrelated to GraniteSpeech PEFT checkpoint loading." + }, + { + "left": "issue:44843", + "right": "issue:45081", + "accept": false, + "reason": "Both touch _patch_mistral_regex, but one is an offline model_info call and the other is a backend_tokenizer attribute crash; different concrete bugs." + }, + { + "left": "issue:43716", + "right": "issue:44008", + "accept": false, + "reason": "Blip/Mistral dtype mismatch and Gemma3n variable collision are unrelated." + }, + { + "left": "issue:43720", + "right": "issue:43819", + "accept": false, + "reason": "BitNet packed-weight loading and DAC from_latents/STE mismatch are different model-specific issues." + }, + { + "left": "issue:43653", + "right": "issue:43720", + "accept": false, + "reason": "BigBirdTokenizer special-token decode issue is unrelated to BitNet accelerate loading." + }, + { + "left": "issue:43335", + "right": "issue:43756", + "accept": false, + "reason": "Different models and mechanisms: SwitchTransformers sparse-layer config vs Smollm3 RoPE-layer dropping." + }, + { + "left": "issue:44062", + "right": "issue:44488", + "accept": false, + "reason": "Tokenizer AddedToken special-arg duplication is not the same as failing to load cjvt/sleng-bert." + }, + { + "left": "issue:30064", + "right": "issue:36010", + "accept": false, + "reason": "ImageProcessor segmentation-map handling and GenerationMixin import failure are unrelated." + }, + { + "left": "issue:44186", + "right": "issue:44898", + "accept": false, + "reason": "LayoutLMv2 tokenizer padding/NER crash and Perceiver interpolation failure are different bugs." + }, + { + "left": "issue:43421", + "right": "issue:44811", + "accept": false, + "reason": "Runtime special-token post-processor update and Whisper batch_decode skip_special_tokens are separate tokenizer/processor behaviors." + }, + { + "left": "issue:43550", + "right": "issue:45072", + "accept": false, + "reason": "Different failures in different models: torch.compile+SDPA vs bfloat16 dtype mismatch." + }, + { + "left": "issue:43540", + "right": "issue:45381", + "accept": false, + "reason": "Both involve video inputs, but the models and reported failures differ (Qwen3OmniMoe processing vs Qwen2.5-VL vision_position_ids)." + }, + { + "left": "issue:43577", + "right": "issue:43818", + "accept": false, + "reason": "BLIP2 dtype retention and Video-LLaVA temporal-attention/weight-sharing issues are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43927", + "accept": false, + "reason": "Bamba torch.compile/SDPA failure is unrelated to DiaConfig custom token ID loss on save/load." + }, + { + "left": "issue:43643", + "right": "issue:44877", + "accept": false, + "reason": "AutoConfig missing fields with trust_remote_code and strict granite_speech config loading are different config-path issues." + }, + { + "left": "issue:43577", + "right": "issue:43582", + "accept": false, + "reason": "Model dtype retention and Apple Silicon caching_allocator_warmup TypeError are unrelated." + }, + { + "left": "issue:43531", + "right": "issue:43643", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and trust_remote_code AutoConfig fields are unrelated." + }, + { + "left": "issue:43782", + "right": "issue:44977", + "accept": false, + "reason": "Qwen3VL weight_only load error and Qwen3.5 flash-attention generation issue are different failures." + }, + { + "left": "issue:44448", + "right": "issue:44671", + "accept": false, + "reason": "Both are v4/v5 output regressions, but they concern different models and tasks." + }, + { + "left": "issue:43644", + "right": "issue:43749", + "accept": false, + "reason": "Non-persistent buffer initialization junk and FSDP CPU RAM efficient loading are separate loading/state issues." + }, + { + "left": "issue:44561", + "right": "issue:45020", + "accept": false, + "reason": "45020 is a broad umbrella around remote_code regressions; 44561 is one specific trust_remote_code breakage from a removed helper, not the same issue." + }, + { + "left": "issue:43653", + "right": "issue:43819", + "accept": false, + "reason": "BigBirdTokenizer special-token decode and DAC latent reconstruction mismatch are unrelated." + }, + { + "left": "issue:43825", + "right": "issue:43994", + "accept": false, + "reason": "Pipeline translation error messaging and SigLIP2 nonsensical outputs are unrelated." + }, + { + "left": "issue:44821", + "right": "issue:45084", + "accept": false, + "reason": "AutoImageProcessor URL loading and template-node compilation errors are unrelated." + }, + { + "left": "issue:43550", + "right": "issue:43819", + "accept": false, + "reason": "Different model families and different code paths: SDPA compile failure vs DAC latent decoding mismatch." + }, + { + "left": "issue:44186", + "right": "issue:44442", + "accept": false, + "reason": "LayoutLMv2 tokenizer crash and AutoTokenizer failing to load FastSpeech2ConformerTokenizer are different tokenizer issues." + }, + { + "left": "issue:43825", + "right": "issue:43827", + "accept": false, + "reason": "Both are about pipeline deprecation fallout, but one is a bad runtime error message and the other is outdated docs; separate fixes." + } + ] + }, + "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": 13, + "review_unit_index": 13, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:30064", + "issue:34634", + "issue:36010", + "issue:36331", + "issue:41720", + "issue:42175", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43452", + "issue:43454", + "issue:43526", + "issue:43577", + "issue:43582", + "issue:43606", + "issue:43638", + "issue:43650", + "issue:43653", + "issue:43697", + "issue:43720", + "issue:43746", + "issue:43749", + "issue:43782", + "issue:43819", + "issue:43906", + "issue:44060", + "issue:44291", + "issue:44368", + "issue:44485", + "issue:44509", + "issue:44554", + "issue:44568", + "issue:44589", + "issue:44610", + "issue:44671", + "issue:44805", + "issue:44811", + "issue:44857", + "issue:44912", + "issue:44936", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45081", + "issue:45325", + "issue:45406", + "issue:45488" + ], + "soft_pairs": [ + "issue:43650|issue:43906", + "issue:41720|issue:44991", + "issue:44805|issue:44936", + "issue:45081|issue:45406", + "issue:30064|issue:36331", + "issue:42175|issue:45081", + "issue:43606|issue:45072", + "issue:44610|issue:45072", + "issue:43782|issue:45325", + "issue:44485|issue:44509", + "issue:34634|issue:36010", + "issue:44060|issue:44368", + "issue:43720|issue:45072", + "issue:43262|issue:43819", + "issue:43582|issue:44857", + "issue:43749|issue:44291", + "issue:43653|issue:44671", + "issue:43526|issue:43697", + "issue:43526|issue:44554", + "issue:43653|issue:44912", + "issue:43454|issue:43577", + "issue:44568|issue:45488", + "issue:43452|issue:44991", + "issue:34634|issue:36331", + "issue:43638|issue:44936", + "issue:43257|issue:43329", + "issue:43746|issue:44811", + "issue:44589|issue:45042" + ], + "prepared_review_unit_hash": "185078575abf880b9b189dca2165dfef5743d6fad9c4ce2f230db3e63c3d21ca", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 28, + "serialized_chars": 36088, + "estimated_input_tokens": 9022, + "estimated_eval_tokens": 18300 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These are mostly unrelated bug reports spanning tokenizers, model loading, processors, training, and backend-specific failures. None of the soft pairs look like the same underlying bug/change closely enough to merge, so all soft edges should be rejected. The clearest standalone issue is #45081, which has a concrete traceback and a specific tokenizer code path.", + "confidence": 0.91, + "canonical_issue_reason": "#45081 is the most canonical issue in this set: it is specific, has an explicit traceback, names the exact failing function and flag, and is scoped to a concrete tokenizer loading path.", + "canonical_pr_reason": null, + "best_issue_reason": "#45081 is the strongest issue candidate because it is actionable, well-scoped, and tied to a reproducible AttributeError in a single code path.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43650", + "right": "issue:43906", + "accept": false, + "reason": "Different problems: a vague content issue vs an isolated reproduction of another bug; no shared code path or fix." + }, + { + "left": "issue:41720", + "right": "issue:44991", + "accept": false, + "reason": "One is CUDA assert during auto device mapping for Qwen3, the other is tokenizer loading for EMBEDDIA/est-roberta; unrelated failures." + }, + { + "left": "issue:44805", + "right": "issue:44936", + "accept": false, + "reason": "Mask/tensor shape mismatch in one report versus trainer evaluate-after-train behavior in the other; not the same bug." + }, + { + "left": "issue:45081", + "right": "issue:45406", + "accept": false, + "reason": "Both are processor/tokenizer errors, but they fail in different classes and code paths: Mistral regex patching vs Gemma4Processor missing _tokenizer." + }, + { + "left": "issue:30064", + "right": "issue:36331", + "accept": false, + "reason": "Image segmentation map processing and CustomTrainer.compute_loss signature mismatch are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:45081", + "accept": false, + "reason": "Package/backend dependency issue versus a tokenizer patch crash; no common concrete bug." + }, + { + "left": "issue:43606", + "right": "issue:45072", + "accept": false, + "reason": "CPU offload device mismatch for Bark-small is unrelated to bfloat16 dtype mismatches in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:44610", + "right": "issue:45072", + "accept": false, + "reason": "Different domains and failures: OmDet-Turbo image size mismatch versus dtype mismatch in inference." + }, + { + "left": "issue:43782", + "right": "issue:45325", + "accept": false, + "reason": "Both involve Qwen/VL models, but one is a from_pretrained weight_only error and the other is a RoPE position-id scaling bug." + }, + { + "left": "issue:44485", + "right": "issue:44509", + "accept": false, + "reason": "GLM-5 RoPE implementation bug and docs mentioning removed pipeline tasks are separate issues." + }, + { + "left": "issue:34634", + "right": "issue:36010", + "accept": false, + "reason": "BarkProcessor voice_preset bug is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:44060", + "right": "issue:44368", + "accept": false, + "reason": "Both mention tied-weights warnings, but they concern different models and different symptoms; not a single concrete bug." + }, + { + "left": "issue:43720", + "right": "issue:45072", + "accept": false, + "reason": "BitNet accelerate-loading unpacking bug is unrelated to dtype mismatches in other models." + }, + { + "left": "issue:43262", + "right": "issue:43819", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting and DAC.from_latents vs forward mismatch are separate code paths." + }, + { + "left": "issue:43582", + "right": "issue:44857", + "accept": false, + "reason": "Apple Silicon TypeError in caching allocator warmup is unrelated to CUDA AMP crashes in LwDetrImageLoss." + }, + { + "left": "issue:43749", + "right": "issue:44291", + "accept": false, + "reason": "Both touch loading, but one is FSDP CPU RAM efficient loading and the other is init_empty_weights passing an unexpected argument; not the same bug." + }, + { + "left": "issue:43653", + "right": "issue:44671", + "accept": false, + "reason": "Tokenizer special-token registration and masked LM prediction drift are different issues." + }, + { + "left": "issue:43526", + "right": "issue:43697", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug does not match RTDetrV2 output differences in v5." + }, + { + "left": "issue:43526", + "right": "issue:44554", + "accept": false, + "reason": "Image processor label reduction and an MPS attention correctness issue are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:44912", + "accept": false, + "reason": "BigBirdTokenizer decode bug and MXFP4 quantization loading fallback are different code paths." + }, + { + "left": "issue:43454", + "right": "issue:43577", + "accept": false, + "reason": "AyaVision lm_head tying and Blip2 dtype propagation are separate model-loading problems." + }, + { + "left": "issue:44568", + "right": "issue:45488", + "accept": false, + "reason": "Tokenizer special-token handling for mdeberta-v3-base and LlamaTokenizer pre-tokenizer override for DeepSeek are not the same fix." + }, + { + "left": "issue:43452", + "right": "issue:44991", + "accept": false, + "reason": "gguf_file loading breakage and tokenizer loading failure for a specific model are unrelated." + }, + { + "left": "issue:43638", + "right": "issue:44936", + "accept": false, + "reason": "DeepSpeed ZeRO-3 + non-pretrained Bert index error is not the same as evaluate-after-train behavior." + }, + { + "left": "issue:43746", + "right": "issue:44811", + "accept": false, + "reason": "PEFT local checkpoint loading and Whisper batch_decode skip_special_tokens handling are unrelated." + }, + { + "left": "issue:44589", + "right": "issue:45042", + "accept": false, + "reason": "Float8 storage lookup failure is unrelated to PIL backend processors requiring torchvision." + } + ] + }, + "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": 13, + "review_unit_index": 14, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:33357", + "issue:34567", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:41720", + "issue:42175", + "issue:42915", + "issue:43257", + "issue:43262", + "issue:43329", + "issue:43450", + "issue:43452", + "issue:43531", + "issue:43540", + "issue:43550", + "issue:43577", + "issue:43582", + "issue:43645", + "issue:43696", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43819", + "issue:43854", + "issue:43927", + "issue:43931", + "issue:44062", + "issue:44291", + "issue:44479", + "issue:44560", + "issue:44568", + "issue:44610", + "issue:44756", + "issue:44811", + "issue:44821", + "issue:44912", + "issue:44938", + "issue:44977", + "issue:44991", + "issue:45042", + "issue:45072", + "issue:45356", + "issue:45357", + "issue:45381" + ], + "soft_pairs": [ + "issue:43854|issue:44977", + "issue:38175|issue:44610", + "issue:39692|issue:44912", + "issue:36010|issue:43645", + "issue:43257|issue:43696", + "issue:43819|issue:43927", + "issue:43262|issue:44811", + "issue:43720|issue:44610", + "issue:43550|issue:44912", + "issue:44568|issue:45356", + "issue:42175|issue:44938", + "issue:43749|issue:44756", + "issue:43531|issue:45042", + "issue:43329|issue:44479", + "issue:36010|issue:36331", + "issue:43257|issue:43931", + "issue:43257|issue:43756", + "issue:43329|issue:43450", + "issue:43577|issue:45072", + "issue:43697|issue:45357", + "issue:43452|issue:43582", + "issue:41720|issue:43540", + "issue:44479|issue:45381", + "issue:44291|issue:45042", + "issue:44062|issue:44991", + "issue:43540|issue:44560", + "issue:42915|issue:43742", + "issue:43927|issue:44912", + "issue:43540|issue:44479", + "issue:42915|issue:44821", + "issue:33357|issue:34567" + ], + "prepared_review_unit_hash": "b2d1458166ec460135b1a93115be94b3cf7ef1da0944bc385202760abd6d2bbc", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37099, + "estimated_input_tokens": 9275, + "estimated_eval_tokens": 18806 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most candidate pairs are unrelated and should stay separate. The only clear overlaps are the Qwen video-input regressions around v5.3.0, which appear to share the same underlying multimodal/video code-path issue.", + "confidence": 0.76, + "canonical_issue_reason": "issue:44479 is the broadest and clearest report of the shared Qwen video-input regression, covering multiple affected Qwen VL variants and matching the other video-related soft edges.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44479 is the best canonical issue because it describes the user-facing regression most generally and is broad enough to subsume the narrower video-position/tokenization reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43854", + "right": "issue:44977", + "accept": false, + "reason": "Different models and symptoms; CLIP load crash vs Qwen3.5 flash-attention generation bug." + }, + { + "left": "issue:38175", + "right": "issue:44610", + "accept": false, + "reason": "Unrelated subsystems: SigLIP2 probability issue vs OmDet-Turbo image-size mismatch." + }, + { + "left": "issue:39692", + "right": "issue:44912", + "accept": false, + "reason": "Documentation/example errors vs MXFP4 quantization fallback; not the same bug." + }, + { + "left": "issue:36010", + "right": "issue:43645", + "accept": false, + "reason": "Both are version-5 breakages, but one is a GenerationMixin import error and the other is custom-model notebook init behavior." + }, + { + "left": "issue:43257", + "right": "issue:43696", + "accept": false, + "reason": "Qwen3 MOE weight conversion/loading bug vs GPT-oss OOM; different failure modes." + }, + { + "left": "issue:43819", + "right": "issue:43927", + "accept": false, + "reason": "Different components entirely: DAC latent conversion vs DiaConfig token-id persistence." + }, + { + "left": "issue:43262", + "right": "issue:44811", + "accept": false, + "reason": "Audio chat-template sample-rate default vs Whisper batch_decode skip_special_tokens handling." + }, + { + "left": "issue:43720", + "right": "issue:44610", + "accept": false, + "reason": "Accelerate packed-weight loading bug vs OmDet processor input-size mismatch." + }, + { + "left": "issue:43550", + "right": "issue:44912", + "accept": false, + "reason": "torch.compile/SDPA runtime bug vs quantization fallback; different code paths." + }, + { + "left": "issue:44568", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer-related, but the titles indicate different models and different tokenizer regressions." + }, + { + "left": "issue:42175", + "right": "issue:44938", + "accept": false, + "reason": "Packaging/backend dependency issue vs Python 3.14 import failure." + }, + { + "left": "issue:43749", + "right": "issue:44756", + "accept": false, + "reason": "Both involve loading/memory, but one is FSDP CPU RAM efficient loading and the other is mmap/OOM on Strix Halo." + }, + { + "left": "issue:43531", + "right": "issue:45042", + "accept": false, + "reason": "Qwen3-MoE sliding window bug vs PIL backend torchvision dependency regression." + }, + { + "left": "issue:43329", + "right": "issue:44479", + "accept": true, + "reason": "Both point to the same Qwen video-input/multimodal-token regression in the video code path; one is the specific undefined-symbol bug, the other the broader downstream regression report." + }, + { + "left": "issue:36010", + "right": "issue:36331", + "accept": false, + "reason": "Both are API breakages, but one is GenerationMixin import and the other is compute_loss signature mismatch." + }, + { + "left": "issue:43257", + "right": "issue:43931", + "accept": false, + "reason": "Different Qwen3 model-loading failures with different symptoms and causes." + }, + { + "left": "issue:43257", + "right": "issue:43756", + "accept": false, + "reason": "Qwen3 MOE loading vs Smollm3 RoPE layer dropping; unrelated." + }, + { + "left": "issue:43329", + "right": "issue:43450", + "accept": false, + "reason": "Both are video-related, but one is token counting in multimodal utilities and the other is a batched video processor shape bug." + }, + { + "left": "issue:43577", + "right": "issue:45072", + "accept": false, + "reason": "Different dtype mismatch regressions in different model families." + }, + { + "left": "issue:43697", + "right": "issue:45357", + "accept": false, + "reason": "RTDetrV2 output regression vs Qwen3.5 save_pretrained visual-key regression; unrelated." + }, + { + "left": "issue:43452", + "right": "issue:43582", + "accept": false, + "reason": "Tokenizer/model gguf loading issue vs Apple Silicon allocator warmup TypeError." + }, + { + "left": "issue:41720", + "right": "issue:43540", + "accept": false, + "reason": "Qwen3 auto device-mapping CUDA assert vs Qwen3OmniMoe video ValueError; different bugs." + }, + { + "left": "issue:44479", + "right": "issue:45381", + "accept": true, + "reason": "Both are Qwen2.5-VL/Qwen video-input regressions; 45381 looks like a narrower manifestation of the same bad video position handling." + }, + { + "left": "issue:44291", + "right": "issue:45042", + "accept": false, + "reason": "init_empty_weights argument handling vs PIL/torchvision dependency issue." + }, + { + "left": "issue:44062", + "right": "issue:44991", + "accept": false, + "reason": "Both are tokenizer loading problems, but the titles do not show the same concrete failure or model/config." + }, + { + "left": "issue:43540", + "right": "issue:44560", + "accept": false, + "reason": "Different Qwen model variants and different video failures; not clearly the same bug." + }, + { + "left": "issue:42915", + "right": "issue:43742", + "accept": false, + "reason": "FineGrainedFP8Config issue vs MobileLLM key error; unrelated." + }, + { + "left": "issue:43927", + "right": "issue:44912", + "accept": false, + "reason": "Custom token ID persistence vs quantization fallback; unrelated." + }, + { + "left": "issue:43540", + "right": "issue:44479", + "accept": false, + "reason": "Related by model family and video inputs, but the reported failures and model targets are different enough to keep separate." + }, + { + "left": "issue:42915", + "right": "issue:44821", + "accept": false, + "reason": "Unrelated: FP8 Qwen3Moe loading vs AutoImageProcessor-from-URL loading." + }, + { + "left": "issue:33357", + "right": "issue:34567", + "accept": false, + "reason": "MacOS bus error on CLIP vs TrainerState token counter not updating; unrelated." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The canonical issue choice is supported by the titles, and the only accepted soft-edge links are the clearly video-regression-related Qwen pairs, which look plausibly mergeable into a single bugfix. No obvious overstatement stands out." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 15, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:28282", + "issue:29127", + "issue:34567", + "issue:34689", + "issue:36010", + "issue:36331", + "issue:38175", + "issue:42175", + "issue:43452", + "issue:43493", + "issue:43526", + "issue:43531", + "issue:43577", + "issue:43582", + "issue:43643", + "issue:43646", + "issue:43688", + "issue:43697", + "issue:43742", + "issue:43749", + "issue:43756", + "issue:43782", + "issue:43824", + "issue:43906", + "issue:43994", + "issue:44062", + "issue:44291", + "issue:44484", + "issue:44488", + "issue:44514", + "issue:44561", + "issue:44589", + "issue:44704", + "issue:44792", + "issue:44821", + "issue:44843", + "issue:44898", + "issue:44964", + "issue:45042", + "issue:45081", + "issue:45137", + "issue:45216", + "issue:45290", + "issue:45325", + "issue:45375", + "issue:45381", + "issue:45405", + "issue:45479" + ], + "soft_pairs": [ + "issue:42175|issue:43824", + "issue:45137|issue:45405", + "issue:42175|issue:44488", + "issue:43906|issue:45381", + "issue:43526|issue:45479", + "issue:38175|issue:44898", + "issue:43582|issue:43782", + "issue:44484|issue:44704", + "issue:28282|issue:34689", + "issue:43646|issue:43749", + "issue:36010|issue:42175", + "issue:43742|issue:44291", + "issue:43493|issue:43994", + "issue:43643|issue:45375", + "issue:28282|issue:29127", + "issue:43756|issue:45325", + "issue:44062|issue:45081", + "issue:44291|issue:44821", + "issue:43452|issue:44488", + "issue:34567|issue:36331", + "issue:44792|issue:44964", + "issue:43688|issue:43749", + "issue:44514|issue:45290", + "issue:43742|issue:44589", + "issue:43452|issue:43577", + "issue:43643|issue:44561", + "issue:45042|issue:45081", + "issue:43531|issue:44821", + "issue:44514|issue:45381", + "issue:43697|issue:45216", + "issue:43452|issue:44843" + ], + "prepared_review_unit_hash": "6f74cd32eb2dd254ef3829a0a33f4c068071ea5c66d70804a2b23130d23aa0bb", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37287, + "estimated_input_tokens": 9322, + "estimated_eval_tokens": 18900 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster remains heterogeneous overall: most items are unrelated loading, dependency, training, or model-specific bugs. I only accept one soft-edge duplicate pair, the two SigLIP2 reports, because they point to the same model-family output regression. The other soft edges are rejected as separate failure modes, with the Qwen2.5-VL and tool-call pairs kept conservative and grounded in the packet text.", + "confidence": 0.95, + "canonical_issue_reason": "No single issue cleanly represents the cluster; the items span many distinct code paths and bug classes, so any one issue would be a poor canonical for the whole set.", + "canonical_pr_reason": null, + "best_issue_reason": "No global best issue: the only clear local duplicate is the SigLIP2 pair, but the cluster as a whole is too mixed for one issue to serve as a reliable representative.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42175", + "right": "issue:43824", + "accept": false, + "reason": "Backend packaging/dependency problem vs missing Qwen2.5-VL symbol import; not the same bug." + }, + { + "left": "issue:45137", + "right": "issue:45405", + "accept": false, + "reason": "DeepSpeed ZeRO3 deque failure is unrelated to a PEFT version pin regression." + }, + { + "left": "issue:42175", + "right": "issue:44488", + "accept": false, + "reason": "Torch backend packaging issue vs a specific model-loading failure; different failures." + }, + { + "left": "issue:43906", + "right": "issue:45381", + "accept": false, + "reason": "The packet only shows an isolated repro and a Qwen2.5-VL video position_ids bug; no grounded evidence they are the same defect." + }, + { + "left": "issue:43526", + "right": "issue:45479", + "accept": false, + "reason": "BEiT fast processor label-reduction bug vs sequence-classification zero-loss bug; different code paths." + }, + { + "left": "issue:38175", + "right": "issue:44898", + "accept": false, + "reason": "Both are vision-model related, but the model and symptom are different enough that they do not look like one bug." + }, + { + "left": "issue:43582", + "right": "issue:43782", + "accept": false, + "reason": "Apple Silicon allocator warmup TypeError is unrelated to Qwen3VL weight_only loading." + }, + { + "left": "issue:44484", + "right": "issue:44704", + "accept": false, + "reason": "save_pretrained shard sizing vs AutoProcessor kwargs forwarding are separate issues." + }, + { + "left": "issue:28282", + "right": "issue:34689", + "accept": false, + "reason": "Missing PyTorch ImportError is not the same as a Llama 3.2 Vision loading regression." + }, + { + "left": "issue:43646", + "right": "issue:43749", + "accept": false, + "reason": "Custom model initialization breakage and FSDP CPU RAM-efficient loading are different code paths." + }, + { + "left": "issue:36010", + "right": "issue:42175", + "accept": false, + "reason": "GenerationMixin import failure and TensorFlow backend packaging are unrelated." + }, + { + "left": "issue:43742", + "right": "issue:44291", + "accept": false, + "reason": "MobileLLM key error and init_empty_weights/_is_hf_initialized mismatch are separate problems." + }, + { + "left": "issue:43493", + "right": "issue:43994", + "accept": true, + "reason": "Both reports describe the same SigLIP2 model producing incorrect or nonsensical outputs, indicating the same underlying bug." + }, + { + "left": "issue:43643", + "right": "issue:45375", + "accept": false, + "reason": "Both involve config fields, but they concern different models and different failure mechanisms." + }, + { + "left": "issue:28282", + "right": "issue:29127", + "accept": false, + "reason": "PyTorch-missing ImportError and LayoutLMv3 error-message clarity are not duplicates." + }, + { + "left": "issue:43756", + "right": "issue:45325", + "accept": false, + "reason": "Both touch RoPE/position handling, but they are different models and distinct bugs." + }, + { + "left": "issue:44062", + "right": "issue:45081", + "accept": false, + "reason": "AddedToken constructor conflict is unrelated to the Mistral regex crash." + }, + { + "left": "issue:44291", + "right": "issue:44821", + "accept": false, + "reason": "init_empty_weights argument handling and AutoImageProcessor URL loading are separate failures." + }, + { + "left": "issue:43452", + "right": "issue:44488", + "accept": false, + "reason": "gguf loading breakage and a specific model load failure for cjvt/sleng-bert are different issues." + }, + { + "left": "issue:34567", + "right": "issue:36331", + "accept": false, + "reason": "TrainerState token-counting update and CustomTrainer signature mismatch are different trainer bugs." + }, + { + "left": "issue:44792", + "right": "issue:44964", + "accept": false, + "reason": "A janus test failure is not the same as loading Phi-4 multimodal instruct." + }, + { + "left": "issue:43688", + "right": "issue:43749", + "accept": false, + "reason": "Auxiliary-loss normalization and FSDP CPU RAM-efficient loading are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:45290", + "accept": false, + "reason": "The packet describes batched-input/padding=False crashes on one side and tool-call assistant-message crashes on the other; not the same bug." + }, + { + "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:43577", + "accept": false, + "reason": "gguf loading breakage and Blip2 dtype propagation are separate issues." + }, + { + "left": "issue:43643", + "right": "issue:44561", + "accept": false, + "reason": "Both affect trust_remote_code models, but the concrete failure modes differ." + }, + { + "left": "issue:45042", + "right": "issue:45081", + "accept": false, + "reason": "PIL backend torchvision dependency issue is unrelated to the Mistral regex crash." + }, + { + "left": "issue:43531", + "right": "issue:44821", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and AutoImageProcessor URL loading are unrelated." + }, + { + "left": "issue:44514", + "right": "issue:45381", + "accept": false, + "reason": "Batched-input/padding=False apply_chat_template crash and Qwen2.5-VL video position_ids bug are different problems." + }, + { + "left": "issue:43697", + "right": "issue:45216", + "accept": false, + "reason": "RTDetrV2 output regression and Qwen3.5 checkpoint corruption are not the same bug." + }, + { + "left": "issue:43452", + "right": "issue:44843", + "accept": false, + "reason": "gguf loading breakage and offline model_info calls in _patch_mistral_regex are different code paths." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The overall heterogeneity summary is fine, but at least a couple of soft-edge reasons overreach the packet. In particular, issue:44514 is described in the packet as a batched-input/padding=False crash, not a tool-call crash, and issue:43906 relies on unknown external context (issue 38071) to justify the comparison to issue:45381. That makes the soft verdicts insufficiently grounded/conservative, so I can\u2019t accept this clustering." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 16, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:24643", + "issue:28282", + "issue:29127", + "issue:30064", + "issue:32090", + "issue:33357", + "issue:33666", + "issue:35141", + "issue:38175", + "issue:40444", + "issue:41720", + "issue:41762", + "issue:43122", + "issue:43295", + "issue:43329", + "issue:43525", + "issue:43531", + "issue:43577", + "issue:43643", + "issue:43653", + "issue:43688", + "issue:43697", + "issue:43720", + "issue:43742", + "issue:43819", + "issue:43906", + "issue:43927", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44466", + "issue:44479", + "issue:44554", + "issue:44568", + "issue:44625", + "issue:44743", + "issue:44811", + "issue:44843", + "issue:44877", + "issue:44898", + "issue:44936", + "issue:44977", + "issue:44991", + "issue:45005", + "issue:45325", + "issue:45356", + "issue:45440" + ], + "soft_pairs": [ + "issue:43688|issue:45440", + "issue:44625|issue:44877", + "issue:43819|issue:44811", + "issue:43720|issue:44898", + "issue:41762|issue:44936", + "issue:43653|issue:44568", + "issue:44464|issue:44743", + "issue:44466|issue:45005", + "issue:43295|issue:45356", + "issue:24643|issue:29127", + "issue:41720|issue:43906", + "issue:43525|issue:43643", + "issue:40444|issue:44811", + "issue:44843|issue:44991", + "issue:44361|issue:44442", + "issue:43577|issue:43697", + "issue:38175|issue:43295", + "issue:43653|issue:44361", + "issue:44479|issue:44936", + "issue:28282|issue:32090", + "issue:28282|issue:33357", + "issue:30064|issue:33666", + "issue:28282|issue:35141", + "issue:43122|issue:43906", + "issue:44977|issue:45356", + "issue:44464|issue:44977", + "issue:43531|issue:43742", + "issue:43577|issue:44554", + "issue:43653|issue:43927", + "issue:43329|issue:45325" + ], + "prepared_review_unit_hash": "20f9973b334efb4625fd81110350eb911477a185c65a853d07fc522688fd360f", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36486, + "estimated_input_tokens": 9122, + "estimated_eval_tokens": 18500 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly unrelated issues spanning tokenizer/config problems, multimodal regressions, trainer/DeepSpeed failures, and model-specific bugs. Only one plausible duplicate edge stands out: the tied-weights v5 regressions around 44466 and 45005.", + "confidence": 0.91, + "canonical_issue_reason": "If a canonical issue must be chosen, issue 45005 is the broadest representative of the tied-weights v5 regression family and best covers the one plausible duplicate pair (44466/45005).", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 45005 is the best representative issue in this set because it describes the broader tied-weights regression, while 44466 looks like a narrower symptom of the same underlying bug.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43688", + "right": "issue:45440", + "accept": false, + "reason": "Both involve MoE-related model behavior, but one is auxiliary-loss normalization and the other is DeepSeekV3 implementation divergence; different bug surfaces and code paths." + }, + { + "left": "issue:44625", + "right": "issue:44877", + "accept": false, + "reason": "Both are config-related, but one is num_labels propagation in Qwen3.5 and the other is strict loading for granite_speech; not the same underlying issue." + }, + { + "left": "issue:43819", + "right": "issue:44811", + "accept": false, + "reason": "DAC latent decoding mismatch and Whisper batch_decode skip_special_tokens are unrelated tokenizer/model behaviors." + }, + { + "left": "issue:43720", + "right": "issue:44898", + "accept": false, + "reason": "Packed-weight loading for BitNet and Perceiver interpolation failures are different model paths with no shared bug." + }, + { + "left": "issue:41762", + "right": "issue:44936", + "accept": false, + "reason": "Gemma3 ZeRO-3 loading index error and trainer.evaluate-after-train failure are unrelated trainer/loading issues." + }, + { + "left": "issue:43653", + "right": "issue:44568", + "accept": false, + "reason": "Both touch special-token handling, but they affect different tokenizers and different behaviors (mask token registration vs BOS/EOS insertion)." + }, + { + "left": "issue:44464", + "right": "issue:44743", + "accept": false, + "reason": "Chunked generation with compiled forward and Qwen3.5 recurrent-state reset are different generation bugs." + }, + { + "left": "issue:44466", + "right": "issue:45005", + "accept": true, + "reason": "Both describe tied-weights regressions in v5; 45005 is the broader formulation and 44466 reads like a concrete manifestation of the same serialization/tied-weight bug." + }, + { + "left": "issue:43295", + "right": "issue:45356", + "accept": false, + "reason": "Processor.tokenizer regression and Kimi-K2.5 codec/regex issues are unrelated tokenizer regressions." + }, + { + "left": "issue:24643", + "right": "issue:29127", + "accept": false, + "reason": "DeepSpeed weight shape error and LayoutLMv3 box-validation messaging are unrelated." + }, + { + "left": "issue:41720", + "right": "issue:43906", + "accept": false, + "reason": "Qwen3 auto device-mapping cuda assert and an isolated reproduction of a different issue are not the same bug." + }, + { + "left": "issue:43525", + "right": "issue:43643", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config and missing fields from trust_remote_code AutoConfig are distinct config-loading problems." + }, + { + "left": "issue:40444", + "right": "issue:44811", + "accept": false, + "reason": "IterableDataset multi-image fine-tuning and Whisper batch_decode token skipping are unrelated." + }, + { + "left": "issue:44843", + "right": "issue:44991", + "accept": false, + "reason": "Offline model_info call in AutoTokenizer patching and tokenizer load failure for EMBEDDIA/est-roberta are different failures." + }, + { + "left": "issue:44361", + "right": "issue:44442", + "accept": false, + "reason": "MLukeTokenizer AttributeError and FastSpeech2ConformerTokenizer load failure are separate tokenizer-specific bugs." + }, + { + "left": "issue:43577", + "right": "issue:43697", + "accept": false, + "reason": "BLIP2 dtype propagation and RTDetrV2 output differences are unrelated model-behavior bugs." + }, + { + "left": "issue:38175", + "right": "issue:43295", + "accept": false, + "reason": "Zero probabilities in SigLIP2 and processor.tokenizer regression are unrelated." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and DiaConfig custom token ID loss are different tokenizer/config issues." + }, + { + "left": "issue:44479", + "right": "issue:44936", + "accept": false, + "reason": "Video-input regression across Qwen models and trainer.evaluate-after-train are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:32090", + "accept": false, + "reason": "Missing PyTorch import and a NoneType GPU broadcast error are unrelated." + }, + { + "left": "issue:28282", + "right": "issue:33357", + "accept": false, + "reason": "ImportError for missing torch and MacOS bus error are unrelated." + }, + { + "left": "issue:30064", + "right": "issue:33666", + "accept": false, + "reason": "Void segmentation map processing and Qwen2-VL multi-GPU training are not the same problem." + }, + { + "left": "issue:28282", + "right": "issue:35141", + "accept": false, + "reason": "Torch import failure and tied-embedding reinitialization are unrelated." + }, + { + "left": "issue:43122", + "right": "issue:43906", + "accept": false, + "reason": "Tokenization differences across versions and an isolated reproduction of another issue are not enough to establish the same bug." + }, + { + "left": "issue:44977", + "right": "issue:45356", + "accept": false, + "reason": "Qwen3.5 flash-attention generation failure and Kimi-K2.5 tokenizer regression are unrelated." + }, + { + "left": "issue:44464", + "right": "issue:44977", + "accept": false, + "reason": "Both are generation-related, but one is compiled-forward chunked generation and the other is flash-attention model behavior; different code paths." + }, + { + "left": "issue:43531", + "right": "issue:43742", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior and MobileLLM-125M loading key errors are unrelated." + }, + { + "left": "issue:43577", + "right": "issue:44554", + "accept": false, + "reason": "BLIP2 dtype issue and MPS attention correctness issue are different model/runtime problems." + }, + { + "left": "issue:43653", + "right": "issue:43927", + "accept": false, + "reason": "Tokenizer special-token registration and DiaConfig save/load token-ID loss are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:45325", + "accept": false, + "reason": "Video-branch undefined variables in multimodal token counting and still-image temporal position scaling are both multimodal bugs, but not the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet and is mostly conservative. The only accepted soft edge, 44466/45005, is plausibly the same tied-weights v5 regression family based on the titles, and the other pairs are appropriately rejected as different bug surfaces." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 17, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:32090", + "issue:33666", + "issue:34567", + "issue:36010", + "issue:41720", + "issue:43012", + "issue:43257", + "issue:43262", + "issue:43278", + "issue:43299", + "issue:43329", + "issue:43334", + "issue:43452", + "issue:43493", + "issue:43504", + "issue:43525", + "issue:43540", + "issue:43550", + "issue:43572", + "issue:43577", + "issue:43606", + "issue:43688", + "issue:43716", + "issue:43720", + "issue:43742", + "issue:43756", + "issue:43825", + "issue:43844", + "issue:43873", + "issue:43927", + "issue:44361", + "issue:44451", + "issue:44464", + "issue:44479", + "issue:44488", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44821", + "issue:44843", + "issue:44849", + "issue:44913", + "issue:44977", + "issue:45084", + "issue:45200", + "issue:45229", + "issue:45357", + "issue:45520" + ], + "soft_pairs": [ + "issue:43452|issue:44451", + "issue:44479|issue:44625", + "issue:43012|issue:43577", + "issue:41720|issue:44464", + "issue:43504|issue:43720", + "issue:43550|issue:44361", + "issue:43504|issue:43550", + "issue:32090|issue:33666", + "issue:32090|issue:34567", + "issue:43504|issue:44610", + "issue:43262|issue:45084", + "issue:43278|issue:43688", + "issue:45200|issue:45229", + "issue:44849|issue:45357", + "issue:43334|issue:43572", + "issue:43742|issue:44821", + "issue:43540|issue:43716", + "issue:43493|issue:43825", + "issue:43756|issue:44913", + "issue:43525|issue:43572", + "issue:33666|issue:36010", + "issue:43688|issue:43873", + "issue:43720|issue:43927", + "issue:43606|issue:43927", + "issue:43278|issue:43844", + "issue:44488|issue:44843", + "issue:43257|issue:43299", + "issue:44589|issue:45520", + "issue:43540|issue:44977", + "issue:43720|issue:44361", + "issue:43329|issue:43606" + ], + "prepared_review_unit_hash": "f4f03343c43b047ee764ce8209e2e1900ff7dfb5d977960f0fe85cbe85e9d552", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 36961, + "estimated_input_tokens": 9241, + "estimated_eval_tokens": 18738 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a grab bag of unrelated Transformers issues; the soft pairs share broad vocabulary but not the same concrete bug or fix path. No single issue stands out as a safe canonical duplicate target.", + "confidence": 0.95, + "canonical_issue_reason": "No single issue is a clear canonical representative: these items span unrelated regressions in loading, training, quantization, multimodal processing, config serialization, and warning/error-message behavior.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is strong enough to serve as a global best canonical for the cluster, since the pairwise similarities are mostly superficial and the underlying failures differ.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43452", + "right": "issue:44451", + "accept": false, + "reason": "Both are model-loading failures, but they involve different APIs/models and likely different root causes; too broad to merge." + }, + { + "left": "issue:44479", + "right": "issue:44625", + "accept": false, + "reason": "One is a v5 video-input regression across Qwen VL variants; the other is a Qwen3.5 num_labels propagation bug. Different code paths." + }, + { + "left": "issue:43012", + "right": "issue:43577", + "accept": false, + "reason": "Both mention dtype/precision, but one is a compile-time warning and the other is incorrect load-time dtypes for Blip2; not the same bug." + }, + { + "left": "issue:41720", + "right": "issue:44464", + "accept": false, + "reason": "CUDA assert with auto device mapping vs inconsistent chunked generation under compiled forward are unrelated failures." + }, + { + "left": "issue:43504", + "right": "issue:43720", + "accept": false, + "reason": "Legacy-field load failure for BEiT segmentation and unpacked-weights accelerate loading for BitNet are distinct issues." + }, + { + "left": "issue:43550", + "right": "issue:44361", + "accept": false, + "reason": "A torch.compile/SDPA model bug and a tokenizer AttributeError are unrelated." + }, + { + "left": "issue:43504", + "right": "issue:43550", + "accept": false, + "reason": "Different models, different subsystems, and different failure modes; no common concrete bug." + }, + { + "left": "issue:32090", + "right": "issue:33666", + "accept": false, + "reason": "Generic Trainer broadcast NoneType error vs Qwen2-VL multi-GPU training problem; too little evidence of the same root cause." + }, + { + "left": "issue:32090", + "right": "issue:34567", + "accept": false, + "reason": "Trainer broadcast TypeError and TrainerState token counter not updating are separate training bugs." + }, + { + "left": "issue:43504", + "right": "issue:44610", + "accept": false, + "reason": "Pretrained preset loading regression and OmDet-Turbo processor/model size mismatch are unrelated." + }, + { + "left": "issue:43262", + "right": "issue:45084", + "accept": false, + "reason": "Audio chat-template sampling-rate default and non-template-node compilation error are different parts of the stack." + }, + { + "left": "issue:43278", + "right": "issue:43688", + "accept": false, + "reason": "Embedding dtype drift in eval and auxiliary-loss normalization are unrelated training behaviors." + }, + { + "left": "issue:45200", + "right": "issue:45229", + "accept": false, + "reason": "Text-only mm_token_type_ids defaulting and Gemma4 multi-GPU inference OOM are separate issues." + }, + { + "left": "issue:44849", + "right": "issue:45357", + "accept": false, + "reason": "Both are Qwen3.5 regressions, but one is output_hidden_states behavior and the other is save_pretrained key serialization; different bugs." + }, + { + "left": "issue:43334", + "right": "issue:43572", + "accept": false, + "reason": "Missing pad_token_id on Qwen3-VL text config vs missing pad_token_idx on StableLmConfig are similar symptom classes but not the same bug." + }, + { + "left": "issue:43742", + "right": "issue:44821", + "accept": false, + "reason": "KeyError while loading a model vs AutoImageProcessor URL loading are unrelated loader paths." + }, + { + "left": "issue:43540", + "right": "issue:43716", + "accept": false, + "reason": "Video-processing ValueError in Qwen3OmniMoe and image-preprocessor dtype mismatch in Mistral-3 are different multimodal bugs." + }, + { + "left": "issue:43493", + "right": "issue:43825", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and pipeline task-support error message are not the same underlying issue." + }, + { + "left": "issue:43756", + "right": "issue:44913", + "accept": false, + "reason": "Both touch RoPE-related configuration, but they affect different models and different mechanisms; not a mergeable duplicate." + }, + { + "left": "issue:43525", + "right": "issue:43572", + "accept": false, + "reason": "Both are missing token-id fields after v5 changes, but they concern different configs and different field names." + }, + { + "left": "issue:33666", + "right": "issue:36010", + "accept": false, + "reason": "Qwen2-VL multi-GPU training and GenerationMixin import failure are unrelated." + }, + { + "left": "issue:43688", + "right": "issue:43873", + "accept": false, + "reason": "Auxiliary-loss normalization and quantization/offloading behavior are distinct issues." + }, + { + "left": "issue:43720", + "right": "issue:43927", + "accept": false, + "reason": "Packed-weight handling in accelerate loading and DiaConfig save/load token IDs are unrelated regression classes." + }, + { + "left": "issue:43606", + "right": "issue:43927", + "accept": false, + "reason": "CPU offload device mismatch and config token-ID persistence are different bugs." + }, + { + "left": "issue:43278", + "right": "issue:43844", + "accept": false, + "reason": "Eval dtype drift vs exploding gradients under ZeRO-3 have no concrete overlap." + }, + { + "left": "issue:44488", + "right": "issue:44843", + "accept": false, + "reason": "Model loading failure and offline-mode hub lookup bug are different problems." + }, + { + "left": "issue:43257", + "right": "issue:43299", + "accept": false, + "reason": "Both concern Qwen3 MOE loading, but one is accelerate+deepspeed conversion and the other is a transformer-dev version breakage; not clearly the same fix." + }, + { + "left": "issue:44589", + "right": "issue:45520", + "accept": false, + "reason": "Missing Float8 storage object and flash_attn import key error are unrelated environment/compatibility issues." + }, + { + "left": "issue:43540", + "right": "issue:44977", + "accept": false, + "reason": "Video-input processing error and flash-attention generation issue in Qwen3.5 are different code paths." + }, + { + "left": "issue:43720", + "right": "issue:44361", + "accept": false, + "reason": "Accelerate loading of packed BitNet weights and MLukeTokenizer task AttributeError are unrelated." + }, + { + "left": "issue:43329", + "right": "issue:43606", + "accept": false, + "reason": "Undefined multimodal token helper in video branch and Bark CPU-offload device mismatch are different bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 18, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:15354", + "issue:33357", + "issue:33666", + "issue:36331", + "issue:38175", + "issue:39692", + "issue:43257", + "issue:43278", + "issue:43295", + "issue:43493", + "issue:43504", + "issue:43540", + "issue:43572", + "issue:43575", + "issue:43582", + "issue:43606", + "issue:43646", + "issue:43653", + "issue:43688", + "issue:43701", + "issue:43746", + "issue:43756", + "issue:43824", + "issue:43844", + "issue:43927", + "issue:43994", + "issue:44060", + "issue:44112", + "issue:44291", + "issue:44361", + "issue:44423", + "issue:44466", + "issue:44493", + "issue:44521", + "issue:44554", + "issue:44589", + "issue:44734", + "issue:44792", + "issue:44964", + "issue:44977", + "issue:45072", + "issue:45084", + "issue:45198", + "issue:45200", + "issue:45290", + "issue:45405", + "issue:45446", + "issue:45520" + ], + "soft_pairs": [ + "issue:39692|issue:43493", + "issue:44964|issue:45405", + "issue:33666|issue:36331", + "issue:43493|issue:43844", + "issue:43756|issue:44060", + "issue:43646|issue:44466", + "issue:44423|issue:44734", + "issue:43688|issue:43746", + "issue:43824|issue:44589", + "issue:44521|issue:45290", + "issue:33357|issue:33666", + "issue:43572|issue:44291", + "issue:39692|issue:43295", + "issue:43504|issue:43606", + "issue:43994|issue:44493", + "issue:43575|issue:45405", + "issue:43504|issue:43653", + "issue:43927|issue:44361", + "issue:43257|issue:45405", + "issue:43278|issue:43701", + "issue:44792|issue:45405", + "issue:43606|issue:44112", + "issue:45084|issue:45520", + "issue:15354|issue:38175", + "issue:44554|issue:45446", + "issue:43540|issue:45200", + "issue:43582|issue:44977", + "issue:45072|issue:45198" + ], + "prepared_review_unit_hash": "cd996b56b5fbe79d03a432894d5fb5561c7b2b6ec420c4a2c595b4b774bc99ac", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 28, + "serialized_chars": 36304, + "estimated_input_tokens": 9076, + "estimated_eval_tokens": 18408 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is heterogeneous and does not present a single duplicate theme. The soft-pair candidates share only broad subsystem keywords at most, not the same concrete bug or fix.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39692", + "right": "issue:43493", + "accept": false, + "reason": "Both involve SigLIP2, but one is a documentation/example error and the other is a model implementation discrepancy; different bugs and fixes." + }, + { + "left": "issue:44964", + "right": "issue:45405", + "accept": false, + "reason": "Phi-4 multimodal loading failure vs. a PEFT version bump problem; unrelated failure modes." + }, + { + "left": "issue:33666", + "right": "issue:36331", + "accept": false, + "reason": "Multi-GPU training question for Qwen2-VL is unrelated to a custom trainer signature regression." + }, + { + "left": "issue:43493", + "right": "issue:43844", + "accept": false, + "reason": "SigLIP2 implementation mismatch and ZeRO-3 gradient inflation are different code paths and symptoms." + }, + { + "left": "issue:43756", + "right": "issue:44060", + "accept": false, + "reason": "Smollm3 RoPE-layer bug and Qwen3-Next tied-weights warning are unrelated model-specific issues." + }, + { + "left": "issue:43646", + "right": "issue:44466", + "accept": false, + "reason": "Both are v5-related, but one breaks model initialization while the other is about state dict serialization/tied weights." + }, + { + "left": "issue:44423", + "right": "issue:44734", + "accept": false, + "reason": "Both are serve crashes, but they hit different pathways: multimodal continuous batching vs KV-cache continuation indexing." + }, + { + "left": "issue:43688", + "right": "issue:43746", + "accept": false, + "reason": "Aux-loss normalization bug and PEFT-local-checkpoint loading bug are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:44589", + "accept": false, + "reason": "ImportError for a model class vs missing float8 storage type; no shared underlying defect." + }, + { + "left": "issue:44521", + "right": "issue:45290", + "accept": false, + "reason": "Both touch apply_chat_template, but one returns bad masks for multimodal inputs while the other crashes on tool-call assistant messages." + }, + { + "left": "issue:33357", + "right": "issue:33666", + "accept": false, + "reason": "MacOS bus error on a CLIP model is unrelated to Qwen2-VL multi-GPU training." + }, + { + "left": "issue:43572", + "right": "issue:44291", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression and init_empty_weights unexpected argument are separate v5 breakages." + }, + { + "left": "issue:39692", + "right": "issue:43295", + "accept": false, + "reason": "Doc example errors for SigLIP2 do not match the processor.tokenizer regression in multimodal custom model code." + }, + { + "left": "issue:43504", + "right": "issue:43606", + "accept": false, + "reason": "Legacy field load failure in Beit differs from CPU-offload device mismatch in bark-small." + }, + { + "left": "issue:43994", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 bad outputs and unexpected position-id key warnings are distinct issues." + }, + { + "left": "issue:43575", + "right": "issue:45405", + "accept": false, + "reason": "Tensor-parallel OOM when loading Qwen2-57B is unrelated to the PEFT version constant bump." + }, + { + "left": "issue:43504", + "right": "issue:43653", + "accept": false, + "reason": "Model preset loading with a legacy field is unrelated to BigBirdTokenizer special-token registration." + }, + { + "left": "issue:43927", + "right": "issue:44361", + "accept": false, + "reason": "Custom token IDs lost on save/load and MLukeTokenizer task AttributeError are different tokenizer bugs." + }, + { + "left": "issue:43257", + "right": "issue:45405", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate+deepspeed is unrelated to the PEFT release-version mismatch." + }, + { + "left": "issue:43278", + "right": "issue:43701", + "accept": false, + "reason": "Embedding dtype drift during eval and resume_from_checkpoint key mismatch are unrelated regressions." + }, + { + "left": "issue:44792", + "right": "issue:45405", + "accept": false, + "reason": "Janus image-generation test failure has nothing to do with the PEFT version bump." + }, + { + "left": "issue:43606", + "right": "issue:44112", + "accept": false, + "reason": "Both mention GraniteSpeech only indirectly, but one is a real CPU-offload device mismatch and the other is a stale CI test." + }, + { + "left": "issue:45084", + "right": "issue:45520", + "accept": false, + "reason": "Template-node compilation error and flash_attn import KeyError on Python 3.13 are unrelated." + }, + { + "left": "issue:15354", + "right": "issue:38175", + "accept": false, + "reason": "TorchScript GeneratorExp export failure and unexpected zero probabilities in SigLIP2 are unrelated." + }, + { + "left": "issue:44554", + "right": "issue:45446", + "accept": false, + "reason": "Attention correctness with mismatched head dims is a different bug from a PyTorch version check in flex_attention imports." + }, + { + "left": "issue:43540", + "right": "issue:45200", + "accept": false, + "reason": "Video-input processing error in Qwen3OmniMoe is unrelated to Gemma 4 text-only fine-tuning defaults." + }, + { + "left": "issue:43582", + "right": "issue:44977", + "accept": false, + "reason": "Apple Silicon allocator TypeError and Qwen3.5 flash-attention generation problems are different issues." + }, + { + "left": "issue:45072", + "right": "issue:45198", + "accept": false, + "reason": "bfloat16 dtype mismatches in inference and Wav2Vec2 save/tokenization failures 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": 13, + "review_unit_index": 19, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:37428", + "issue:38175", + "issue:39401", + "issue:40990", + "issue:42915", + "issue:43232", + "issue:43278", + "issue:43295", + "issue:43334", + "issue:43388", + "issue:43450", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43572", + "issue:43575", + "issue:43701", + "issue:43824", + "issue:43844", + "issue:43856", + "issue:43873", + "issue:43957", + "issue:43975", + "issue:43994", + "issue:44315", + "issue:44360", + "issue:44442", + "issue:44464", + "issue:44466", + "issue:44514", + "issue:44521", + "issue:44554", + "issue:44561", + "issue:44568", + "issue:44589", + "issue:44610", + "issue:44625", + "issue:44849", + "issue:44898", + "issue:44936", + "issue:44964", + "issue:44987", + "issue:45127", + "issue:45198", + "issue:45200", + "issue:45325", + "issue:45356", + "issue:45412" + ], + "soft_pairs": [ + "issue:43232|issue:44849", + "issue:43844|issue:44964", + "issue:39401|issue:44568", + "issue:37428|issue:43957", + "issue:44315|issue:44849", + "issue:44554|issue:45356", + "issue:43334|issue:43525", + "issue:43295|issue:44610", + "issue:45200|issue:45325", + "issue:43493|issue:43701", + "issue:44898|issue:45412", + "issue:44466|issue:45127", + "issue:43873|issue:44464", + "issue:40990|issue:44625", + "issue:37428|issue:43824", + "issue:43388|issue:43526", + "issue:43572|issue:44987", + "issue:44514|issue:44521", + "issue:43844|issue:44936", + "issue:43572|issue:44589", + "issue:43575|issue:44464", + "issue:43824|issue:43975", + "issue:44442|issue:45198", + "issue:43295|issue:43450", + "issue:43295|issue:43873", + "issue:38175|issue:45412", + "issue:43278|issue:43856", + "issue:44360|issue:44561", + "issue:43295|issue:43994", + "issue:42915|issue:43824" + ], + "prepared_review_unit_hash": "8d01796bb013ead3f5dd809762577d0b7db38e0655c69843e165c717a4eb1c01", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36838, + "estimated_input_tokens": 9210, + "estimated_eval_tokens": 18676 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "I don\u2019t see any true duplicate pairs here; the list spans many unrelated issues across loading, tokenization, generation, training, and multimodal processing. No PRs are present.", + "confidence": 0.92, + "canonical_issue_reason": "No clear canonical issue emerges because the items are heterogeneous and do not share one underlying bug or code path.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick one representative, issue:44521 is the most active/referenced recent bug, but it is only a narrow multimodal chat-template regression and not representative of the whole set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:44849", + "accept": false, + "reason": "Different problems: generation kwargs after sync_gpus vs Qwen3.5 hidden-states behavior." + }, + { + "left": "issue:43844", + "right": "issue:44964", + "accept": false, + "reason": "ZeRO-3 gradient anomaly and Phi-4 multimodal loading failure are unrelated." + }, + { + "left": "issue:39401", + "right": "issue:44568", + "accept": false, + "reason": "Both are tokenizer bugs, but one is offset mapping and the other is special-token insertion; different code paths." + }, + { + "left": "issue:37428", + "right": "issue:43957", + "accept": false, + "reason": "ImportError for a missing flash-attention helper is not the same as meta-device model-loading breakage." + }, + { + "left": "issue:44315", + "right": "issue:44849", + "accept": false, + "reason": "Liger Kernel application in model_init is unrelated to Qwen3.5 output_hidden_states." + }, + { + "left": "issue:44554", + "right": "issue:45356", + "accept": false, + "reason": "MPS attention correctness issue and Kimi tokenizer regression do not share a concrete bug." + }, + { + "left": "issue:43334", + "right": "issue:43525", + "accept": false, + "reason": "Similar symptom (missing pad_token_id), but different models/configs and likely separate fixes." + }, + { + "left": "issue:43295", + "right": "issue:44610", + "accept": false, + "reason": "Processor.tokenizer regression is unrelated to OmDet-Turbo image size mismatch." + }, + { + "left": "issue:45200", + "right": "issue:45325", + "accept": false, + "reason": "Different multimodal bugs: token_type_ids defaulting vs rope/position_ids scaling." + }, + { + "left": "issue:43493", + "right": "issue:43701", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and checkpoint key mismatch are unrelated." + }, + { + "left": "issue:44898", + "right": "issue:45412", + "accept": false, + "reason": "Perceiver resolution handling and RT-DETR memory reclamation are different issues." + }, + { + "left": "issue:44466", + "right": "issue:45127", + "accept": false, + "reason": "Serialization of tied weights is separate from LoRA merge collapse with extended vocab." + }, + { + "left": "issue:43873", + "right": "issue:44464", + "accept": false, + "reason": "Quantization offloading and compiled-forward chunked generation are different failure modes." + }, + { + "left": "issue:40990", + "right": "issue:44625", + "accept": false, + "reason": "High perplexity on gpt-oss-20b is unrelated to Qwen3.5 num_labels propagation." + }, + { + "left": "issue:37428", + "right": "issue:43824", + "accept": false, + "reason": "Two different missing-import symbols; not the same underlying import regression." + }, + { + "left": "issue:43388", + "right": "issue:43526", + "accept": false, + "reason": "Both involve labels, but gather_for_metrics batching and BeitImageProcessorFast reduce_labels are distinct bugs." + }, + { + "left": "issue:43572", + "right": "issue:44987", + "accept": false, + "reason": "StableLm pad_token_idx regression and loading physical-intelligence/fast on 5.1 are not the same bug." + }, + { + "left": "issue:44514", + "right": "issue:44521", + "accept": false, + "reason": "Both hit apply_chat_template, but one crashes on batched padding=False input while the other produces wrong assistant masks; too different to merge." + }, + { + "left": "issue:43844", + "right": "issue:44936", + "accept": false, + "reason": "Gradient growth under ZeRO-3 and trainer.evaluate after train are unrelated training issues." + }, + { + "left": "issue:43572", + "right": "issue:44589", + "accept": false, + "reason": "Missing pad_token_idx and Float8 storage lookup failure are different compatibility problems." + }, + { + "left": "issue:43575", + "right": "issue:44464", + "accept": false, + "reason": "TP OOM on Qwen2-57B and chunked-generation inconsistency are unrelated." + }, + { + "left": "issue:43824", + "right": "issue:43975", + "accept": false, + "reason": "Import failure for Qwen2.5-VL is unrelated to DeepSeek-Coder detokenization." + }, + { + "left": "issue:44442", + "right": "issue:45198", + "accept": false, + "reason": "AutoTokenizer loading failure for one model and save_pretrained/tokenization failure for another are different bugs." + }, + { + "left": "issue:43295", + "right": "issue:43450", + "accept": false, + "reason": "Processor.tokenizer regression is not the same as batched video-shape handling." + }, + { + "left": "issue:43295", + "right": "issue:43873", + "accept": false, + "reason": "Custom processor regression and quantization offloading are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 zero probabilities and RT-DETR memory retention are entirely different." + }, + { + "left": "issue:43278", + "right": "issue:43856", + "accept": false, + "reason": "Embedding dtype drift in eval is not the same as Qwen3 MoE memory inefficiency." + }, + { + "left": "issue:44360", + "right": "issue:44561", + "accept": false, + "reason": "DSA indexer/ReLU discussion is unrelated to removal of is_torch_fx_available." + }, + { + "left": "issue:43295", + "right": "issue:43994", + "accept": false, + "reason": "Processor.tokenizer regression and SigLIP2 nonsensical outputs do not share the same root cause." + }, + { + "left": "issue:42915", + "right": "issue:43824", + "accept": false, + "reason": "Qwen3Moe FP8 failure is unrelated to Qwen2.5-VL import errors." + } + ] + }, + "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": 13, + "review_unit_index": 20, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:29942", + "issue:32090", + "issue:37428", + "issue:38175", + "issue:39692", + "issue:43262", + "issue:43278", + "issue:43335", + "issue:43404", + "issue:43425", + "issue:43450", + "issue:43493", + "issue:43526", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43611", + "issue:43653", + "issue:43819", + "issue:43825", + "issue:43827", + "issue:43901", + "issue:43931", + "issue:44112", + "issue:44292", + "issue:44351", + "issue:44448", + "issue:44462", + "issue:44493", + "issue:44561", + "issue:44568", + "issue:44610", + "issue:44743", + "issue:44779", + "issue:44849", + "issue:44855", + "issue:44857", + "issue:44991", + "issue:45003", + "issue:45083", + "issue:45127", + "issue:45200", + "issue:45245", + "issue:45290", + "issue:45310", + "issue:45381", + "issue:45478", + "issue:45520" + ], + "soft_pairs": [ + "issue:43827|issue:43901", + "issue:44849|issue:45003", + "issue:38175|issue:43493", + "issue:43653|issue:44112", + "issue:44568|issue:44991", + "issue:43450|issue:43493", + "issue:45127|issue:45245", + "issue:45290|issue:45381", + "issue:43819|issue:45083", + "issue:44462|issue:44561", + "issue:43450|issue:43575", + "issue:44561|issue:44779", + "issue:44610|issue:44857", + "issue:37428|issue:43425", + "issue:43819|issue:44112", + "issue:44448|issue:44743", + "issue:43262|issue:43450", + "issue:43278|issue:43825", + "issue:43606|issue:45083", + "issue:29942|issue:32090", + "issue:45310|issue:45478", + "issue:43611|issue:43931", + "issue:37428|issue:44351", + "issue:43819|issue:44493", + "issue:39692|issue:43526", + "issue:43335|issue:43550", + "issue:45200|issue:45381", + "issue:43450|issue:43526", + "issue:44855|issue:45083", + "issue:43404|issue:45200", + "issue:44292|issue:45520", + "issue:39692|issue:44493", + "issue:44448|issue:44779" + ], + "prepared_review_unit_hash": "b21fb352f604897138bef0b6b5e734c3ef62600a300126d2c12909e6cc2dbd5c", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37621, + "estimated_input_tokens": 9406, + "estimated_eval_tokens": 19068 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most pairs are broad topical similarities, not true duplicates. The only clear duplicate-looking pair is the repeated Qwen3.5 MoE from_pretrained error (45310/45478), with 45310 the better canonical target.", + "confidence": 0.89, + "canonical_issue_reason": "Issue 45310 is the clearer, more detailed report of the same Qwen3.5 MoE from_pretrained failure; 45478 reads like a later duplicate/restatement.", + "canonical_pr_reason": null, + "best_issue_reason": "45310 is the most triage-worthy issue here: it names the exact failing model, version area, and concrete loading error, and it cleanly anchors the only strong duplicate pair.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43827", + "right": "issue:43901", + "accept": false, + "reason": "Both are pipeline-v5 docs/problems, but one is about translation pipeline references and the other is about TextClassificationPipeline return_all_scores; different concrete issues." + }, + { + "left": "issue:44849", + "right": "issue:45003", + "accept": false, + "reason": "Qwen3.5 hidden_states behavior vs unsafe sys.modules access are unrelated code paths." + }, + { + "left": "issue:38175", + "right": "issue:43493", + "accept": false, + "reason": "Same model family, but one reports zero probabilities and the other an HF-vs-JAX implementation discrepancy; not the same bug." + }, + { + "left": "issue:43653", + "right": "issue:44112", + "accept": false, + "reason": "BigBirdTokenizer special-token registration and GraniteSpeech CI device override are unrelated." + }, + { + "left": "issue:44568", + "right": "issue:44991", + "accept": false, + "reason": "Both involve tokenizers after v5, but they affect different models and different failure modes." + }, + { + "left": "issue:43450", + "right": "issue:43493", + "accept": false, + "reason": "Video processor batching shape bug vs SigLIP2 model discrepancy are different subsystems and symptoms." + }, + { + "left": "issue:45127", + "right": "issue:45245", + "accept": false, + "reason": "LoRA/vocab merge collapse is unrelated to the categories > 2^24 runtime error." + }, + { + "left": "issue:45290", + "right": "issue:45381", + "accept": false, + "reason": "Chat-template tool-call crash and Qwen2.5-VL video vision_position_ids bug are different code paths." + }, + { + "left": "issue:43819", + "right": "issue:45083", + "accept": false, + "reason": "DAC latent/STE mismatch is unrelated to qwen3_omni_moe feature-extract length behavior." + }, + { + "left": "issue:44462", + "right": "issue:44561", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and v5 removal of is_torch_fx_available are unrelated." + }, + { + "left": "issue:43450", + "right": "issue:43575", + "accept": false, + "reason": "Batched video shape output and tensor-parallel OOM are not the same underlying bug." + }, + { + "left": "issue:44561", + "right": "issue:44779", + "accept": false, + "reason": "Removed helper breaking trust_remote_code models vs Deepseek tokenizer regression are different issues." + }, + { + "left": "issue:44610", + "right": "issue:44857", + "accept": false, + "reason": "OmDet-Turbo input-size mismatch and LwDetrImageLoss AMP/CUDA crash are unrelated." + }, + { + "left": "issue:37428", + "right": "issue:43425", + "accept": false, + "reason": "ImportError for a flash-attention helper and Torch 2.10 incompatibility are only loosely version-related, not the same bug." + }, + { + "left": "issue:43819", + "right": "issue:44112", + "accept": false, + "reason": "DAC latent mismatch and GraniteSpeech CI stale device override are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44743", + "accept": false, + "reason": "Pegasus v4/v5 output difference and qwen3_5 recurrent-state reset bug do not describe the same failure." + }, + { + "left": "issue:43262", + "right": "issue:43450", + "accept": false, + "reason": "Audio chat-template sampling-rate default and video processor batch-shape bug are different processor issues." + }, + { + "left": "issue:43278", + "right": "issue:43825", + "accept": false, + "reason": "Embedding dtype drift in evaluate vs pipeline translation messaging/docs are unrelated." + }, + { + "left": "issue:43606", + "right": "issue:45083", + "accept": false, + "reason": "CPU-offload device mismatch and qwen3_omni_moe feat-length helper behavior are different problems." + }, + { + "left": "issue:29942", + "right": "issue:32090", + "accept": false, + "reason": "Flash Attention 2 test failures and Trainer NoneType GPU broadcast errors are unrelated." + }, + { + "left": "issue:45310", + "right": "issue:45478", + "accept": true, + "reason": "These are the same reported Qwen3.5 MoE from_pretrained error; 45478 appears to be a duplicate of 45310." + }, + { + "left": "issue:43611", + "right": "issue:43931", + "accept": false, + "reason": "base_model_prefix loading breakage and Qwen3-VL weight-shape mismatch are different load failures." + }, + { + "left": "issue:37428", + "right": "issue:44351", + "accept": false, + "reason": "Different import errors affecting different symbols; not the same underlying change." + }, + { + "left": "issue:43819", + "right": "issue:44493", + "accept": false, + "reason": "DAC STE mismatch and unexpected position-id keys are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:43526", + "accept": false, + "reason": "SigLIP2 doc-example errors and BeitImageProcessorFast label reduction bug are unrelated." + }, + { + "left": "issue:43335", + "right": "issue:43550", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug and Bamba torch.compile SDPA failure are different models and code paths." + }, + { + "left": "issue:45200", + "right": "issue:45381", + "accept": false, + "reason": "mm_token_type_ids defaulting for text-only fine-tuning and Qwen2.5-VL video position-id bug are not the same issue." + }, + { + "left": "issue:43450", + "right": "issue:43526", + "accept": false, + "reason": "Video batching shape bug and BeitImageProcessorFast label reduction bug are unrelated processor issues." + }, + { + "left": "issue:44855", + "right": "issue:45083", + "accept": false, + "reason": "Python 3.13 IndentationError in DebertaV2Model import and qwen3_omni_moe helper behavior are unrelated." + }, + { + "left": "issue:43404", + "right": "issue:45200", + "accept": false, + "reason": "Mistral3 lm_head tie issue and Gemma 4 mm_token_type_ids defaulting are different generation/fine-tuning bugs." + }, + { + "left": "issue:44292", + "right": "issue:45520", + "accept": false, + "reason": "Qwen-3-8B-NVFP4 runtime error and Python 3.13 flash_attn import KeyError are different failures." + }, + { + "left": "issue:39692", + "right": "issue:44493", + "accept": false, + "reason": "SigLIP2 doc-example problems and unexpected position-id key behavior are unrelated." + }, + { + "left": "issue:44448", + "right": "issue:44779", + "accept": false, + "reason": "Pegasus output drift and Deepseek tokenizer regression are different model-specific bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "Grounded and conservative overall. The only accepted duplicate pair, issue:45310/issue:45478, is well supported by the identical Qwen3.5 MoE from_pretrained failure title/version. The remaining soft-edge rejects are appropriately cautious and do not overstate duplicate evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 21, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:25251", + "issue:29127", + "issue:29942", + "issue:32090", + "issue:33357", + "issue:34689", + "issue:35141", + "issue:37428", + "issue:38175", + "issue:40444", + "issue:41720", + "issue:43065", + "issue:43066", + "issue:43278", + "issue:43329", + "issue:43335", + "issue:43504", + "issue:43526", + "issue:43550", + "issue:43606", + "issue:43653", + "issue:43701", + "issue:43720", + "issue:43723", + "issue:43873", + "issue:43976", + "issue:43994", + "issue:44112", + "issue:44448", + "issue:44451", + "issue:44462", + "issue:44466", + "issue:44493", + "issue:44554", + "issue:44568", + "issue:44857", + "issue:44977", + "issue:44987", + "issue:45072", + "issue:45083", + "issue:45229", + "issue:45372", + "issue:45406", + "issue:45412", + "issue:45440", + "issue:45459", + "issue:45491", + "issue:45520" + ], + "soft_pairs": [ + "issue:29127|issue:32090", + "issue:44977|issue:45229", + "issue:43335|issue:43653", + "issue:38175|issue:44493", + "issue:43335|issue:45083", + "issue:43335|issue:43504", + "issue:43550|issue:44112", + "issue:43335|issue:45072", + "issue:44568|issue:45491", + "issue:43720|issue:44112", + "issue:44448|issue:44977", + "issue:40444|issue:43526", + "issue:43723|issue:44554", + "issue:43335|issue:44112", + "issue:34689|issue:35141", + "issue:45372|issue:45406", + "issue:43606|issue:44857", + "issue:44857|issue:45072", + "issue:43066|issue:44462", + "issue:43873|issue:44493", + "issue:43873|issue:45229", + "issue:37428|issue:44987", + "issue:43278|issue:43329", + "issue:43701|issue:45083", + "issue:29942|issue:35141", + "issue:41720|issue:45229", + "issue:43550|issue:45083", + "issue:44451|issue:45520", + "issue:43976|issue:45440", + "issue:32090|issue:33357", + "issue:45372|issue:45459", + "issue:44448|issue:44466", + "issue:43065|issue:45412", + "issue:43065|issue:43994", + "issue:25251|issue:32090" + ], + "prepared_review_unit_hash": "5ec1634ce9c4b0c07e981ab9c9c37ba317cbbeb35547b0e3aa7bd5954da83e8e", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 35, + "serialized_chars": 38376, + "estimated_input_tokens": 9594, + "estimated_eval_tokens": 19444 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly a grab bag of unrelated Transformer bugs across different models and subsystems. None of the soft pairs look like the same underlying issue, so all soft edges are rejected. Issue 43335 is the strongest representative because it is the most central item in the soft-link graph and describes a concrete, reproducible bug.", + "confidence": 0.72, + "canonical_issue_reason": "Issue 43335 is the most cluster-central item, appearing in multiple soft-similarity links, and it has a precise, actionable bug description about SwitchTransformers config behavior.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43335 is the best representative of this set because it is the most connected and most clearly scoped bug among otherwise unrelated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:29127", + "right": "issue:32090", + "accept": false, + "reason": "Different failures: LayoutLMv3 error messaging vs Trainer GPU broadcast of NoneType." + }, + { + "left": "issue:44977", + "right": "issue:45229", + "accept": false, + "reason": "Different problems: flash-attention generation bug vs multi-GPU OOM." + }, + { + "left": "issue:43335", + "right": "issue:43653", + "accept": false, + "reason": "SwitchTransformers config bug and BigBirdTokenizer special-token registration are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44493", + "accept": false, + "reason": "Siglip2 zero probabilities and position-id key warnings are different symptoms in different code paths." + }, + { + "left": "issue:43335", + "right": "issue:45083", + "accept": false, + "reason": "Sparse-layer construction bug vs multimodal feature-length helper behavior; no shared underlying bug." + }, + { + "left": "issue:43335", + "right": "issue:43504", + "accept": false, + "reason": "Different models and failure modes: SwitchTransformers config vs Beit legacy-field loading." + }, + { + "left": "issue:43550", + "right": "issue:44112", + "accept": false, + "reason": "Bamba torch.compile/SDPA issue is unrelated to a stale GraniteSpeech CI test." + }, + { + "left": "issue:43335", + "right": "issue:45072", + "accept": false, + "reason": "SwitchTransformers sparse-layer bug and bfloat16 dtype mismatch in CI are unrelated." + }, + { + "left": "issue:44568", + "right": "issue:45491", + "accept": false, + "reason": "Tokenizer special-token addition bug is unrelated to Gemma3 NaN embeddings from sliding-window attention." + }, + { + "left": "issue:43720", + "right": "issue:44112", + "accept": false, + "reason": "BitNet packed-weight loading issue is unrelated to GraniteSpeech device-override CI failure." + }, + { + "left": "issue:44448", + "right": "issue:44977", + "accept": false, + "reason": "Pegasus output differences across versions and Qwen3.5 flash-attention failure are distinct bugs." + }, + { + "left": "issue:40444", + "right": "issue:43526", + "accept": false, + "reason": "Qwen2.5-VL multi-image dataset failure and BeitImageProcessorFast label reduction bug are unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44554", + "accept": false, + "reason": "Tokenizer loading regression and MPS attention correctness issue do not share the same code path." + }, + { + "left": "issue:43335", + "right": "issue:44112", + "accept": false, + "reason": "SwitchTransformers config bug and GraniteSpeech stale test failure are unrelated." + }, + { + "left": "issue:34689", + "right": "issue:35141", + "accept": false, + "reason": "Model loading regression for Llama vision and token-embedding resizing/reinit bug are different issues." + }, + { + "left": "issue:45372", + "right": "issue:45406", + "accept": false, + "reason": "Gemma4 processor import dependency failure and Gemma4Processor missing _tokenizer are separate loading bugs." + }, + { + "left": "issue:43606", + "right": "issue:44857", + "accept": false, + "reason": "CPU-offload device mismatch and float16 AMP crash are different failure modes." + }, + { + "left": "issue:44857", + "right": "issue:45072", + "accept": false, + "reason": "Different bugs: LwDetr loss AMP crash vs dtype mismatch in SwitchTransformers/TimmWrapperModel inference." + }, + { + "left": "issue:43066", + "right": "issue:44462", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 is unrelated to AutoTokenizer ignoring tokenizer.json." + }, + { + "left": "issue:43873", + "right": "issue:44493", + "accept": false, + "reason": "Quantization/offloading behavior and unexpected position-id keys are unrelated." + }, + { + "left": "issue:43873", + "right": "issue:45229", + "accept": false, + "reason": "Quantization offloading issue is not the same as Gemma4 multi-GPU OOM." + }, + { + "left": "issue:37428", + "right": "issue:44987", + "accept": false, + "reason": "Flash-attention import error and loading a specific model in v5 are not the same bug." + }, + { + "left": "issue:43278", + "right": "issue:43329", + "accept": false, + "reason": "Embedding dtype drift and multimodal token-count helper bug are unrelated." + }, + { + "left": "issue:43701", + "right": "issue:45083", + "accept": false, + "reason": "Checkpoint resume key mismatch and qwen3_omni_moe length helper issue are different code paths." + }, + { + "left": "issue:29942", + "right": "issue:35141", + "accept": false, + "reason": "Flash Attention 2 test failures are unrelated to output embedding reinitialization after resizing embeddings." + }, + { + "left": "issue:41720", + "right": "issue:45229", + "accept": false, + "reason": "cudaErrorAssert in Qwen3 auto device mapping is unrelated to Gemma4 OOM." + }, + { + "left": "issue:43550", + "right": "issue:45083", + "accept": false, + "reason": "Bamba SDPA compile failure and qwen3_omni_moe helper behavior are unrelated." + }, + { + "left": "issue:44451", + "right": "issue:45520", + "accept": false, + "reason": "ScandiBERT loading failure and flash_attn KeyError on Python 3.13 are different import/load issues." + }, + { + "left": "issue:43976", + "right": "issue:45440", + "accept": false, + "reason": "Python version compatibility bug and DeepseekV3MoE divergence are unrelated." + }, + { + "left": "issue:32090", + "right": "issue:33357", + "accept": false, + "reason": "Trainer NoneType broadcast error and MacOS bus error on CLIP are different runtime failures." + }, + { + "left": "issue:45372", + "right": "issue:45459", + "accept": false, + "reason": "Gemma4 processor import failure and tokenizer error masking when protobuf is absent are separate issues." + }, + { + "left": "issue:44448", + "right": "issue:44466", + "accept": false, + "reason": "Pegasus output differences and lm_head.weight serialization inconsistency are not the same bug." + }, + { + "left": "issue:43065", + "right": "issue:45412", + "accept": false, + "reason": "Dummy Conv2d in Sam3PixelDecoder and RT-DETR memory leak are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:43994", + "accept": false, + "reason": "Sam3PixelDecoder dummy layer bug and Siglip2 nonsensical outputs are unrelated." + }, + { + "left": "issue:25251", + "right": "issue:32090", + "accept": false, + "reason": "Pipeline top_k nesting bug and Trainer GPU broadcast NoneType 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": 13, + "review_unit_index": 22, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:29127", + "issue:29942", + "issue:30064", + "issue:32090", + "issue:33290", + "issue:33357", + "issue:34567", + "issue:35141", + "issue:36010", + "issue:39692", + "issue:42915", + "issue:42994", + "issue:43257", + "issue:43262", + "issue:43278", + "issue:43334", + "issue:43381", + "issue:43421", + "issue:43504", + "issue:43526", + "issue:43531", + "issue:43653", + "issue:43701", + "issue:43756", + "issue:43824", + "issue:43994", + "issue:44186", + "issue:44206", + "issue:44265", + "issue:44279", + "issue:44361", + "issue:44442", + "issue:44464", + "issue:44479", + "issue:44610", + "issue:44743", + "issue:44857", + "issue:44871", + "issue:44977", + "issue:45072", + "issue:45103", + "issue:45198", + "issue:45229", + "issue:45245", + "issue:45372", + "issue:45412", + "issue:45440", + "issue:45491" + ], + "soft_pairs": [ + "issue:44186|issue:45198", + "issue:43278|issue:43381", + "issue:43653|issue:44743", + "issue:44265|issue:45198", + "issue:44361|issue:45198", + "issue:43257|issue:43278", + "issue:42915|issue:42994", + "issue:43421|issue:45491", + "issue:43994|issue:45412", + "issue:44871|issue:44977", + "issue:44206|issue:44479", + "issue:43756|issue:45440", + "issue:43504|issue:44857", + "issue:44610|issue:45412", + "issue:29942|issue:30064", + "issue:45103|issue:45372", + "issue:33357|issue:39692", + "issue:29127|issue:34567", + "issue:32090|issue:33290", + "issue:45072|issue:45229", + "issue:43824|issue:44279", + "issue:43381|issue:43701", + "issue:43653|issue:44442", + "issue:43334|issue:43531", + "issue:43526|issue:44186", + "issue:44464|issue:45491", + "issue:43262|issue:44206", + "issue:44977|issue:45245", + "issue:29942|issue:36010", + "issue:29127|issue:33357", + "issue:29127|issue:35141" + ], + "prepared_review_unit_hash": "b83316dc1edb8d9e1d660aece1e253b350628850b2906f8d7734d2ca59b65348", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 31, + "serialized_chars": 37060, + "estimated_input_tokens": 9265, + "estimated_eval_tokens": 18786 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is a loose semantic grab bag across unrelated bug reports (tokenizers, vision processors, training/runtime, generation, and model-loading issues). None of the soft pairs look like the same underlying defect, so all soft edges are rejected.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue: the items span unrelated bugs in different subsystems, and the soft pairs do not read as true duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "No clear best issue representative for the same reason; choosing any one would be arbitrary across these unrelated reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44186", + "right": "issue:45198", + "accept": false, + "reason": "Different bugs: LayoutLMv2 NER/padding crash vs Wav2Vec2 save_pretrained/tokenization failure." + }, + { + "left": "issue:43278", + "right": "issue:43381", + "accept": false, + "reason": "Embedding dtype regression and eval-mode gradient-checkpointing are unrelated failures." + }, + { + "left": "issue:43653", + "right": "issue:44743", + "accept": false, + "reason": "BigBird tokenizer special-token decode issue vs Qwen3-MoE recurrent-state reset bug." + }, + { + "left": "issue:44265", + "right": "issue:45198", + "accept": false, + "reason": "torch.export failure with torch_compilable_check is unrelated to Wav2Vec2 save/tokenization errors." + }, + { + "left": "issue:44361", + "right": "issue:45198", + "accept": false, + "reason": "MLuke tokenizer AttributeError is a different code path from Wav2Vec2 save_pretrained/tokenization." + }, + { + "left": "issue:43257", + "right": "issue:43278", + "accept": false, + "reason": "Qwen3 MoE weight conversion under accelerate/deepspeed is not the same as BF16-to-FP32 embedding dtype drift." + }, + { + "left": "issue:42915", + "right": "issue:42994", + "accept": false, + "reason": "FineGrainedFP8 failure on Qwen3Moe is distinct from quantized model saving." + }, + { + "left": "issue:43421", + "right": "issue:45491", + "accept": false, + "reason": "Runtime special-token post-processor updates and Gemma3 NaN embeddings are different problems." + }, + { + "left": "issue:43994", + "right": "issue:45412", + "accept": false, + "reason": "SigLIP2 inference mismatch/quantization issue is unrelated to RT-DETR memory release behavior." + }, + { + "left": "issue:44871", + "right": "issue:44977", + "accept": false, + "reason": "Gemma-3 eos_token_id inconsistency is separate from Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:44206", + "right": "issue:44479", + "accept": false, + "reason": "Unsupported center arg in LasrFeatureExtractor is unrelated to the Qwen video-input regression." + }, + { + "left": "issue:43756", + "right": "issue:45440", + "accept": false, + "reason": "Smollm3 RoPE layer omission and DeepseekV3 divergence are different model-specific bugs." + }, + { + "left": "issue:43504", + "right": "issue:44857", + "accept": false, + "reason": "BEiT legacy-field loading failure is unrelated to LwDetrImageLoss AMP/CUDA crashes." + }, + { + "left": "issue:44610", + "right": "issue:45412", + "accept": false, + "reason": "Processor size mismatch for OmDet-Turbo is not the same as RT-DETR memory leakage." + }, + { + "left": "issue:29942", + "right": "issue:30064", + "accept": false, + "reason": "Flash Attention 2 test failures and void segmentation-map processing are unrelated." + }, + { + "left": "issue:45103", + "right": "issue:45372", + "accept": false, + "reason": "auto_docstring annotation handling and mistral_common import failures have different root causes." + }, + { + "left": "issue:33357", + "right": "issue:39692", + "accept": false, + "reason": "MacOS bus error for CLIP is unrelated to SigLIP2 documentation/model-processor mismatch." + }, + { + "left": "issue:29127", + "right": "issue:34567", + "accept": false, + "reason": "LayoutLMv3 error messaging and TrainerState token counter updating are different issues." + }, + { + "left": "issue:32090", + "right": "issue:33290", + "accept": false, + "reason": "NoneType broadcast in Trainer is unrelated to Adafactor/deepspeed OOM." + }, + { + "left": "issue:45072", + "right": "issue:45229", + "accept": false, + "reason": "bfloat16 inference dtype mismatches do not match a Gemma4 multi-GPU OOM." + }, + { + "left": "issue:43824", + "right": "issue:44279", + "accept": false, + "reason": "Qwen2.5-VL import error is broader than the dependency issue and not the same concrete bug." + }, + { + "left": "issue:43381", + "right": "issue:43701", + "accept": false, + "reason": "Gradient checkpointing in eval mode is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43653", + "right": "issue:44442", + "accept": false, + "reason": "BigBirdTokenizer decode output bug and AutoTokenizer loading FastSpeech2ConformerTokenizer are distinct." + }, + { + "left": "issue:43334", + "right": "issue:43531", + "accept": false, + "reason": "Qwen3-VL pad_token_id import failure and Qwen3-MoE sliding-window bug are different code paths." + }, + { + "left": "issue:43526", + "right": "issue:44186", + "accept": false, + "reason": "BeitImageProcessorFast reduce_labels bug is unrelated to LayoutLMv2 tokenizer crashes." + }, + { + "left": "issue:44464", + "right": "issue:45491", + "accept": false, + "reason": "Chunked generation with compiled forward and Gemma3 sliding-window NaNs are different generation bugs." + }, + { + "left": "issue:43262", + "right": "issue:44206", + "accept": false, + "reason": "Audio chat-template sampling-rate defaulting is unrelated to the LasrFeatureExtractor center-arg crash." + }, + { + "left": "issue:44977", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3.5 flash-attention generation and category-count overflow are unrelated." + }, + { + "left": "issue:29942", + "right": "issue:36010", + "accept": false, + "reason": "Flash Attention 2 tests do not match the GenerationMixin import regression." + }, + { + "left": "issue:29127", + "right": "issue:33357", + "accept": false, + "reason": "LayoutLMv3 error clarity and CLIP MacOS bus error are different artifacts." + }, + { + "left": "issue:29127", + "right": "issue:35141", + "accept": false, + "reason": "LayoutLMv3 error messaging and token-embedding reinitialization in post_init 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": 13, + "review_unit_index": 23, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:29942", + "issue:33290", + "issue:33666", + "issue:34567", + "issue:34689", + "issue:35141", + "issue:39692", + "issue:41628", + "issue:42915", + "issue:43232", + "issue:43316", + "issue:43381", + "issue:43388", + "issue:43493", + "issue:43525", + "issue:43526", + "issue:43575", + "issue:43611", + "issue:43646", + "issue:43673", + "issue:43701", + "issue:43723", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43931", + "issue:43994", + "issue:44077", + "issue:44079", + "issue:44188", + "issue:44206", + "issue:44387", + "issue:44442", + "issue:44479", + "issue:44610", + "issue:44743", + "issue:44811", + "issue:44857", + "issue:44898", + "issue:44964", + "issue:45003", + "issue:45072", + "issue:45216", + "issue:45290", + "issue:45372", + "issue:45440", + "issue:45479", + "issue:45491" + ], + "soft_pairs": [ + "issue:43761|issue:44079", + "issue:44442|issue:45072", + "issue:42915|issue:43931", + "issue:43232|issue:43388", + "issue:43526|issue:44857", + "issue:43526|issue:44610", + "issue:44479|issue:45290", + "issue:45003|issue:45440", + "issue:43723|issue:45479", + "issue:44743|issue:44811", + "issue:43575|issue:43701", + "issue:43316|issue:43525", + "issue:43931|issue:45216", + "issue:43646|issue:44077", + "issue:43761|issue:44188", + "issue:44964|issue:45440", + "issue:43575|issue:44387", + "issue:43994|issue:44206", + "issue:29942|issue:33666", + "issue:39692|issue:45491", + "issue:43673|issue:44743", + "issue:44442|issue:44898", + "issue:33290|issue:34567", + "issue:34689|issue:43611", + "issue:43493|issue:43827", + "issue:43381|issue:43873", + "issue:44077|issue:44479", + "issue:33290|issue:34689", + "issue:33290|issue:35141", + "issue:41628|issue:45372" + ], + "prepared_review_unit_hash": "14a1f2fbdb367af83375774584380cdb5083a00a41c9049d164d92c6add0ce5c", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 30, + "serialized_chars": 36720, + "estimated_input_tokens": 9180, + "estimated_eval_tokens": 18616 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a loose grab-bag of unrelated Transformers issues spanning model loading, generation, quantization, processors, docs, and training behavior. None of the soft pairs look like true duplicates or the same underlying fix, so all soft edges are rejected. The most umbrella-like issue is the generic Transformers 5.0.0 custom model initialization regression.", + "confidence": 0.27, + "canonical_issue_reason": "issue:43646 is the broadest, most umbrella-style report in the set and best fits as a weak cluster representative, even though the cluster items are not true duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43646 is the least specific and most generally applicable issue here, so it is the best single issue to anchor the cluster despite weak semantic cohesion.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:44079", + "accept": false, + "reason": "Both touch output/model-result handling, but one is a CLIPVisionModel regression and the other is a generic ModelOutput key assignment bug; different code paths." + }, + { + "left": "issue:44442", + "right": "issue:45072", + "accept": false, + "reason": "Tokenizer loading for FastSpeech2Conformer is unrelated to dtype mismatch failures in SwitchTransformers/TimmWrapperModel." + }, + { + "left": "issue:42915", + "right": "issue:43931", + "accept": false, + "reason": "Both are Qwen-related loading problems, but one is a FineGrainedFP8 config failure and the other is a shape mismatch for a different model family; not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:43388", + "accept": false, + "reason": "Generation kwargs synchronization and metric gathering are unrelated subsystems with different failure modes." + }, + { + "left": "issue:43526", + "right": "issue:44857", + "accept": false, + "reason": "Beit label reduction and LwDetr AMP/CUDA crashes are separate image-processing vs loss-computation issues." + }, + { + "left": "issue:43526", + "right": "issue:44610", + "accept": false, + "reason": "Both involve vision processors, but one is a label-reduction bug and the other is a processor/model size mismatch; different fixes." + }, + { + "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." + }, + { + "left": "issue:45003", + "right": "issue:45440", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils is unrelated to DeepseekV3MoE implementation divergence." + }, + { + "left": "issue:43723", + "right": "issue:45479", + "accept": false, + "reason": "Tokenizer loading in v5 and zero-loss sequence classification with num_labels=1 are different bugs in different APIs." + }, + { + "left": "issue:44743", + "right": "issue:44811", + "accept": false, + "reason": "Recurrent-state cache reset and Whisper batch_decode skip_special_tokens handling are unrelated." + }, + { + "left": "issue:43575", + "right": "issue:43701", + "accept": false, + "reason": "Tensor-parallel OOM when loading Qwen2-57B is unrelated to resume_from_checkpoint key mismatch." + }, + { + "left": "issue:43316", + "right": "issue:43525", + "accept": false, + "reason": "Both are config API discrepancies, but they concern different models and different missing attributes." + }, + { + "left": "issue:43931", + "right": "issue:45216", + "accept": false, + "reason": "Model weight-shape mismatch during load is not the same as an incorrect save_pretrained checkpoint regression." + }, + { + "left": "issue:43646", + "right": "issue:44077", + "accept": false, + "reason": "Generic custom-model initialization breakage and patchtsmixer post_init policy are not the same concrete bug." + }, + { + "left": "issue:43761", + "right": "issue:44188", + "accept": false, + "reason": "CLIP hidden_states regression and attention-kernel divergence under torch.compile are separate issues." + }, + { + "left": "issue:44964", + "right": "issue:45440", + "accept": false, + "reason": "Phi-4 multimodal loading failure and DeepSeekV3MoE divergence are unrelated model-specific problems." + }, + { + "left": "issue:43575", + "right": "issue:44387", + "accept": false, + "reason": "OOM from tensor-parallel loading and increased reserved memory under int4 quantization are both memory-related but not the same bug." + }, + { + "left": "issue:43994", + "right": "issue:44206", + "accept": false, + "reason": "SigLIP2 nonsensical outputs and LasrFeatureExtractor center-arg crash are different failure modes." + }, + { + "left": "issue:29942", + "right": "issue:33666", + "accept": false, + "reason": "Flash Attention 2 test failures and Qwen2-VL multi-GPU training are unrelated." + }, + { + "left": "issue:39692", + "right": "issue:45491", + "accept": false, + "reason": "SigLIP2 docs/example problems and Gemma3 mixed-length NaN embeddings are unrelated." + }, + { + "left": "issue:43673", + "right": "issue:44743", + "accept": false, + "reason": "Both mention cache, but one is a GenerationMixin cache regression and the other is Qwen3_5 recurrent-state behavior; not the same fix." + }, + { + "left": "issue:44442", + "right": "issue:44898", + "accept": false, + "reason": "FastSpeech2Conformer tokenizer loading and Perceiver resolution handling are unrelated vision/audio issues." + }, + { + "left": "issue:33290", + "right": "issue:34567", + "accept": false, + "reason": "Adafactor/deepspeed OOM and TrainerState token-count tracking are completely different bugs." + }, + { + "left": "issue:34689", + "right": "issue:43611", + "accept": false, + "reason": "Model loading breakage for a specific Llama vision model and a base_model_prefix regression are different loading paths and root causes." + }, + { + "left": "issue:43493", + "right": "issue:43827", + "accept": false, + "reason": "SigLIP2 implementation discrepancy and docs still referencing removed pipeline() APIs do not share a concrete code-path problem." + }, + { + "left": "issue:43381", + "right": "issue:43873", + "accept": false, + "reason": "Gradient checkpointing in eval mode and quantization offloading issues are unrelated training/runtime bugs." + }, + { + "left": "issue:44077", + "right": "issue:44479", + "accept": false, + "reason": "patchtsmixer post_init policy and Qwen VL video-input regression are unrelated." + }, + { + "left": "issue:33290", + "right": "issue:34689", + "accept": false, + "reason": "OOM in optimizer setup and model-loading failure for Llama vision are different failure classes." + }, + { + "left": "issue:33290", + "right": "issue:35141", + "accept": false, + "reason": "Optimizer OOM and embedding reinitialization in post_init are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45372", + "accept": false, + "reason": "Import errors for AutoImageProcessor and ReasoningEffort are distinct symbols and contexts." + } + ] + }, + "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": 13, + "review_unit_index": 24, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:16998", + "issue:22355", + "issue:25251", + "issue:28282", + "issue:33290", + "issue:34689", + "issue:36010", + "issue:36331", + "issue:39692", + "issue:41950", + "issue:42175", + "issue:43065", + "issue:43295", + "issue:43316", + "issue:43381", + "issue:43388", + "issue:43404", + "issue:43441", + "issue:43475", + "issue:43540", + "issue:43643", + "issue:43688", + "issue:43716", + "issue:43720", + "issue:43749", + "issue:43761", + "issue:43825", + "issue:43856", + "issue:43873", + "issue:43931", + "issue:44162", + "issue:44190", + "issue:44206", + "issue:44442", + "issue:44479", + "issue:44496", + "issue:44514", + "issue:44617", + "issue:44734", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44877", + "issue:44913", + "issue:45127", + "issue:45137", + "issue:45200", + "issue:45406" + ], + "soft_pairs": [ + "issue:43720|issue:44442", + "issue:43749|issue:43931", + "issue:43643|issue:44913", + "issue:43873|issue:44871", + "issue:39692|issue:44206", + "issue:44190|issue:44496", + "issue:33290|issue:36331", + "issue:43381|issue:43856", + "issue:43316|issue:44877", + "issue:44855|issue:45137", + "issue:43540|issue:44514", + "issue:34689|issue:36010", + "issue:42175|issue:43316", + "issue:44162|issue:44792", + "issue:43388|issue:43688", + "issue:43404|issue:43441", + "issue:43316|issue:43716", + "issue:16998|issue:22355", + "issue:25251|issue:28282", + "issue:41950|issue:43475", + "issue:43065|issue:44617", + "issue:45127|issue:45200", + "issue:43295|issue:43825", + "issue:43761|issue:44479", + "issue:44734|issue:45406" + ], + "prepared_review_unit_hash": "330bd7d447d69758e69e984510d677329bdcaef5bf4574a575121d1ffe5242e7", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 25, + "serialized_chars": 35228, + "estimated_input_tokens": 8807, + "estimated_eval_tokens": 17870 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: the items cover unrelated bugs across model loading, training, pipelines, configs, and docs, so there is no credible duplicate canonical here. All soft-edge pairs should be rejected.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43720", + "right": "issue:44442", + "accept": false, + "reason": "Different failures: one is BitNet packed-weight unpacking during accelerate loading, the other is AutoTokenizer loading a FastSpeech2Conformer tokenizer." + }, + { + "left": "issue:43749", + "right": "issue:43931", + "accept": false, + "reason": "Different code paths and symptoms: FSDP CPU RAM-efficient loading vs a Qwen3-VL model weight-shape mismatch." + }, + { + "left": "issue:43643", + "right": "issue:44913", + "accept": false, + "reason": "Both are config-related, but one is trust_remote_code returning incomplete AutoConfig fields and the other is GPTNeoX rotary_pct not persisting on reload." + }, + { + "left": "issue:43873", + "right": "issue:44871", + "accept": false, + "reason": "Quantization/offloading behavior is unrelated to Gemma-3 eos_token_id configuration inconsistency." + }, + { + "left": "issue:39692", + "right": "issue:44206", + "accept": false, + "reason": "A documentation example with model/processor mismatch is not the same bug as LaserFeatureExtractor passing an unsupported center argument." + }, + { + "left": "issue:44190", + "right": "issue:44496", + "accept": false, + "reason": "Different subsystems: local dataset loading in an image-classification training script vs config loading failure for an unrecognized model." + }, + { + "left": "issue:33290", + "right": "issue:36331", + "accept": false, + "reason": "One is OOM with Adafactor under DeepSpeed; the other is a CustomTrainer method signature change." + }, + { + "left": "issue:43381", + "right": "issue:43856", + "accept": false, + "reason": "Gradient checkpointing in eval mode and Qwen3 MoE memory usage are different training behaviors, not one underlying bug." + }, + { + "left": "issue:43316", + "right": "issue:44877", + "accept": false, + "reason": "Both mention config APIs, but the former is a Gemma3TextConfig discrepancy and the latter is strict config blocking granite_speech loading." + }, + { + "left": "issue:44855", + "right": "issue:45137", + "accept": false, + "reason": "Python 3.13 torch.jit parsing/IndentationError is unrelated to a DeepSpeed ZeRO3 deque pop error." + }, + { + "left": "issue:43540", + "right": "issue:44514", + "accept": false, + "reason": "Different processor/input bugs: Qwen3OmniMoe video processing ValueError vs batched chat-template padding crash in Qwen2_5_VLProcessor." + }, + { + "left": "issue:34689", + "right": "issue:36010", + "accept": false, + "reason": "Model-loading regression for Llama 3.2 Vision is unrelated to the GenerationMixin import error." + }, + { + "left": "issue:42175", + "right": "issue:43316", + "accept": false, + "reason": "Package/backend installation issue with pip install '.[torch]' is unrelated to a Gemma3TextConfig API discrepancy." + }, + { + "left": "issue:44162", + "right": "issue:44792", + "accept": false, + "reason": "ESM2 being broken and a janus image-generation test failure are different models and failure modes." + }, + { + "left": "issue:43388", + "right": "issue:43688", + "accept": false, + "reason": "Gathering metrics drops tuple labels in the last batch, while auxiliary-loss normalization is a separate training-loss bug." + }, + { + "left": "issue:43404", + "right": "issue:43441", + "accept": false, + "reason": "lm_head weight tying in Mistral3ForConditionalGeneration is a different concrete issue than FlashAttention failures in Ministral-3." + }, + { + "left": "issue:43316", + "right": "issue:43716", + "accept": false, + "reason": "Gemma3TextConfig API mismatch is unrelated to the Mistral-3 image-preprocessor dtype mismatch." + }, + { + "left": "issue:16998", + "right": "issue:22355", + "accept": false, + "reason": "A model_max_length question for DeBERTa-V3 is unrelated to a missing transformers.onnx import." + }, + { + "left": "issue:25251", + "right": "issue:28282", + "accept": false, + "reason": "Pipeline top_k output shape change is unrelated to AutoModel requiring PyTorch to be installed." + }, + { + "left": "issue:41950", + "right": "issue:43475", + "accept": false, + "reason": "Video-classification pipeline using image processors is a different bug from Sam3VisionEncoderOutput missing fpn_position_embeddings." + }, + { + "left": "issue:43065", + "right": "issue:44617", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d and Sam3Video CUDA OOM are both SAM3-related but not the same underlying change or code path." + }, + { + "left": "issue:45127", + "right": "issue:45200", + "accept": false, + "reason": "LoRA merge collapse with tied embeddings is unrelated to Gemma 4 mm_token_type_ids defaulting behavior." + }, + { + "left": "issue:43295", + "right": "issue:43825", + "accept": false, + "reason": "processor.tokenizer regression in v4.57.5 is unrelated to pipeline translation-task error messaging in v5." + }, + { + "left": "issue:43761", + "right": "issue:44479", + "accept": false, + "reason": "CLIPVisionModel hidden_states regression and Qwen video-input regression are different models and different breakages." + }, + { + "left": "issue:44734", + "right": "issue:45406", + "accept": false, + "reason": "Transformers serve KV-cache continuation indexing is unrelated to Gemma4Processor missing _tokenizer." + } + ] + }, + "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": 13, + "review_unit_index": 25, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:32090", + "issue:35141", + "issue:41720", + "issue:42617", + "issue:43065", + "issue:43295", + "issue:43388", + "issue:43452", + "issue:43525", + "issue:43531", + "issue:43550", + "issue:43575", + "issue:43606", + "issue:43646", + "issue:43650", + "issue:43701", + "issue:43742", + "issue:43761", + "issue:43827", + "issue:43873", + "issue:43906", + "issue:43931", + "issue:43994", + "issue:44060", + "issue:44077", + "issue:44079", + "issue:44190", + "issue:44220", + "issue:44242", + "issue:44464", + "issue:44479", + "issue:44496", + "issue:44521", + "issue:44534", + "issue:44610", + "issue:44734", + "issue:44792", + "issue:44871", + "issue:44913", + "issue:44933", + "issue:44964", + "issue:44977", + "issue:45127", + "issue:45245", + "issue:45290", + "issue:45357", + "issue:45362", + "issue:45381" + ], + "soft_pairs": [ + "issue:44077|issue:44534", + "issue:43550|issue:43701", + "issue:43295|issue:44871", + "issue:43452|issue:44496", + "issue:43606|issue:43701", + "issue:43646|issue:44913", + "issue:43906|issue:45290", + "issue:43761|issue:44242", + "issue:42617|issue:43575", + "issue:43575|issue:43742", + "issue:43931|issue:44060", + "issue:44521|issue:45381", + "issue:43525|issue:43531", + "issue:43575|issue:43873", + "issue:44792|issue:44977", + "issue:44479|issue:44734", + "issue:44464|issue:44871", + "issue:32090|issue:35141", + "issue:44933|issue:45362", + "issue:43065|issue:44610", + "issue:43827|issue:43994", + "issue:44534|issue:44913", + "issue:41720|issue:45245", + "issue:43650|issue:44220", + "issue:44190|issue:44964", + "issue:43388|issue:44079", + "issue:45127|issue:45357" + ], + "prepared_review_unit_hash": "e2cc9d58ad9b1ebbc62d4188589a55c7f308e44ad4dfd28e28c10ed06ecba4c7", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 27, + "serialized_chars": 35662, + "estimated_input_tokens": 8916, + "estimated_eval_tokens": 18088 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These items are mostly unrelated issue reports across different regressions and subsystems. None of the soft pairs look like the same concrete bug, so I would not merge any of them as duplicates. If I had to anchor the cluster, I\u2019d use issue 44521 because it is current, well-documented, and has the strongest external signal.", + "confidence": 0.86, + "canonical_issue_reason": "issue:44521 is the strongest representative: it is open, detailed, and has multiple inbound references, making it the most useful anchor among these unrelated issues.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44521 is the best global issue candidate for triage because it has the clearest repro and the most activity/references.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44077", + "right": "issue:44534", + "accept": false, + "reason": "Both involve model init/buffer behavior in v5, but one is about post_init being disallowed for patchtsmixer and the other about junk non-persistent buffers; different bugs." + }, + { + "left": "issue:43550", + "right": "issue:43701", + "accept": false, + "reason": "One is a torch.compile/SDPA failure for Bamba, the other is a resume_from_checkpoint key mismatch; no shared code-path bug." + }, + { + "left": "issue:43295", + "right": "issue:44871", + "accept": false, + "reason": "Custom processor/tokenizer regression and Gemma eos_token_id inconsistency are distinct model/config issues." + }, + { + "left": "issue:43452", + "right": "issue:44496", + "accept": false, + "reason": "Both are loading-related, but gguf_file handling and missing model_type validation are different root causes." + }, + { + "left": "issue:43606", + "right": "issue:43701", + "accept": false, + "reason": "CPU offload device mismatch is unrelated to resume checkpoint key mismatch." + }, + { + "left": "issue:43646", + "right": "issue:44913", + "accept": false, + "reason": "Generic custom initialization breakage and GPTNeoX rotary_pct reload regression are separate problems." + }, + { + "left": "issue:43906", + "right": "issue:45290", + "accept": false, + "reason": "An isolated reproduction of one past issue does not match a chat-template tool-call crash; different functionality and failure mode." + }, + { + "left": "issue:43761", + "right": "issue:44242", + "accept": false, + "reason": "CLIP hidden_states regression and MoE router loss omission are unrelated model-output vs training-loss bugs." + }, + { + "left": "issue:42617", + "right": "issue:43575", + "accept": false, + "reason": "3d_parallel startup failure and tensor-parallel OOM are both parallelism-related but not the same concrete bug." + }, + { + "left": "issue:43575", + "right": "issue:43742", + "accept": false, + "reason": "OOM on a large Qwen2 model with tp is not the same as a key error loading MobileLLM-125M." + }, + { + "left": "issue:43931", + "right": "issue:44060", + "accept": false, + "reason": "Shape mismatch loading Qwen3-VL differs from a tied-weights warning in Qwen3-Next." + }, + { + "left": "issue:44521", + "right": "issue:45381", + "accept": false, + "reason": "Both touch multimodal chat/vision input handling, but assistant mask generation and video vision_position_ids are different code paths and symptoms." + }, + { + "left": "issue:43525", + "right": "issue:43531", + "accept": false, + "reason": "Missing pad_token_id in Llama4Config and Qwen3-MoE sliding_window behavior are unrelated config bugs." + }, + { + "left": "issue:43575", + "right": "issue:43873", + "accept": false, + "reason": "Tensor-parallel OOM and offloading/quantization behavior are different memory-management issues." + }, + { + "left": "issue:44792", + "right": "issue:44977", + "accept": false, + "reason": "Janus image-generation test failure and Qwen3.5 flash-attention generation issues are unrelated." + }, + { + "left": "issue:44479", + "right": "issue:44734", + "accept": false, + "reason": "Video-input regression in Qwen VL models and serve KV-cache continuation indexing are different failures." + }, + { + "left": "issue:44464", + "right": "issue:44871", + "accept": false, + "reason": "Compiled chunked generation inconsistency is not the same as a Gemma eos_token_id config mismatch." + }, + { + "left": "issue:32090", + "right": "issue:35141", + "accept": false, + "reason": "Trainer GPU broadcast NoneType error and embedding reinitialization after resize are different bugs." + }, + { + "left": "issue:44933", + "right": "issue:45362", + "accept": false, + "reason": "A missing import in image_utils and a Qwen3.5 chat crash are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:44610", + "accept": false, + "reason": "Sam3PixelDecoder dummy Conv2d and OmDet-Turbo processor image-size mismatch are different vision-model issues." + }, + { + "left": "issue:43827", + "right": "issue:43994", + "accept": false, + "reason": "Documentation still referencing pipeline() is unrelated to SigLIP2 producing bad outputs with AutoModel/pipeline." + }, + { + "left": "issue:44534", + "right": "issue:44913", + "accept": false, + "reason": "Non-persistent buffer junk and GPTNeoX rotary_pct not persisting are both v5 regressions, but not the same underlying serialization bug." + }, + { + "left": "issue:41720", + "right": "issue:45245", + "accept": false, + "reason": "Qwen3 auto device mapping cuda assert and category-count overflow are unrelated runtime errors." + }, + { + "left": "issue:43650", + "right": "issue:44220", + "accept": false, + "reason": "A vague data request is not the same as a fbank feature extraction bug." + }, + { + "left": "issue:44190", + "right": "issue:44964", + "accept": false, + "reason": "Local dataset loading in a training script and loading a multimodal model are different loading paths and failures." + }, + { + "left": "issue:43388", + "right": "issue:44079", + "accept": false, + "reason": "Dropping label elements in gather_for_metrics and ModelOutput key assignment with None are distinct tuple/ModelOutput handling issues." + }, + { + "left": "issue:45127", + "right": "issue:45357", + "accept": false, + "reason": "LoRA merge with extended vocabulary causing collapse and incorrect visual encoder keys in save_pretrained are separate model-saving bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 26, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:36246", + "issue:38175", + "issue:38617", + "issue:39692", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:43065", + "issue:43116", + "issue:43295", + "issue:43299", + "issue:43335", + "issue:43404", + "issue:43479", + "issue:43540", + "issue:43550", + "issue:43643", + "issue:43644", + "issue:43761", + "issue:43854", + "issue:43976", + "issue:44016", + "issue:44077", + "issue:44079", + "issue:44112", + "issue:44186", + "issue:44190", + "issue:44220", + "issue:44246", + "issue:44279", + "issue:44355", + "issue:44466", + "issue:44496", + "issue:44683", + "issue:44734", + "issue:44743", + "issue:44792", + "issue:44855", + "issue:44871", + "issue:44913", + "issue:45003", + "issue:45216", + "issue:45276", + "issue:45335", + "issue:45372", + "issue:45375", + "issue:45446" + ], + "soft_pairs": [ + "issue:43761|issue:44743", + "issue:44355|issue:44855", + "issue:43644|issue:44913", + "issue:44496|issue:45372", + "issue:43976|issue:44855", + "issue:43540|issue:44734", + "issue:44077|issue:44079", + "issue:43550|issue:44190", + "issue:43295|issue:45276", + "issue:43335|issue:44855", + "issue:43404|issue:45216", + "issue:43116|issue:44190", + "issue:41950|issue:44016", + "issue:36246|issue:41628", + "issue:39692|issue:44190", + "issue:43644|issue:44871", + "issue:44186|issue:44734", + "issue:43479|issue:45375", + "issue:36246|issue:38617", + "issue:44016|issue:44220", + "issue:44246|issue:45003", + "issue:38175|issue:43065", + "issue:44466|issue:44913", + "issue:44683|issue:45446", + "issue:42175|issue:44279", + "issue:44112|issue:44792", + "issue:43854|issue:44792", + "issue:43295|issue:45335", + "issue:43643|issue:44496", + "issue:43299|issue:45216" + ], + "prepared_review_unit_hash": "64143b45b764f86bec0d77ee3c552500287cec0ca90c1dfadde5fd942773908b", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 30, + "serialized_chars": 36364, + "estimated_input_tokens": 9091, + "estimated_eval_tokens": 18438 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This set is a loose collection of unrelated Transformers issues; none of the soft-edge pairs look like the same underlying bug or change, so all are rejected. The best anchor issue is the broad v5 regression about non-persistent buffers being filled incorrectly, which is the closest to the recurring serialization/regression theme in the set.", + "confidence": 0.88, + "canonical_issue_reason": "issue:43644 is the strongest representative of the cluster\u2019s repeated v5/regression/serialization theme and is broad enough to serve as an anchor, even though the other items are not true duplicates of it.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43644 has the most general core-runtime impact and is more central than the narrower model- or example-specific reports elsewhere in the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43761", + "right": "issue:44743", + "accept": false, + "reason": "Different models and different failures: CLIPVisionModel hidden_states regression vs Qwen3.5 recurrent-state reset with cache." + }, + { + "left": "issue:44355", + "right": "issue:44855", + "accept": false, + "reason": "Both are Python/runtime-related, but one is about compiled Python files and the other is a DebertaV2 import parsing issue on Python 3.13; not the same bug." + }, + { + "left": "issue:43644", + "right": "issue:44913", + "accept": false, + "reason": "Both involve reload/serialization behavior, but one is about non-persistent buffers getting junk values and the other is GPTNeoX rotary_pct not persisting; different code paths." + }, + { + "left": "issue:44496", + "right": "issue:45372", + "accept": false, + "reason": "Both are model-loading/import problems, but they concern different missing symbols and different models; not the same concrete issue." + }, + { + "left": "issue:43976", + "right": "issue:44855", + "accept": false, + "reason": "General Python-version compatibility complaints are not enough; these point to different failure modes and different code paths." + }, + { + "left": "issue:43540", + "right": "issue:44734", + "accept": false, + "reason": "Qwen3OmniMoe video-input validation and transformers serve KV-cache indexing are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:44079", + "accept": false, + "reason": "One is a config/post_init rule for patchtsmixer, the other is ModelOutput key assignment logic; different bugs." + }, + { + "left": "issue:43550", + "right": "issue:44190", + "accept": false, + "reason": "torch.compile+SDPA failure in Bamba is unrelated to the image-classification no-trainer local dataset loading issue." + }, + { + "left": "issue:43295", + "right": "issue:45276", + "accept": false, + "reason": "Processor/tokenizer regression and resize_token_embeddings propagation are different problems." + }, + { + "left": "issue:43335", + "right": "issue:44855", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug is unrelated to the Python 3.13 import parsing error." + }, + { + "left": "issue:43404", + "right": "issue:45216", + "accept": false, + "reason": "Both mention weight tying/checkpointing, but they affect different models and different failure modes; not mergeable as one bug." + }, + { + "left": "issue:43116", + "right": "issue:44190", + "accept": false, + "reason": "Example script multi-label output bug vs local dataset loading failure in a different script." + }, + { + "left": "issue:41950", + "right": "issue:44016", + "accept": false, + "reason": "Video-classification pipeline processor lookup and notebook syntax error are unrelated." + }, + { + "left": "issue:36246", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors around image processor APIs, but they are different missing names and not the same underlying breakage." + }, + { + "left": "issue:39692", + "right": "issue:44190", + "accept": false, + "reason": "Docs example errors for SigLIP2 are unrelated to the no-trainer image-classification dataset issue." + }, + { + "left": "issue:43644", + "right": "issue:44871", + "accept": false, + "reason": "Buffer serialization corruption and Gemma3 eos_token_id mismatch are different issues." + }, + { + "left": "issue:44186", + "right": "issue:44734", + "accept": false, + "reason": "LayoutLMv2Tokenizer NER/padding crash is unrelated to serve KV-cache continuation indexing." + }, + { + "left": "issue:43479", + "right": "issue:45375", + "accept": false, + "reason": "Different multimodal config bugs in different models; not the same defect." + }, + { + "left": "issue:36246", + "right": "issue:38617", + "accept": false, + "reason": "Qwen2_5_VLImageProcessor import error and layer_type_validation import error are unrelated import failures." + }, + { + "left": "issue:44016", + "right": "issue:44220", + "accept": false, + "reason": "Notebook syntax error is unrelated to _torch_extract_fbank_features() behavior." + }, + { + "left": "issue:44246", + "right": "issue:45003", + "accept": false, + "reason": "Import slowdown and unsafe sys.modules access are different import-time issues." + }, + { + "left": "issue:38175", + "right": "issue:43065", + "accept": false, + "reason": "SigLIP2 zero probabilities and Sam3PixelDecoder dummy Conv2d are unrelated model-specific bugs." + }, + { + "left": "issue:44466", + "right": "issue:44913", + "accept": false, + "reason": "Device-dependent lm_head serialization inconsistency and rotary_pct reload regression are different serialization issues." + }, + { + "left": "issue:44683", + "right": "issue:45446", + "accept": false, + "reason": "Compiled flex_attention failure on torch>=2.9 and a wrong version check for AuxRequest import are related only by subsystem, not the same bug." + }, + { + "left": "issue:42175", + "right": "issue:44279", + "accept": false, + "reason": "Package extras/backend dependency issue is not the same as a generic dependency problem report." + }, + { + "left": "issue:44112", + "right": "issue:44792", + "accept": false, + "reason": "CI stale test failure in GraniteSpeech and janus image-generation test failure are unrelated." + }, + { + "left": "issue:43854", + "right": "issue:44792", + "accept": false, + "reason": "Unable to load a specific model in unit tests is not the same as a failed janus image-generation test." + }, + { + "left": "issue:43295", + "right": "issue:45335", + "accept": false, + "reason": "Different resize_token_embeddings propagation bugs in different models/code paths." + }, + { + "left": "issue:43643", + "right": "issue:44496", + "accept": false, + "reason": "trust_remote_code missing fields and unrecognized model_type are both config-loading issues, but they are distinct root causes." + }, + { + "left": "issue:43299", + "right": "issue:45216", + "accept": false, + "reason": "Both are Qwen-related regressions, but one is a load-time failure and the other is a save_pretrained checkpoint corruption; not the same concrete problem." + } + ] + }, + "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": 13, + "review_unit_index": 27, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:38175", + "issue:41950", + "issue:42907", + "issue:43295", + "issue:43299", + "issue:43441", + "issue:43479", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43577", + "issue:43650", + "issue:43673", + "issue:43824", + "issue:43854", + "issue:43901", + "issue:43976", + "issue:44016", + "issue:44062", + "issue:44075", + "issue:44112", + "issue:44242", + "issue:44279", + "issue:44297", + "issue:44336", + "issue:44351", + "issue:44368", + "issue:44462", + "issue:44464", + "issue:44496", + "issue:44521", + "issue:44617", + "issue:44779", + "issue:44871", + "issue:44877", + "issue:44933", + "issue:44964", + "issue:44987", + "issue:45020", + "issue:45042", + "issue:45081", + "issue:45200", + "issue:45216", + "issue:45245", + "issue:45278", + "issue:45325", + "issue:45357", + "issue:45405" + ], + "soft_pairs": [ + "issue:42907|issue:43441", + "issue:44016|issue:44062", + "issue:44112|issue:44964", + "issue:44871|issue:45245", + "issue:43673|issue:44336", + "issue:43976|issue:44987", + "issue:44496|issue:45042", + "issue:43526|issue:43901", + "issue:43824|issue:44933", + "issue:43299|issue:45357", + "issue:43976|issue:44279", + "issue:43577|issue:45216", + "issue:44075|issue:44368", + "issue:44521|issue:45325", + "issue:43901|issue:44496", + "issue:43854|issue:44964", + "issue:44242|issue:45200", + "issue:41950|issue:43650", + "issue:41950|issue:45020", + "issue:44464|issue:45245", + "issue:43976|issue:45405", + "issue:44351|issue:45278", + "issue:43479|issue:43525", + "issue:44462|issue:45081", + "issue:44496|issue:44877", + "issue:44297|issue:44462", + "issue:44279|issue:45405", + "issue:43550|issue:44964", + "issue:43299|issue:43854", + "issue:44462|issue:44779", + "issue:43295|issue:44779", + "issue:38175|issue:44617" + ], + "prepared_review_unit_hash": "6e57d4beb410dbc043d7a31442e98fbebf42ce6dc94e3ad7f87753ddb4194c9c", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 37250, + "estimated_input_tokens": 9313, + "estimated_eval_tokens": 18882 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is mostly a grab bag of unrelated Transformers bug reports (model loading, tokenizer/config regressions, docs, CI, and pipeline issues). None of the soft pairs look like the same concrete bug or fix.", + "confidence": 0.88, + "canonical_issue_reason": "issue:45020 is the broadest open, actionable report in the set and the closest umbrella for recent-version model-loading regressions involving remote_code.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45020 is the best global anchor because it is open, clearly user-facing, and broad enough to represent the regression theme without overfitting to a single model.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42907", + "right": "issue:43441", + "accept": false, + "reason": "Different problems: saving a dequantized model vs FlashAttention load failure for Ministral-3." + }, + { + "left": "issue:44016", + "right": "issue:44062", + "accept": false, + "reason": "Notebook syntax error and AddedToken keyword conflict are unrelated bugs." + }, + { + "left": "issue:44112", + "right": "issue:44964", + "accept": false, + "reason": "CI stale device-override test in GraniteSpeech is not the same as Phi-4 multimodal loading failure." + }, + { + "left": "issue:44871", + "right": "issue:45245", + "accept": false, + "reason": "eos_token_id config mismatch is unrelated to the categories-limit runtime error." + }, + { + "left": "issue:43673", + "right": "issue:44336", + "accept": false, + "reason": "Chunked-prefill cache regression and ANSI logging output are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44987", + "accept": false, + "reason": "Python-version compatibility issue is unrelated to loading physical-intelligence/fast." + }, + { + "left": "issue:44496", + "right": "issue:45042", + "accept": false, + "reason": "Unrecognized model_type and PIL backend torchvision requirement are different code paths." + }, + { + "left": "issue:43526", + "right": "issue:43901", + "accept": false, + "reason": "BeitImageProcessorFast label reduction bug is unrelated to a docs mismatch." + }, + { + "left": "issue:43824", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but they concern different symbols/modules and different failures." + }, + { + "left": "issue:43299", + "right": "issue:45357", + "accept": false, + "reason": "Qwen3VL loading failure and Qwen3.5 save_pretrained key regression are different bugs." + }, + { + "left": "issue:43976", + "right": "issue:44279", + "accept": false, + "reason": "Python version/support problem is unrelated to a vague dependency issue." + }, + { + "left": "issue:43577", + "right": "issue:45216", + "accept": false, + "reason": "Blip2 dtype propagation and Qwen3.5 checkpoint saving are unrelated." + }, + { + "left": "issue:44075", + "right": "issue:44368", + "accept": false, + "reason": "Unused SGD args and tie_word_embeddings warning are unrelated." + }, + { + "left": "issue:44521", + "right": "issue:45325", + "accept": false, + "reason": "Multimodal assistant mask bug and Qwen2.5-VL rope-index bug are different generation-path issues." + }, + { + "left": "issue:43901", + "right": "issue:44496", + "accept": false, + "reason": "Docs about return_all_scores do not match an unrecognized-model loading error." + }, + { + "left": "issue:43854", + "right": "issue:44964", + "accept": false, + "reason": "Different model families and distinct load failures; not the same bug." + }, + { + "left": "issue:44242", + "right": "issue:45200", + "accept": false, + "reason": "Load-balancing loss gating and mm_token_type_ids defaults are unrelated." + }, + { + "left": "issue:41950", + "right": "issue:43650", + "accept": false, + "reason": "Video-classification processor lookup is unrelated to the vague 'ADD THE DATA' issue." + }, + { + "left": "issue:41950", + "right": "issue:45020", + "accept": false, + "reason": "A pipeline processor-selection bug is not the same as remote_code model-loading regressions." + }, + { + "left": "issue:44464", + "right": "issue:45245", + "accept": false, + "reason": "Chunked generation inconsistency and category-count limit error are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:45405", + "accept": false, + "reason": "Python-version support and minimum PEFT version bump are unrelated dependency issues." + }, + { + "left": "issue:44351", + "right": "issue:45278", + "accept": false, + "reason": "A missing exported symbol and general import errors are related only at a very high level, not the same bug." + }, + { + "left": "issue:43479", + "right": "issue:43525", + "accept": false, + "reason": "Phi4MultimodalConfig default init and Llama4Config missing pad_token_id are different config bugs." + }, + { + "left": "issue:44462", + "right": "issue:45081", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and Mistral regex patch crashing are distinct tokenizer-loading paths." + }, + { + "left": "issue:44496", + "right": "issue:44877", + "accept": false, + "reason": "Unrecognized model_type and strict granite_speech config loading are different configuration issues." + }, + { + "left": "issue:44297", + "right": "issue:44462", + "accept": false, + "reason": "save_pretrained writing the wrong tokenizer_class is not the same as AutoTokenizer ignoring tokenizer.json." + }, + { + "left": "issue:44279", + "right": "issue:45405", + "accept": false, + "reason": "Generic dependency issue and PEFT version bump are not the same underlying problem." + }, + { + "left": "issue:43550", + "right": "issue:44964", + "accept": false, + "reason": "torch.compile + SDPA failure in Bamba is unrelated to Phi-4 multimodal loading." + }, + { + "left": "issue:43299", + "right": "issue:43854", + "accept": false, + "reason": "Qwen3VL model-loading regression and GLM-4.7-Flash load failure are different models and code paths." + }, + { + "left": "issue:44462", + "right": "issue:44779", + "accept": false, + "reason": "Tokenizer.json being ignored and Deepseek tokenizer incorrect results are different tokenizer regressions." + }, + { + "left": "issue:43295", + "right": "issue:44779", + "accept": false, + "reason": "Processor.tokenizer regression and Deepseek tokenizer output issues are unrelated." + }, + { + "left": "issue:38175", + "right": "issue:44617", + "accept": false, + "reason": "Zero probabilities in SigLIP2 and Sam3Video OOM are entirely different failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 28, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:39692", + "issue:41628", + "issue:41950", + "issue:42175", + "issue:42994", + "issue:43065", + "issue:43122", + "issue:43262", + "issue:43295", + "issue:43317", + "issue:43352", + "issue:43441", + "issue:43450", + "issue:43475", + "issue:43525", + "issue:43650", + "issue:43673", + "issue:43723", + "issue:43761", + "issue:43824", + "issue:43881", + "issue:43901", + "issue:43976", + "issue:43994", + "issue:44016", + "issue:44162", + "issue:44220", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44279", + "issue:44291", + "issue:44297", + "issue:44466", + "issue:44521", + "issue:44589", + "issue:44617", + "issue:44623", + "issue:44671", + "issue:44683", + "issue:44749", + "issue:44779", + "issue:44898", + "issue:45250", + "issue:45276", + "issue:45290", + "issue:45362", + "issue:45397" + ], + "soft_pairs": [ + "issue:39692|issue:43901", + "issue:43650|issue:45290", + "issue:44016|issue:44291", + "issue:42994|issue:43317", + "issue:43441|issue:44683", + "issue:43450|issue:44279", + "issue:43065|issue:44898", + "issue:44466|issue:45276", + "issue:43262|issue:44521", + "issue:43723|issue:44162", + "issue:44261|issue:44263", + "issue:44623|issue:45397", + "issue:43122|issue:43650", + "issue:44623|issue:44749", + "issue:44671|issue:44779", + "issue:43673|issue:43761", + "issue:43901|issue:43994", + "issue:42175|issue:43976", + "issue:44297|issue:45397", + "issue:43824|issue:43976", + "issue:43475|issue:44016", + "issue:43295|issue:44617", + "issue:44589|issue:45362", + "issue:43352|issue:45250", + "issue:43901|issue:44246", + "issue:43525|issue:43881", + "issue:41950|issue:44220", + "issue:43450|issue:44617", + "issue:41628|issue:43976" + ], + "prepared_review_unit_hash": "34ed06ea54caeadd4acdee71ba53c0ae3265d1ed862a1f8ea02e37820c125a94", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 35995, + "estimated_input_tokens": 8999, + "estimated_eval_tokens": 18254 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is highly heterogeneous; the soft pairs mostly share only broad transformer-related vocabulary, not the same underlying bug. I found no safe duplicates to merge, and the only weak thematic overlap is around SigLIP2, which is still not enough to accept a pair.", + "confidence": 0.96, + "canonical_issue_reason": "If forced to anchor the cluster, issue 43994 is the closest substantive user-facing bug in the only loosely shared area (SigLIP2), but there is no true canonical duplicate set here.", + "canonical_pr_reason": null, + "best_issue_reason": "43994 is the most concrete and broadly applicable issue among this set; the others are mostly unrelated subsystems or narrow regressions. It is still only a weak representative, not a genuine duplicate hub.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:39692", + "right": "issue:43901", + "accept": false, + "reason": "SigLIP2 docs/example bug vs a text-classification docs mismatch; different feature area and failure mode." + }, + { + "left": "issue:43650", + "right": "issue:45290", + "accept": false, + "reason": "Completely unrelated: one is a vague data request, the other is a chat-template tool-call crash." + }, + { + "left": "issue:44016", + "right": "issue:44291", + "accept": false, + "reason": "Notebook syntax error vs init_empty_weights constructor regression; unrelated issues." + }, + { + "left": "issue:42994", + "right": "issue:43317", + "accept": false, + "reason": "Both involve quantization, but one is saving and the other is dequantized model loading/offload; different bugs." + }, + { + "left": "issue:43441", + "right": "issue:44683", + "accept": false, + "reason": "Both mention Flash Attention, but one is model support gating and the other is compiled flex_attention on newer torch; not the same bug." + }, + { + "left": "issue:43450", + "right": "issue:44279", + "accept": false, + "reason": "Video processor batch-shape bug vs a generic dependency issue; no shared code-path." + }, + { + "left": "issue:43065", + "right": "issue:44898", + "accept": false, + "reason": "Dummy Conv2d in SAM3 pixel decoder vs Perceiver interpolation failure; unrelated model internals." + }, + { + "left": "issue:44466", + "right": "issue:45276", + "accept": false, + "reason": "Serialization of tied weights vs resize_token_embeddings not updating Gemma4 embeddings; different concrete problems." + }, + { + "left": "issue:43262", + "right": "issue:44521", + "accept": false, + "reason": "Audio chat-template sampling-rate default vs multimodal assistant-mask bug; same API surface, different behavior." + }, + { + "left": "issue:43723", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer import/loading issue vs ESM2 regression; unrelated models and symptoms." + }, + { + "left": "issue:44261", + "right": "issue:44263", + "accept": false, + "reason": "MLA layernorm epsilon precision issue vs torch.split return behavior in a different component; not the same bug." + }, + { + "left": "issue:44623", + "right": "issue:45397", + "accept": false, + "reason": "processor.save_pretrained missing files vs Gemma-4 zero3 from_pretrained failure; save/load bugs but distinct." + }, + { + "left": "issue:43122", + "right": "issue:43650", + "accept": false, + "reason": "Tokenizer behavior regression vs unrelated data request." + }, + { + "left": "issue:44623", + "right": "issue:44749", + "accept": false, + "reason": "Processor save bug vs a Chinese report about slow filtering; unrelated." + }, + { + "left": "issue:44671", + "right": "issue:44779", + "accept": false, + "reason": "CamemBERT MLM predictions vs Deepseek tokenization; different models and failure modes." + }, + { + "left": "issue:43673", + "right": "issue:43761", + "accept": false, + "reason": "Generation cache regression vs CLIPVisionModel hidden_states regression; both v5 bugs but different code paths." + }, + { + "left": "issue:43901", + "right": "issue:43994", + "accept": false, + "reason": "Docs wording about return_all_scores vs a SigLIP2 inference/result bug; no same underlying issue." + }, + { + "left": "issue:42175", + "right": "issue:43976", + "accept": false, + "reason": "Package/backend dependency mismatch vs Python version compatibility; unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45397", + "accept": false, + "reason": "Tokenizer class metadata mismatch on save vs Gemma-4 deepspeed zero3 loading failure; different layers." + }, + { + "left": "issue:43824", + "right": "issue:43976", + "accept": false, + "reason": "Missing Qwen2.5-VL import vs Python support issue; unrelated." + }, + { + "left": "issue:43475", + "right": "issue:44016", + "accept": false, + "reason": "SAM3 video encoder output attribute bug vs notebook syntax error; unrelated." + }, + { + "left": "issue:43295", + "right": "issue:44617", + "accept": false, + "reason": "Processor/tokenizer regression vs CUDA OOM in Sam3Video; not the same bug." + }, + { + "left": "issue:44589", + "right": "issue:45362", + "accept": false, + "reason": "Missing Float8 storage type vs Qwen3.5 chat crash; different failures." + }, + { + "left": "issue:43352", + "right": "issue:45250", + "accept": false, + "reason": "Specific model lacking FlashAttention2 support vs a generic Flash Attention 2.0 issue title; too vague to merge." + }, + { + "left": "issue:43901", + "right": "issue:44246", + "accept": false, + "reason": "Pipeline docs mismatch vs import sometimes slow; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:43881", + "accept": false, + "reason": "Llama4Config pad_token_id attribute error vs glm-4v-9b loading failure; different models and root causes." + }, + { + "left": "issue:41950", + "right": "issue:44220", + "accept": false, + "reason": "Video-classification pipeline image-processor lookup vs audio feature extraction bug; unrelated." + }, + { + "left": "issue:43450", + "right": "issue:44617", + "accept": false, + "reason": "Incorrect batched video processor shape vs SAM3Video OOM; distinct symptoms and fixes." + }, + { + "left": "issue:41628", + "right": "issue:43976", + "accept": false, + "reason": "AutoImageProcessor import failure vs Python version compatibility; not the same problem." + } + ] + }, + "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": 13, + "review_unit_index": 29, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:36246", + "issue:41950", + "issue:42491", + "issue:42757", + "issue:42886", + "issue:43065", + "issue:43066", + "issue:43257", + "issue:43299", + "issue:43408", + "issue:43441", + "issue:43526", + "issue:43531", + "issue:43540", + "issue:43596", + "issue:43673", + "issue:43901", + "issue:43994", + "issue:44162", + "issue:44188", + "issue:44242", + "issue:44246", + "issue:44295", + "issue:44336", + "issue:44351", + "issue:44373", + "issue:44462", + "issue:44485", + "issue:44496", + "issue:44561", + "issue:44749", + "issue:44779", + "issue:44857", + "issue:44871", + "issue:44877", + "issue:44987", + "issue:44995", + "issue:45003", + "issue:45137", + "issue:45200", + "issue:45229", + "issue:45276", + "issue:45292", + "issue:45356", + "issue:45362", + "issue:45375", + "issue:45397", + "issue:45459" + ], + "soft_pairs": [ + "issue:44188|issue:44336", + "issue:44295|issue:45397", + "issue:42491|issue:43299", + "issue:44246|issue:44496", + "issue:44336|issue:44373", + "issue:44462|issue:45356", + "issue:43526|issue:45003", + "issue:43994|issue:45276", + "issue:43531|issue:45362", + "issue:36246|issue:42757", + "issue:36246|issue:44351", + "issue:43673|issue:44561", + "issue:44336|issue:44485", + "issue:36246|issue:43596", + "issue:44242|issue:44857", + "issue:44162|issue:45200", + "issue:44246|issue:44871", + "issue:44779|issue:45292", + "issue:44987|issue:45459", + "issue:44295|issue:44749", + "issue:43441|issue:43540", + "issue:44877|issue:45375", + "issue:41950|issue:43596", + "issue:43066|issue:44779", + "issue:42886|issue:45229", + "issue:43901|issue:45003", + "issue:44995|issue:45200", + "issue:43065|issue:43408", + "issue:43257|issue:45137" + ], + "prepared_review_unit_hash": "bd82b06ebbeef5758a0a57b9cfbe00f0836e80d9a5ebf0c061cb604396c884fc", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 29, + "serialized_chars": 36191, + "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 heterogeneous: the suggested pairs span unrelated bugs, regressions, docs, config, tokenizer, and model-loading issues. I found no soft edge that safely collapses to the same underlying change, and there are no PRs in the set.", + "confidence": 0.97, + "canonical_issue_reason": "No single canonical issue fits this cluster well; the items are not one duplicate family but many distinct problems across different subsystems and models.", + "canonical_pr_reason": null, + "best_issue_reason": "No issue is a strong global representative for deduping here, because the cluster lacks a shared concrete code path or fix target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44188", + "right": "issue:44336", + "accept": false, + "reason": "Torch.compile attention-kernel branching and ANSI output/reporting are unrelated code paths." + }, + { + "left": "issue:44295", + "right": "issue:45397", + "accept": false, + "reason": "Reading registered position_ids and Gemma-4 ZeRO3 from_pretrained are different failures." + }, + { + "left": "issue:42491", + "right": "issue:43299", + "accept": false, + "reason": "Both involve Qwen3-MoE, but one is LoRA portability and the other is model loading; not the same concrete bug." + }, + { + "left": "issue:44246", + "right": "issue:44496", + "accept": false, + "reason": "Import latency and an unrecognized model/config error are unrelated issues." + }, + { + "left": "issue:44336", + "right": "issue:44373", + "accept": false, + "reason": "Terminal ANSI output generation and a wrong docstring for position_ids do not share a bug." + }, + { + "left": "issue:44462", + "right": "issue:45356", + "accept": false, + "reason": "AutoTokenizer ignoring tokenizer.json and Kimi-K2.5 codec/warning regressions are different tokenizer bugs." + }, + { + "left": "issue:43526", + "right": "issue:45003", + "accept": false, + "reason": "BeitImageProcessorFast label reduction and unsafe sys.modules access are unrelated." + }, + { + "left": "issue:43994", + "right": "issue:45276", + "accept": false, + "reason": "Siglip2 model/pipeline output issues and resize_token_embeddings propagation are different problems." + }, + { + "left": "issue:43531", + "right": "issue:45362", + "accept": false, + "reason": "Both are Qwen3-family bugs, but sliding_window handling and chat crashes are distinct code paths." + }, + { + "left": "issue:36246", + "right": "issue:42757", + "accept": false, + "reason": "Different missing imports from different packages; same symptom class, not the same bug." + }, + { + "left": "issue:36246", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they concern different symbols and likely different fixes." + }, + { + "left": "issue:43673", + "right": "issue:44561", + "accept": false, + "reason": "Generation cache chunked_prefill and removal of is_torch_fx_available are unrelated regressions." + }, + { + "left": "issue:44336", + "right": "issue:44485", + "accept": false, + "reason": "ANSI/reporting behavior and GLM-5 RoPE implementation discussion are not the same issue." + }, + { + "left": "issue:36246", + "right": "issue:43596", + "accept": false, + "reason": "Qwen2.5-VL import failure and DeepSpeed ZeRO3/BertModel index error are unrelated." + }, + { + "left": "issue:44242", + "right": "issue:44857", + "accept": false, + "reason": "Missing MoE load-balancing loss and an AMP/CUDA crash are different bugs." + }, + { + "left": "issue:44162", + "right": "issue:45200", + "accept": false, + "reason": "ESM2 breakage and Gemma 4 mm_token_type_ids defaults are unrelated model-specific issues." + }, + { + "left": "issue:44246", + "right": "issue:44871", + "accept": false, + "reason": "Slow import timing and eos_token_id config mismatch are unrelated." + }, + { + "left": "issue:44779", + "right": "issue:45292", + "accept": false, + "reason": "Tokenizer regression and resize_token_embeddings not updating outputs are different code paths." + }, + { + "left": "issue:44987", + "right": "issue:45459", + "accept": false, + "reason": "Model loading failure and protobuf error masking are unrelated tokenizer/loading issues." + }, + { + "left": "issue:44295", + "right": "issue:44749", + "accept": false, + "reason": "Position_ids buffer access and a general slowdown regression are not the same bug." + }, + { + "left": "issue:43441", + "right": "issue:43540", + "accept": false, + "reason": "Ministral FlashAttention failure and Qwen3OmniMoe video input errors are unrelated." + }, + { + "left": "issue:44877", + "right": "issue:45375", + "accept": false, + "reason": "Both are config-validation problems, but for different models and different missing fields." + }, + { + "left": "issue:41950", + "right": "issue:43596", + "accept": false, + "reason": "Video-classification pipeline processor selection and ZeRO3/BertModel index errors are unrelated." + }, + { + "left": "issue:43066", + "right": "issue:44779", + "accept": false, + "reason": "Tokenizer decoder type regression and DeepSeek tokenizer regression are not the same concrete issue." + }, + { + "left": "issue:42886", + "right": "issue:45229", + "accept": false, + "reason": "Offline cache loading and Gemma4 multi-GPU OOM are unrelated." + }, + { + "left": "issue:43901", + "right": "issue:45003", + "accept": false, + "reason": "Docs mismatch and unsafe sys.modules access are not the same bug." + }, + { + "left": "issue:44995", + "right": "issue:45200", + "accept": false, + "reason": "Stale indexer cache on second forward and Gemma 4 text-only defaults are unrelated." + }, + { + "left": "issue:43065", + "right": "issue:43408", + "accept": false, + "reason": "SAM3 dummy Conv2d implementation and sam3_video/sam3_tracker config warning are related only by family, not the same fix." + }, + { + "left": "issue:43257", + "right": "issue:45137", + "accept": false, + "reason": "Qwen3-MoE weight conversion with accelerate/deepspeed and ZeRO3 deque underflow are different bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 30, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:28282", + "issue:30990", + "issue:42673", + "issue:43232", + "issue:43262", + "issue:43295", + "issue:43381", + "issue:43441", + "issue:43475", + "issue:43540", + "issue:43606", + "issue:43643", + "issue:43646", + "issue:43653", + "issue:43673", + "issue:43701", + "issue:43704", + "issue:43931", + "issue:44038", + "issue:44060", + "issue:44077", + "issue:44162", + "issue:44246", + "issue:44297", + "issue:44336", + "issue:44361", + "issue:44368", + "issue:44466", + "issue:44496", + "issue:44623", + "issue:44734", + "issue:44749", + "issue:44869", + "issue:44933", + "issue:44977", + "issue:45003", + "issue:45081", + "issue:45103", + "issue:45230", + "issue:45278", + "issue:45292", + "issue:45310", + "issue:45335", + "issue:45356", + "issue:45381", + "issue:45397", + "issue:45459" + ], + "soft_pairs": [ + "issue:43232|issue:44336", + "issue:44496|issue:45003", + "issue:43262|issue:44246", + "issue:43295|issue:45292", + "issue:44060|issue:44077", + "issue:43475|issue:43540", + "issue:44466|issue:45335", + "issue:43653|issue:44162", + "issue:44368|issue:44977", + "issue:43673|issue:44734", + "issue:44933|issue:45459", + "issue:43441|issue:43606", + "issue:44869|issue:45356", + "issue:43232|issue:43673", + "issue:43931|issue:44368", + "issue:43673|issue:45278", + "issue:44297|issue:44623", + "issue:43646|issue:44162", + "issue:43381|issue:45292", + "issue:43381|issue:45335", + "issue:42673|issue:43704", + "issue:28282|issue:30990", + "issue:44361|issue:44749", + "issue:44077|issue:45003", + "issue:44749|issue:45310", + "issue:45230|issue:45397", + "issue:44038|issue:44623", + "issue:43475|issue:45381", + "issue:43475|issue:43643", + "issue:43701|issue:44749", + "issue:45081|issue:45103" + ], + "prepared_review_unit_hash": "77fa6eb05ff4c306902569c5ce267ccea6439d1799aef53bd0cbc01368ae5df4", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 31, + "serialized_chars": 36501, + "estimated_input_tokens": 9126, + "estimated_eval_tokens": 18508 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most candidate pairs are unrelated or only loosely similar. The only clear duplicate is the Qwen3ForCausalLM VRAM leak report; everything else appears to be separate bugs or regressions in different code paths/models.", + "confidence": 0.93, + "canonical_issue_reason": "issue:42673 is the clearer, higher-signal report for the Qwen3ForCausalLM VRAM leak, and issue:43704 appears to be the same bug reported again.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:42673 best represents the only true duplicate pair in this set and is the strongest canonical issue choice.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43232", + "right": "issue:44336", + "accept": false, + "reason": "Different subsystems: generation kwargs vs terminal ANSI output formatting." + }, + { + "left": "issue:44496", + "right": "issue:45003", + "accept": false, + "reason": "Unrecognized model/config loading error is unrelated to unsafe sys.modules access." + }, + { + "left": "issue:43262", + "right": "issue:44246", + "accept": false, + "reason": "Audio sampling-rate defaulting and slow import are unrelated bugs." + }, + { + "left": "issue:43295", + "right": "issue:45292", + "accept": false, + "reason": "Custom processor/tokenizer regression is unrelated to embedding resize behavior." + }, + { + "left": "issue:44060", + "right": "issue:44077", + "accept": false, + "reason": "Weight-tying warning bug and patchtsmixer post_init restriction are different issues." + }, + { + "left": "issue:43475", + "right": "issue:43540", + "accept": false, + "reason": "Both involve multimodal/video models, but they are distinct failures in different model paths." + }, + { + "left": "issue:44466", + "right": "issue:45335", + "accept": false, + "reason": "Both touch tied embeddings, but one is serialization/device-dependent and the other is resize_token_embeddings propagation." + }, + { + "left": "issue:43653", + "right": "issue:44162", + "accept": false, + "reason": "Tokenizer special-token registration and ESM2 breakage are unrelated." + }, + { + "left": "issue:44368", + "right": "issue:44977", + "accept": false, + "reason": "A tied-weights warning and a flash-attention generation bug are not the same concrete defect." + }, + { + "left": "issue:43673", + "right": "issue:44734", + "accept": false, + "reason": "Both concern generation/cache, but the reported failures and code paths differ." + }, + { + "left": "issue:44933", + "right": "issue:45459", + "accept": false, + "reason": "Missing image_utils import and protobuf error masking are different import/runtime problems." + }, + { + "left": "issue:43441", + "right": "issue:43606", + "accept": false, + "reason": "FlashAttention failure for Ministral-3 is unrelated to CPU offload device mismatch for bark-small." + }, + { + "left": "issue:44869", + "right": "issue:45356", + "accept": false, + "reason": "Whisper timestamp decoding crash and Kimi tokenizer codec regression are different tokenizer bugs." + }, + { + "left": "issue:43232", + "right": "issue:43673", + "accept": false, + "reason": "These are different generation-related regressions with distinct triggers and symptoms." + }, + { + "left": "issue:43931", + "right": "issue:44368", + "accept": false, + "reason": "Weight shape mismatch in Qwen3-VL is unrelated to a tied-embeddings warning in Qwen3.5." + }, + { + "left": "issue:43673", + "right": "issue:45278", + "accept": false, + "reason": "A specific chunked_prefill cache bug is not the same as broad import errors after upgrade." + }, + { + "left": "issue:44297", + "right": "issue:44623", + "accept": false, + "reason": "Tokenizer save_pretrained metadata mismatch and processor save_pretrained missing files are separate serialization bugs." + }, + { + "left": "issue:43646", + "right": "issue:44162", + "accept": false, + "reason": "Custom model initialization breakage is not the same as the ESM2-specific failure." + }, + { + "left": "issue:43381", + "right": "issue:45292", + "accept": false, + "reason": "Gradient checkpointing eval-mode behavior is unrelated to embedding resize propagation." + }, + { + "left": "issue:43381", + "right": "issue:45335", + "accept": false, + "reason": "Different features and failure modes: checkpointing vs token embedding resize." + }, + { + "left": "issue:42673", + "right": "issue:43704", + "accept": true, + "reason": "Exact same title and same Qwen3ForCausalLM VRAM leak in multiple dataloader threads." + }, + { + "left": "issue:28282", + "right": "issue:30990", + "accept": false, + "reason": "PyTorch missing ImportError and Sentence Transformers hanging are not the same bug." + }, + { + "left": "issue:44361", + "right": "issue:44749", + "accept": false, + "reason": "Tokenizer AttributeError on tasks and slowdown after upgrade are unrelated." + }, + { + "left": "issue:44077", + "right": "issue:45003", + "accept": false, + "reason": "Patchtsmixer post_init policy and unsafe sys.modules access are different code issues." + }, + { + "left": "issue:44749", + "right": "issue:45310", + "accept": false, + "reason": "Performance regression during filtering is unrelated to Qwen3.5 MoE from_pretrained errors." + }, + { + "left": "issue:45230", + "right": "issue:45397", + "accept": false, + "reason": "Generic bug report is too vague and does not match the specific gemma-4 zero3 failure." + }, + { + "left": "issue:44038", + "right": "issue:44623", + "accept": false, + "reason": "Qwen3-VL-Moe loading error and processor save_pretrained missing files are separate issues." + }, + { + "left": "issue:43475", + "right": "issue:45381", + "accept": false, + "reason": "Both are vision/video-related, but they affect different models and different tensor fields." + }, + { + "left": "issue:43475", + "right": "issue:43643", + "accept": false, + "reason": "A missing vision encoder attribute is unrelated to missing fields from AutoConfig with trust_remote_code." + }, + { + "left": "issue:43701", + "right": "issue:44749", + "accept": false, + "reason": "Checkpoint key mismatch and filtering slowdown are unrelated." + }, + { + "left": "issue:45081", + "right": "issue:45103", + "accept": false, + "reason": "Both are crashes in helper code, but they come from different utilities and different causes." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet and stays conservative. The duplicate call on issue:42673 / issue:43704 is well supported by the exact same title and matching symptom, and the other soft-edge rejections are reasonable based on clearly different titles/code paths. I did not see evidence of overstatement." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 31, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:30990", + "issue:35141", + "issue:36010", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42831", + "issue:42898", + "issue:42994", + "issue:43232", + "issue:43278", + "issue:43377", + "issue:43475", + "issue:43479", + "issue:43525", + "issue:43673", + "issue:43784", + "issue:43824", + "issue:43873", + "issue:43976", + "issue:44038", + "issue:44079", + "issue:44164", + "issue:44188", + "issue:44222", + "issue:44242", + "issue:44351", + "issue:44373", + "issue:44568", + "issue:44623", + "issue:44661", + "issue:44734", + "issue:44792", + "issue:44869", + "issue:44933", + "issue:44945", + "issue:44964", + "issue:45092", + "issue:45103", + "issue:45137", + "issue:45200", + "issue:45310", + "issue:45362", + "issue:45375", + "issue:45397", + "issue:45478" + ], + "soft_pairs": [ + "issue:42371|issue:45092", + "issue:45375|issue:45478", + "issue:41628|issue:44933", + "issue:42994|issue:43525", + "issue:45310|issue:45375", + "issue:43475|issue:44661", + "issue:42222|issue:45200", + "issue:43873|issue:44945", + "issue:42831|issue:44222", + "issue:45362|issue:45375", + "issue:45310|issue:45397", + "issue:43232|issue:44188", + "issue:41628|issue:44351", + "issue:44222|issue:45310", + "issue:42994|issue:44164", + "issue:43232|issue:44792", + "issue:44038|issue:45310", + "issue:30990|issue:35141", + "issue:36010|issue:41628", + "issue:43377|issue:44242", + "issue:43479|issue:44079", + "issue:43673|issue:44188", + "issue:43479|issue:44869", + "issue:42898|issue:44568", + "issue:44964|issue:45137", + "issue:42175|issue:44933", + "issue:43976|issue:44933", + "issue:43784|issue:43824", + "issue:43278|issue:45137", + "issue:44623|issue:45310", + "issue:44734|issue:45103", + "issue:44188|issue:44373" + ], + "prepared_review_unit_hash": "9a226759d146e854f6b650cf2ab9522a2b75b28035c4f442a7554b2b1540a255", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 32, + "serialized_chars": 36335, + "estimated_input_tokens": 9084, + "estimated_eval_tokens": 18424 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is not a true duplicate set; it contains many unrelated Transformers issues spanning imports, config/model loading, generation, quantization, and serving. None of the soft edges are strong enough to merge under a strict duplicate policy.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42371", + "right": "issue:45092", + "accept": false, + "reason": "Different problems: TF32 settings warning vs. old InternVL2 remote-code/meta-init incompatibility." + }, + { + "left": "issue:45375", + "right": "issue:45478", + "accept": false, + "reason": "Same model family, but one is a missing config field under @strict and the other is a generic from_pretrained failure; not clearly the same bug." + }, + { + "left": "issue:41628", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but they target different symbols/modules and likely different code paths." + }, + { + "left": "issue:42994", + "right": "issue:43525", + "accept": false, + "reason": "Quantized saving failure vs. Llama4Config missing pad_token_id are unrelated." + }, + { + "left": "issue:45310", + "right": "issue:45375", + "accept": false, + "reason": "Both mention Qwen3.5 MoE loading, but one is a specific config field issue and the other is a broader from_pretrained error without enough evidence they are the same defect." + }, + { + "left": "issue:43475", + "right": "issue:44661", + "accept": false, + "reason": "SAM 3 vision output attribute error vs. tokenizer mapping failure; unrelated subsystems." + }, + { + "left": "issue:42222", + "right": "issue:45200", + "accept": false, + "reason": "Different models and different failures: VitPose breakage vs. Gemma 4 mm_token_type_ids defaulting." + }, + { + "left": "issue:43873", + "right": "issue:44945", + "accept": false, + "reason": "Quantization offloading vs. pipeline-parallel output corruption are different issues." + }, + { + "left": "issue:42831", + "right": "issue:44222", + "accept": false, + "reason": "Both involve FP8, but one is an accuracy regression and the other is save_pretrained behavior for MoE; not the same concrete bug." + }, + { + "left": "issue:45362", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen3.5-related, but one is a chat crash and the other is a config-field omission; insufficient overlap." + }, + { + "left": "issue:45310", + "right": "issue:45397", + "accept": false, + "reason": "Same broad loading area, but different model families (Qwen3.5 MoE vs Gemma-4 zero3)." + }, + { + "left": "issue:43232", + "right": "issue:44188", + "accept": false, + "reason": "Both are generation-related, but the reported failures involve different mechanisms and code paths." + }, + { + "left": "issue:41628", + "right": "issue:44351", + "accept": false, + "reason": "Different missing imports; no evidence they stem from the same API break." + }, + { + "left": "issue:44222", + "right": "issue:45310", + "accept": false, + "reason": "FP8 save_pretrained bug vs Qwen3.5 MoE from_pretrained error are distinct." + }, + { + "left": "issue:42994", + "right": "issue:44164", + "accept": false, + "reason": "Both concern save/from_pretrained, but one is quantized model saving and the other is extra_state handling; not the same bug." + }, + { + "left": "issue:43232", + "right": "issue:44792", + "accept": false, + "reason": "Generation kwargs sync issue vs. Janus image-generation test failure are unrelated." + }, + { + "left": "issue:44038", + "right": "issue:45310", + "accept": false, + "reason": "Both mention Qwen variants, but the titles indicate different model releases and different failure modes." + }, + { + "left": "issue:30990", + "right": "issue:35141", + "accept": false, + "reason": "SentenceTransformers loading stall vs embedding reinit after resize are unrelated." + }, + { + "left": "issue:36010", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but for different removed/moved names and different compatibility issues." + }, + { + "left": "issue:43377", + "right": "issue:44242", + "accept": false, + "reason": "MIMI batching/padding-mask bug vs MoE load-balancing loss omission are unrelated." + }, + { + "left": "issue:43479", + "right": "issue:44079", + "accept": false, + "reason": "Config default initialization bug vs ModelOutput key assignment bug are different classes of failure." + }, + { + "left": "issue:43673", + "right": "issue:44188", + "accept": false, + "reason": "Chunked prefill cache issue vs torch.compile attention-kernel divergence are different code paths." + }, + { + "left": "issue:43479", + "right": "issue:44869", + "accept": false, + "reason": "Phi4 multimodal config initialization vs Whisper timestamp decoding crash are unrelated." + }, + { + "left": "issue:42898", + "right": "issue:44568", + "accept": false, + "reason": "Both involve tokenizer behavior changes, but one is clean_up_tokenization_spaces and the other is add_special_tokens/BOS-EOS insertion; not the same bug." + }, + { + "left": "issue:44964", + "right": "issue:45137", + "accept": false, + "reason": "Phi-4 multimodal loading failure vs DeepSpeed ZeRO3 deque error are unrelated." + }, + { + "left": "issue:42175", + "right": "issue:44933", + "accept": false, + "reason": "Package extra/backend dependency issue vs missing image_utils import are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44933", + "accept": false, + "reason": "Python version compatibility bug vs missing import are different." + }, + { + "left": "issue:43784", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but they reference different missing symbols and likely different upstream regressions." + }, + { + "left": "issue:43278", + "right": "issue:45137", + "accept": false, + "reason": "Embedding dtype drift in eval vs ZeRO3 deque underflow are unrelated." + }, + { + "left": "issue:44623", + "right": "issue:45310", + "accept": false, + "reason": "Missing files in processor.save_pretrained vs Qwen3.5 MoE loading error are not the same issue." + }, + { + "left": "issue:44734", + "right": "issue:45103", + "accept": false, + "reason": "Serving KV-cache tensor indexing crash vs auto_docstring annotation AttributeError are unrelated." + }, + { + "left": "issue:44188", + "right": "issue:44373", + "accept": false, + "reason": "Runtime attention-kernel divergence vs docstring wording issue are clearly 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": 13, + "review_unit_index": 32, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:22355", + "issue:28282", + "issue:30990", + "issue:38175", + "issue:41628", + "issue:42757", + "issue:42831", + "issue:42898", + "issue:43295", + "issue:43479", + "issue:43644", + "issue:43723", + "issue:43874", + "issue:44060", + "issue:44077", + "issue:44206", + "issue:44222", + "issue:44242", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44442", + "issue:44448", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44514", + "issue:44568", + "issue:44623", + "issue:44779", + "issue:44871", + "issue:44913", + "issue:44977", + "issue:44991", + "issue:45092", + "issue:45103", + "issue:45125", + "issue:45137", + "issue:45276", + "issue:45310", + "issue:45356", + "issue:45357", + "issue:45406", + "issue:45459" + ], + "soft_pairs": [ + "issue:44060|issue:45125", + "issue:42898|issue:45356", + "issue:43723|issue:44568", + "issue:42898|issue:44779", + "issue:45137|issue:45276", + "issue:43874|issue:45092", + "issue:44077|issue:44242", + "issue:42898|issue:44448", + "issue:41628|issue:45459", + "issue:44222|issue:44297", + "issue:44442|issue:45459", + "issue:42898|issue:43295", + "issue:43479|issue:43644", + "issue:44222|issue:44623", + "issue:44206|issue:45103", + "issue:42831|issue:44623", + "issue:22355|issue:28282", + "issue:41628|issue:42757", + "issue:44483|issue:45406", + "issue:44297|issue:45310", + "issue:44222|issue:44484", + "issue:44263|issue:44360", + "issue:43723|issue:44977", + "issue:44263|issue:45092", + "issue:44514|issue:45103", + "issue:43479|issue:44913", + "issue:44295|issue:44623", + "issue:43723|issue:44871", + "issue:44261|issue:45092", + "issue:43723|issue:44991", + "issue:30990|issue:38175", + "issue:43723|issue:44442", + "issue:44373|issue:44485", + "issue:45092|issue:45357" + ], + "prepared_review_unit_hash": "bb42c38a74e3e90b84e736470e49e5c113d2a3669758073a731f3edb7f914a73", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 34, + "serialized_chars": 37224, + "estimated_input_tokens": 9306, + "estimated_eval_tokens": 18868 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous. I found one likely duplicate pocket around tokenizer loading regressions in v5; the rest are different bugs touching unrelated code paths.", + "confidence": 0.58, + "canonical_issue_reason": "issue:43723 is the broadest tokenizer-loading regression here and best serves as the umbrella for the one plausible duplicate-style report (44991).", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43723 is the strongest global issue candidate because its title is generic enough to represent the shared tokenizer-loading failure pattern without overfitting to a single model.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44060", + "right": "issue:45125", + "accept": false, + "reason": "Different Qwen3/Qwen3.5 MoE problems: tied-weights warning vs missing _tp_plan for tensor parallelism." + }, + { + "left": "issue:42898", + "right": "issue:45356", + "accept": false, + "reason": "Both are tokenizer regressions, but the symptoms and likely code paths differ: cleanup-space behavior vs codec handling/warning." + }, + { + "left": "issue:43723", + "right": "issue:44568", + "accept": false, + "reason": "Tokenizer-loading failure vs BOS/EOS insertion bug; different behavior and failure mode." + }, + { + "left": "issue:42898", + "right": "issue:44779", + "accept": false, + "reason": "Both involve tokenizer behavior in v5, but one is a cleanup-space regression and the other is incorrect DeepSeek outputs." + }, + { + "left": "issue:45137", + "right": "issue:45276", + "accept": false, + "reason": "Unrelated bugs: DeepSpeed ZeRO3 deque error vs Gemma4 resize_token_embeddings not updating embeddings." + }, + { + "left": "issue:43874", + "right": "issue:45092", + "accept": false, + "reason": "Different multimodal failures: missing image-patch method vs old InternVL2 remote-code/meta-init incompatibility." + }, + { + "left": "issue:44077", + "right": "issue:44242", + "accept": false, + "reason": "Distinct model logic bugs in different places; no clear shared concrete code-path." + }, + { + "left": "issue:42898", + "right": "issue:44448", + "accept": false, + "reason": "Tokenizer-related regression vs Pegasus output mismatch; too broad and not the same bug." + }, + { + "left": "issue:41628", + "right": "issue:45459", + "accept": false, + "reason": "Import error for AutoImageProcessor vs tokenizer exception masking when protobuf is absent; unrelated failures." + }, + { + "left": "issue:44222", + "right": "issue:44297", + "accept": false, + "reason": "Save-pretrained FP8/MoE issue vs tokenizer_class metadata mismatch; different components." + }, + { + "left": "issue:44442", + "right": "issue:45459", + "accept": false, + "reason": "AutoTokenizer failing to load a specific tokenizer vs protobuf error handling swallowing tokenizer errors; different root causes." + }, + { + "left": "issue:42898", + "right": "issue:43295", + "accept": false, + "reason": "Different regressions: clean_up_tokenization_spaces vs processor.tokenizer / image-to-tokenizer regression." + }, + { + "left": "issue:43479", + "right": "issue:43644", + "accept": false, + "reason": "Different config initialization bug vs non-persistent buffer corruption; not the same underlying change." + }, + { + "left": "issue:44222", + "right": "issue:44623", + "accept": false, + "reason": "Both mention save_pretrained, but one is FP8 MoE weights and the other is missing processor files." + }, + { + "left": "issue:44206", + "right": "issue:45103", + "accept": false, + "reason": "LasrFeatureExtractor center-arg crash is unrelated to auto_docstring failures under future annotations." + }, + { + "left": "issue:42831", + "right": "issue:44623", + "accept": false, + "reason": "Accuracy issue with FineGrainedFP8 vs processor.save_pretrained missing files; different bug reports." + }, + { + "left": "issue:22355", + "right": "issue:28282", + "accept": false, + "reason": "Different import failures: missing transformers.onnx module vs AutoModel requiring PyTorch." + }, + { + "left": "issue:41628", + "right": "issue:42757", + "accept": false, + "reason": "Different missing imports from different packages and different API surfaces." + }, + { + "left": "issue:44483", + "right": "issue:45406", + "accept": false, + "reason": "Chat-completions API request rejection vs Gemma4Processor missing _tokenizer; unrelated." + }, + { + "left": "issue:44297", + "right": "issue:45310", + "accept": false, + "reason": "Tokenizer config save mismatch vs Qwen3.5 MoE from_pretrained error; not the same concrete bug." + }, + { + "left": "issue:44222", + "right": "issue:44484", + "accept": false, + "reason": "FP8 save_pretrained bug vs max_shard_size behavior question; no duplicate evidence." + }, + { + "left": "issue:44263", + "right": "issue:44360", + "accept": false, + "reason": "Different DSA indexer issues: torch.split return values vs missing ReLU." + }, + { + "left": "issue:43723", + "right": "issue:44977", + "accept": false, + "reason": "Both are tokenizer-ish regressions, but one is loading failure and the other is generation correctness with flash-attention." + }, + { + "left": "issue:44263", + "right": "issue:45092", + "accept": false, + "reason": "Different bugs: torch.split return handling vs remote-code/meta-initialization incompatibility." + }, + { + "left": "issue:44514", + "right": "issue:45103", + "accept": false, + "reason": "apply_chat_template batching crash is unrelated to auto_docstring annotation handling." + }, + { + "left": "issue:43479", + "right": "issue:44913", + "accept": false, + "reason": "Config defaulting for multimodal vision/audio settings vs GPTNeoX rotary_pct reload regression." + }, + { + "left": "issue:44295", + "right": "issue:44623", + "accept": false, + "reason": "Buffer read error for position_ids vs processor.save_pretrained missing files; different code paths." + }, + { + "left": "issue:43723", + "right": "issue:44871", + "accept": false, + "reason": "Tokenizer load failure vs eos_token_id inconsistency between tokenizer and model config." + }, + { + "left": "issue:44261", + "right": "issue:45092", + "accept": false, + "reason": "MLA q_a_layernorm epsilon mismatch vs InternVL2 meta-init incompatibility; unrelated." + }, + { + "left": "issue:43723", + "right": "issue:44991", + "accept": true, + "reason": "Both describe v5 AutoTokenizer.from_pretrained tokenizer-loading failures; 44991 is a specific model reproduction of the same underlying issue." + }, + { + "left": "issue:30990", + "right": "issue:38175", + "accept": false, + "reason": "Sentence Transformers loading hang vs SigLIP2 zero probabilities; different symptoms and likely causes." + }, + { + "left": "issue:43723", + "right": "issue:44442", + "accept": false, + "reason": "Both are tokenizer problems, but one is loading failure and the other is AutoTokenizer failing for a specific tokenizer class." + }, + { + "left": "issue:44373", + "right": "issue:44485", + "accept": false, + "reason": "Docstring issue for position_ids vs GLM-5 RoPE implementation bug; not duplicates." + }, + { + "left": "issue:45092", + "right": "issue:45357", + "accept": false, + "reason": "InternVL2 meta-init incompatibility vs Qwen3.5 save_pretrained writing wrong visual encoder keys; different bugs." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary stays grounded in the packet: it treats the cluster as mostly heterogeneous, keeps the tokenizer-loading grouping narrow, and only accepts the 43723/44991 pair where the titles clearly align. The false soft-edge calls are conservative and do not overclaim duplicate status from weak similarity alone." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 33, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:29127", + "issue:30990", + "issue:33357", + "issue:36296", + "issue:39401", + "issue:41628", + "issue:42371", + "issue:42757", + "issue:42831", + "issue:42886", + "issue:42907", + "issue:42994", + "issue:43066", + "issue:43116", + "issue:43122", + "issue:43232", + "issue:43408", + "issue:43452", + "issue:43475", + "issue:43576", + "issue:43784", + "issue:43824", + "issue:43937", + "issue:43976", + "issue:43992", + "issue:44038", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44295", + "issue:44297", + "issue:44360", + "issue:44373", + "issue:44485", + "issue:44488", + "issue:44514", + "issue:44521", + "issue:44625", + "issue:44704", + "issue:44743", + "issue:44779", + "issue:44829", + "issue:44945", + "issue:45200", + "issue:45245", + "issue:45278", + "issue:45290", + "issue:45310" + ], + "soft_pairs": [ + "issue:42886|issue:42994", + "issue:41628|issue:43824", + "issue:42831|issue:45310", + "issue:43576|issue:43784", + "issue:42831|issue:44038", + "issue:43122|issue:44521", + "issue:42757|issue:45278", + "issue:44263|issue:44373", + "issue:42886|issue:43066", + "issue:44117|issue:44704", + "issue:30990|issue:33357", + "issue:43122|issue:44514", + "issue:42907|issue:44945", + "issue:44704|issue:45290", + "issue:43232|issue:45245", + "issue:42371|issue:44263", + "issue:43976|issue:44945", + "issue:43232|issue:44743", + "issue:44295|issue:44297", + "issue:39401|issue:44488", + "issue:42907|issue:43937", + "issue:43408|issue:43475", + "issue:44625|issue:44829", + "issue:29127|issue:30990", + "issue:41628|issue:45278", + "issue:44360|issue:44485", + "issue:43116|issue:45200", + "issue:43992|issue:44704", + "issue:39401|issue:44521", + "issue:43122|issue:44779", + "issue:44038|issue:44297", + "issue:44261|issue:44373", + "issue:42371|issue:42831", + "issue:44038|issue:44295", + "issue:36296|issue:43116", + "issue:42907|issue:43452" + ], + "prepared_review_unit_hash": "cfd69270f48ae30b2f6fe9e4a4c61306e652810cdb202081abebb4c3ff3d3dff", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 37992, + "estimated_input_tokens": 9498, + "estimated_eval_tokens": 19252 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The set is mostly a grab bag of unrelated issues; a few share broad themes like imports, tokenizers, or generation, but none of the soft edges look like the same concrete bug/change. I would keep them separate.", + "confidence": 0.37, + "canonical_issue_reason": "No single issue clearly subsumes the others. If one representative is needed, issue:45278 is the broadest umbrella-style report about post-upgrade import failures, but it is not a true duplicate hub.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45278 is the most umbrella-like and still-open report, so it is the best single issue to represent the cluster if one must be chosen.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42886", + "right": "issue:42994", + "accept": false, + "reason": "Different failures: offline cache loading vs quantized model saving." + }, + { + "left": "issue:41628", + "right": "issue:43824", + "accept": false, + "reason": "Both are import errors, but for different symbols and different model families." + }, + { + "left": "issue:42831", + "right": "issue:45310", + "accept": false, + "reason": "Accuracy regression in FP8 vs Qwen3.5-MoE from_pretrained error; not the same bug." + }, + { + "left": "issue:43576", + "right": "issue:43784", + "accept": false, + "reason": "Broken env command vs sentence-transformers import NameError; unrelated code paths." + }, + { + "left": "issue:42831", + "right": "issue:44038", + "accept": false, + "reason": "Different model-specific issues: FP8 accuracy vs Qwen3-VL-MoE bug." + }, + { + "left": "issue:43122", + "right": "issue:44521", + "accept": false, + "reason": "Tokenizer regression vs multimodal assistant mask bug; only a broad tokenizer theme overlaps." + }, + { + "left": "issue:42757", + "right": "issue:45278", + "accept": false, + "reason": "Specific missing symbol vs broad post-upgrade import failures; not the same concrete defect." + }, + { + "left": "issue:44263", + "right": "issue:44373", + "accept": false, + "reason": "GLM-MoE indexer behavior vs a docstring issue; different type of problem." + }, + { + "left": "issue:42886", + "right": "issue:43066", + "accept": false, + "reason": "Offline cache load failure vs wrong tokenizer decoder type in v5." + }, + { + "left": "issue:44117", + "right": "issue:44704", + "accept": false, + "reason": "Mapping-name fallback bug vs AutoProcessor kwarg propagation; unrelated paths." + }, + { + "left": "issue:30990", + "right": "issue:33357", + "accept": false, + "reason": "Loading hang vs MacOS bus error; both runtime failures but not the same bug." + }, + { + "left": "issue:43122", + "right": "issue:44514", + "accept": false, + "reason": "Tokenization regression vs batched chat-template crash; different surfaces and causes." + }, + { + "left": "issue:42907", + "right": "issue:44945", + "accept": false, + "reason": "Quantized model save failure vs pipeline parallelism output corruption; not one concrete issue." + }, + { + "left": "issue:44704", + "right": "issue:45290", + "accept": false, + "reason": "Processor kwarg forwarding vs chat-template crash on tool-call messages; different code paths." + }, + { + "left": "issue:43232", + "right": "issue:45245", + "accept": false, + "reason": "Generation state handling vs category-count runtime limit; unrelated." + }, + { + "left": "issue:42371", + "right": "issue:44263", + "accept": false, + "reason": "TF32 configuration guidance vs GLM indexer split behavior; no overlap beyond being model-adjacent." + }, + { + "left": "issue:43976", + "right": "issue:44945", + "accept": false, + "reason": "Python version compatibility vs pipeline parallelism output issue; different regressions." + }, + { + "left": "issue:43232", + "right": "issue:44743", + "accept": false, + "reason": "Both are generation-related, but one is sync_gpus kwargs and the other is Qwen3.5 recurrent-state reset." + }, + { + "left": "issue:44295", + "right": "issue:44297", + "accept": false, + "reason": "Buffer access error vs tokenizer_class metadata mismatch; separate tokenizer/model save issues." + }, + { + "left": "issue:39401", + "right": "issue:44488", + "accept": false, + "reason": "Wrong offset mapping vs failure to load a specific model; different symptoms." + }, + { + "left": "issue:42907", + "right": "issue:43937", + "accept": false, + "reason": "Saving dequantized models vs invalid GenerationConfig; unrelated failures." + }, + { + "left": "issue:43408", + "right": "issue:43475", + "accept": false, + "reason": "Model-type warning vs missing attribute on a vision output object; related family, different bug." + }, + { + "left": "issue:44625", + "right": "issue:44829", + "accept": false, + "reason": "num_labels propagation vs flash_attention_3 training degeneration; no shared concrete code path." + }, + { + "left": "issue:29127", + "right": "issue:30990", + "accept": false, + "reason": "LayoutLMv3 error-message clarity vs sentence-transformers loading hang; unrelated." + }, + { + "left": "issue:41628", + "right": "issue:45278", + "accept": false, + "reason": "Specific AutoImageProcessor import error vs broad upgrade-wide import failures; not the same issue." + }, + { + "left": "issue:44360", + "right": "issue:44485", + "accept": false, + "reason": "DSA indexer lacking ReLU vs GLM-5 RoPE implementation; different model internals." + }, + { + "left": "issue:43116", + "right": "issue:45200", + "accept": false, + "reason": "Example-script multi-label output bug vs Gemma 4 token-type defaulting; different behaviors." + }, + { + "left": "issue:43992", + "right": "issue:44704", + "accept": false, + "reason": "Missing encoder weights in from_pretrained vs cached_file kwarg forwarding; different loader bugs." + }, + { + "left": "issue:39401", + "right": "issue:44514", + "accept": false, + "reason": "Tokenizer offset mapping regression vs batched chat-template crash; same general area, not same defect." + }, + { + "left": "issue:43122", + "right": "issue:44779", + "accept": false, + "reason": "Two tokenizer regressions, but for different models and different failure modes." + }, + { + "left": "issue:44038", + "right": "issue:44297", + "accept": false, + "reason": "Qwen3-VL-MoE bug vs tokenizer save metadata mismatch; unrelated." + }, + { + "left": "issue:44261", + "right": "issue:44373", + "accept": false, + "reason": "Precision issue from missing eps config vs wrong docstring; not duplicates." + }, + { + "left": "issue:42371", + "right": "issue:42831", + "accept": false, + "reason": "TF32 API change guidance vs FP8 accuracy regression; different concerns." + }, + { + "left": "issue:44038", + "right": "issue:44295", + "accept": false, + "reason": "Model bug in Qwen3-VL-MoE vs position_ids buffer read error; no concrete overlap." + }, + { + "left": "issue:36296", + "right": "issue:43116", + "accept": false, + "reason": "Tensor-parallel training bug vs example-script classification output issue; different subsystems." + }, + { + "left": "issue:42907", + "right": "issue:43452", + "accept": false, + "reason": "Saving dequantized models vs gguf_file from_pretrained loading breakage; separate loader/saver bugs." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 34, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:22355", + "issue:29127", + "issue:33453", + "issue:38617", + "issue:39401", + "issue:42222", + "issue:42371", + "issue:42673", + "issue:42757", + "issue:42907", + "issue:43064", + "issue:43116", + "issue:43122", + "issue:43278", + "issue:43441", + "issue:43450", + "issue:43493", + "issue:43550", + "issue:43673", + "issue:43784", + "issue:43824", + "issue:43825", + "issue:43827", + "issue:43874", + "issue:43906", + "issue:43931", + "issue:43937", + "issue:43976", + "issue:43986", + "issue:44038", + "issue:44075", + "issue:44222", + "issue:44291", + "issue:44351", + "issue:44485", + "issue:44779", + "issue:44912", + "issue:44933", + "issue:44936", + "issue:44945", + "issue:44977", + "issue:45071", + "issue:45072", + "issue:45083", + "issue:45216", + "issue:45278", + "issue:45446", + "issue:45478" + ], + "soft_pairs": [ + "issue:44038|issue:44222", + "issue:43824|issue:44351", + "issue:43278|issue:44945", + "issue:43122|issue:44977", + "issue:45216|issue:45478", + "issue:42907|issue:44222", + "issue:42757|issue:44351", + "issue:42907|issue:44912", + "issue:42222|issue:43441", + "issue:42673|issue:45072", + "issue:43784|issue:45446", + "issue:43064|issue:43673", + "issue:43784|issue:44933", + "issue:43937|issue:44485", + "issue:43550|issue:43784", + "issue:38617|issue:45278", + "issue:22355|issue:29127", + "issue:43116|issue:43493", + "issue:44075|issue:44936", + "issue:42371|issue:43874", + "issue:43116|issue:43825", + "issue:44222|issue:45478", + "issue:44038|issue:45478", + "issue:43976|issue:45446", + "issue:39401|issue:44779", + "issue:43931|issue:44945", + "issue:33453|issue:39401", + "issue:45083|issue:45446", + "issue:43937|issue:44977", + "issue:43827|issue:45446", + "issue:42673|issue:43906", + "issue:44291|issue:45071", + "issue:43450|issue:43986" + ], + "prepared_review_unit_hash": "6f5d8ccbbeaa83f67240a6095bfd8793e601534243110a06a4a600f8a610e99a", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 33, + "serialized_chars": 37167, + "estimated_input_tokens": 9292, + "estimated_eval_tokens": 18840 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This is a noisy, heterogeneous cluster: import errors, tokenizer regressions, model-specific generation/loading bugs, training issues, docs, and quantization problems. I found no soft pair that looks like the same underlying bug/change.", + "confidence": 0.93, + "canonical_issue_reason": "No single canonical issue for the whole cluster; the items span unrelated failure modes and model families.", + "canonical_pr_reason": null, + "best_issue_reason": "If forced to pick an anchor, issue 45278 is the broadest umbrella for the import-regression subset, but it is still not a good canonical target for the entire cluster.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44038", + "right": "issue:44222", + "accept": false, + "reason": "Both mention MoE/Qwen, but one is a v5 FlashAttention/generation bug and the other is FP8 save_pretrained behavior." + }, + { + "left": "issue:43824", + "right": "issue:44351", + "accept": false, + "reason": "Both are import errors, but they are for different symbols and code paths (`Qwen2_5_VLForConditionalGeneration` vs `HybridCache`)." + }, + { + "left": "issue:43278", + "right": "issue:44945", + "accept": false, + "reason": "Different problems: dtype drift between train/eval vs incorrect output under pipeline parallelism." + }, + { + "left": "issue:43122", + "right": "issue:44977", + "accept": false, + "reason": "Both are v5 regressions, but tokenizer behavior and FlashAttention generation are different bugs." + }, + { + "left": "issue:45216", + "right": "issue:45478", + "accept": false, + "reason": "Both involve Qwen3.5 and save/load, but one is a bad checkpoint from save_pretrained and the other is a from_pretrained load failure." + }, + { + "left": "issue:42907", + "right": "issue:44222", + "accept": false, + "reason": "Both touch saving quantized models, but they concern different model families and different serialization paths." + }, + { + "left": "issue:42757", + "right": "issue:44351", + "accept": false, + "reason": "Both are import failures, but for unrelated packages (`huggingface_hub` vs `transformers`) and different missing names." + }, + { + "left": "issue:42907", + "right": "issue:44912", + "accept": false, + "reason": "Saving dequantized Ministral/Devstral models is unrelated to MXFP4 loading falling back to bf16." + }, + { + "left": "issue:42222", + "right": "issue:43441", + "accept": false, + "reason": "Different model-specific regressions: vitpose breakage vs Ministral-3 FlashAttention failure." + }, + { + "left": "issue:42673", + "right": "issue:45072", + "accept": false, + "reason": "VRAM leakage in Qwen3ForCausalLM dataloader threads is unrelated to dtype mismatches in inference." + }, + { + "left": "issue:43784", + "right": "issue:45446", + "accept": false, + "reason": "One is a missing import in sentence-transformers integration; the other is a PyTorch version check bug for flex_attention." + }, + { + "left": "issue:43064", + "right": "issue:43673", + "accept": false, + "reason": "Trainer/FSDP optimizer state corruption is unrelated to missing GenerationMixin cache during chunked prefill." + }, + { + "left": "issue:43784", + "right": "issue:44933", + "accept": false, + "reason": "Both are import-related, but they are different missing exports from different modules." + }, + { + "left": "issue:43937", + "right": "issue:44485", + "accept": false, + "reason": "GenerationConfig validation and GLM-5 RoPE implementation are distinct issues." + }, + { + "left": "issue:43550", + "right": "issue:43784", + "accept": false, + "reason": "torch.compile/SDPA failure is unrelated to a missing `nn` import." + }, + { + "left": "issue:38617", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing `layer_type_validation` import does not match the broader 'many import errors' report." + }, + { + "left": "issue:22355", + "right": "issue:29127", + "accept": false, + "reason": "No module named transformers.onnx is unrelated to the LayoutLMv3 error-message clarity issue." + }, + { + "left": "issue:43116", + "right": "issue:43493", + "accept": false, + "reason": "Multi-label classification output parsing and SigLIP2 HF-vs-JAX divergence are different bugs." + }, + { + "left": "issue:44075", + "right": "issue:44936", + "accept": false, + "reason": "Unused SGD args and evaluate-after-train failure do not share a concrete code path." + }, + { + "left": "issue:42371", + "right": "issue:43874", + "accept": false, + "reason": "TF32 settings guidance is unrelated to a missing image-processor method in GLM46V." + }, + { + "left": "issue:43116", + "right": "issue:43825", + "accept": false, + "reason": "Example script behavior and pipeline() error messaging are separate issues." + }, + { + "left": "issue:44222", + "right": "issue:45478", + "accept": false, + "reason": "Both mention Qwen3.5/MoE, but one is FP8 save_pretrained and the other is a from_pretrained load error." + }, + { + "left": "issue:44038", + "right": "issue:45478", + "accept": false, + "reason": "Qwen3.5 FlashAttention generation failure is a different code path from Qwen3.5 MoE loading failure." + }, + { + "left": "issue:43976", + "right": "issue:45446", + "accept": false, + "reason": "Python version support regression is unrelated to the flex_attention import/version-check bug." + }, + { + "left": "issue:39401", + "right": "issue:44779", + "accept": false, + "reason": "Both are tokenizer regressions, but they affect different tokenizers and different outputs." + }, + { + "left": "issue:43931", + "right": "issue:44945", + "accept": false, + "reason": "Shape mismatch while loading Qwen3-VL is unrelated to incorrect LLM output under pipeline parallelism." + }, + { + "left": "issue:33453", + "right": "issue:39401", + "accept": false, + "reason": "Tokenizer loading regression and wrong offset mappings are different failure modes." + }, + { + "left": "issue:45083", + "right": "issue:45446", + "accept": false, + "reason": "A multimodal feature-length helper bug is unrelated to the flex_attention import/version issue." + }, + { + "left": "issue:43937", + "right": "issue:44977", + "accept": false, + "reason": "GLM-5 invalid GenerationConfig and Qwen3.5 FlashAttention generation failures are different bugs." + }, + { + "left": "issue:43827", + "right": "issue:45446", + "accept": false, + "reason": "Docs still referencing pipeline() is unrelated to the AuxRequest import/version-check issue." + }, + { + "left": "issue:42673", + "right": "issue:43906", + "accept": false, + "reason": "The isolated reproduction of issue 38071 is not the same bug as Qwen3ForCausalLM VRAM leakage." + }, + { + "left": "issue:44291", + "right": "issue:45071", + "accept": false, + "reason": "Unexpected `_is_hf_initialized` during init_empty_weights is a different code path from PretrainedConfig type checking." + }, + { + "left": "issue:43450", + "right": "issue:43986", + "accept": false, + "reason": "Batched video processor shape errors and AutoProcessor crashes without torchvision are different problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 35, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:38617", + "issue:39401", + "issue:41628", + "issue:42175", + "issue:42222", + "issue:42371", + "issue:42617", + "issue:42673", + "issue:42994", + "issue:43012", + "issue:43066", + "issue:43097", + "issue:43122", + "issue:43408", + "issue:43421", + "issue:43441", + "issue:43452", + "issue:43577", + "issue:43582", + "issue:43673", + "issue:43742", + "issue:43874", + "issue:43937", + "issue:43986", + "issue:44117", + "issue:44220", + "issue:44261", + "issue:44263", + "issue:44410", + "issue:44483", + "issue:44484", + "issue:44485", + "issue:44493", + "issue:44496", + "issue:44829", + "issue:44933", + "issue:45003", + "issue:45071", + "issue:45081", + "issue:45216", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45405", + "issue:45446", + "issue:45459", + "issue:45464", + "issue:45478" + ], + "soft_pairs": [ + "issue:45405|issue:45446", + "issue:44484|issue:45310", + "issue:39401|issue:43421", + "issue:44829|issue:45216", + "issue:38617|issue:41628", + "issue:44117|issue:45081", + "issue:42673|issue:43582", + "issue:42673|issue:43577", + "issue:44410|issue:45310", + "issue:43408|issue:43742", + "issue:38617|issue:44220", + "issue:43937|issue:44496", + "issue:42371|issue:44261", + "issue:44483|issue:45464", + "issue:42175|issue:42222", + "issue:42617|issue:45341", + "issue:44933|issue:45278", + "issue:44484|issue:45478", + "issue:43066|issue:43122", + "issue:43673|issue:45341", + "issue:43097|issue:44493", + "issue:42222|issue:45459", + "issue:43874|issue:44261", + "issue:44410|issue:45478", + "issue:43441|issue:44829", + "issue:43986|issue:44484", + "issue:43012|issue:43408", + "issue:42222|issue:42994", + "issue:44263|issue:44485", + "issue:41628|issue:45071", + "issue:45003|issue:45341", + "issue:43097|issue:43452" + ], + "prepared_review_unit_hash": "3c8f8fcafa03e7dce582abc188ad29ecc04ff38aa7aab39801acdb3d53bebbf0", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 32, + "serialized_chars": 36992, + "estimated_input_tokens": 9248, + "estimated_eval_tokens": 18752 + }, + "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 import/tokenizer/model-loading bugs; there is one likely true duplicate pair around tokenizer behavior in v5, but the rest are too semantically different to merge.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 45278 is the broadest umbrella report in the set: it summarizes many import errors after upgrading to 5.5.0, making it the best single issue to anchor nearby import-regression reports.", + "canonical_pr_reason": null, + "best_issue_reason": "45278 is the most representative global issue because it is open, recent, and broad enough to cover the cluster\u2019s import-regression theme better than the many narrow one-off bug reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45405", + "right": "issue:45446", + "accept": false, + "reason": "Different subsystems: PEFT version bump vs a PyTorch version check bug in flex_attention." + }, + { + "left": "issue:44484", + "right": "issue:45310", + "accept": false, + "reason": "Both mention model loading/versioning, but one is about shard size in save_pretrained and the other is a Qwen3.5 MoE from_pretrained regression." + }, + { + "left": "issue:39401", + "right": "issue:43421", + "accept": false, + "reason": "Both involve tokenizers, but one is a wrong offset_mapping bug and the other is runtime special-token post-processor updates." + }, + { + "left": "issue:44829", + "right": "issue:45216", + "accept": false, + "reason": "FlashAttention training degeneration and Qwen3.5 checkpoint saving are different code paths and failure modes." + }, + { + "left": "issue:38617", + "right": "issue:41628", + "accept": false, + "reason": "Both are import errors, but they concern different missing symbols and different API surfaces." + }, + { + "left": "issue:44117", + "right": "issue:45081", + "accept": false, + "reason": "One is a tokenizer mapping None-handling bug; the other is a Mistral regex patch crash on tokenizer backend access." + }, + { + "left": "issue:42673", + "right": "issue:43582", + "accept": false, + "reason": "VRAM leakage in dataloader threads is unrelated to an Apple Silicon TypeError in caching allocator warmup." + }, + { + "left": "issue:42673", + "right": "issue:43577", + "accept": false, + "reason": "A VRAM leak bug and a dtype propagation bug in Blip2 loading are unrelated." + }, + { + "left": "issue:44410", + "right": "issue:45310", + "accept": false, + "reason": "Both are Qwen-family reports, but one is missing projections in qwen3next and the other is a MoE from_pretrained error." + }, + { + "left": "issue:43408", + "right": "issue:43742", + "accept": false, + "reason": "A model-type mismatch warning for sam3_tracker is not the same bug as a KeyError loading MobileLLM-125M." + }, + { + "left": "issue:38617", + "right": "issue:44220", + "accept": false, + "reason": "An import error for layer_type_validation is unrelated to _torch_extract_fbank_features() behavior." + }, + { + "left": "issue:43937", + "right": "issue:44496", + "accept": false, + "reason": "GenerationConfig validation failure and unrecognized-model config loading are separate problems." + }, + { + "left": "issue:42371", + "right": "issue:44261", + "accept": false, + "reason": "TF32 configuration guidance is unrelated to MLA q_a_layernorm missing rms_norm_eps." + }, + { + "left": "issue:44483", + "right": "issue:45464", + "accept": false, + "reason": "Both concern chat/completions, but one is a general API request rejection and the other is a streaming inference failure on a specific model." + }, + { + "left": "issue:42175", + "right": "issue:42222", + "accept": false, + "reason": "TensorFlow dependency packaging and vitpose model breakage are not the same underlying issue." + }, + { + "left": "issue:42617", + "right": "issue:45341", + "accept": false, + "reason": "A 3d_parallel.py runtime issue and a testing_utils bug are unrelated." + }, + { + "left": "issue:44933", + "right": "issue:45278", + "accept": false, + "reason": "A specific missing import from image_utils is not the same concrete bug as a broad report about many import errors after upgrade." + }, + { + "left": "issue:44484", + "right": "issue:45478", + "accept": false, + "reason": "Both are Qwen3.5/MoE-related only in a broad sense; one is about shard size, the other about from_pretrained loading failure." + }, + { + "left": "issue:43066", + "right": "issue:43122", + "accept": true, + "reason": "These both describe the same tokenizer regression in v5: tokenization output changes because the decoder/type handling is wrong." + }, + { + "left": "issue:43673", + "right": "issue:45341", + "accept": false, + "reason": "Generation cache missing in chunked_prefill is unrelated to a minor testing utility bug." + }, + { + "left": "issue:43097", + "right": "issue:44493", + "accept": false, + "reason": "A removed config field warning and unexpected position-id key errors are different regression classes." + }, + { + "left": "issue:42222", + "right": "issue:45459", + "accept": false, + "reason": "Vitpose model breakage and tokenizer error masking when protobuf is absent are unrelated." + }, + { + "left": "issue:43874", + "right": "issue:44261", + "accept": false, + "reason": "A missing image-patch count method and a missing rms_norm_eps config value are different model/component bugs." + }, + { + "left": "issue:44410", + "right": "issue:45478", + "accept": false, + "reason": "Missing projections in qwen3next is not the same concrete defect as a Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:43441", + "right": "issue:44829", + "accept": false, + "reason": "Both mention FlashAttention, but they target different models and distinct failures, so they should not be merged." + }, + { + "left": "issue:43986", + "right": "issue:44484", + "accept": false, + "reason": "AutoProcessor/video-model torchvision crash is unrelated to shard-size behavior in save_pretrained." + }, + { + "left": "issue:43012", + "right": "issue:43408", + "accept": false, + "reason": "A bfloat16 compilation warning and a sam3 model-type warning are unrelated." + }, + { + "left": "issue:42222", + "right": "issue:42994", + "accept": false, + "reason": "Vitpose model breakage and quantized model saving failure are different bugs." + }, + { + "left": "issue:44263", + "right": "issue:44485", + "accept": false, + "reason": "A torch.split return-value issue and a GLM-5 RoPE implementation discussion are not the same change." + }, + { + "left": "issue:41628", + "right": "issue:45071", + "accept": false, + "reason": "An AutoImageProcessor import failure is different from a PreTrainedConfig type-checking regression." + }, + { + "left": "issue:45003", + "right": "issue:45341", + "accept": false, + "reason": "Unsafe sys.modules access in modeling_utils is unrelated to a testing_utils bug." + }, + { + "left": "issue:43097", + "right": "issue:43452", + "accept": false, + "reason": "Removed tie_embeddings_and_encoder_decoder support is unrelated to gguf_file breakage in from_pretrained." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats the cluster as mostly heterogeneous, selects 45278 as the broad import-regression anchor, and its rejected soft-pairs are conservative. The one accepted tokenizer pair is plausibly mergeable as a single v5 tokenizer regression fix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 36, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:38617", + "issue:42757", + "issue:42886", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43097", + "issue:43335", + "issue:43352", + "issue:43381", + "issue:43408", + "issue:43502", + "issue:43531", + "issue:43550", + "issue:43582", + "issue:43825", + "issue:43873", + "issue:43874", + "issue:43901", + "issue:43906", + "issue:43931", + "issue:43976", + "issue:43986", + "issue:43992", + "issue:44038", + "issue:44062", + "issue:44117", + "issue:44261", + "issue:44263", + "issue:44315", + "issue:44351", + "issue:44368", + "issue:44380", + "issue:44410", + "issue:44464", + "issue:44484", + "issue:44485", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44987", + "issue:45071", + "issue:45103", + "issue:45245", + "issue:45290", + "issue:45341", + "issue:45372", + "issue:45375" + ], + "soft_pairs": [ + "issue:44117|issue:44987", + "issue:43986|issue:45341", + "issue:43931|issue:44410", + "issue:43335|issue:45071", + "issue:43408|issue:44368", + "issue:43931|issue:44038", + "issue:43986|issue:45372", + "issue:43874|issue:44263", + "issue:44829|issue:45245", + "issue:43976|issue:45071", + "issue:44464|issue:45071", + "issue:43531|issue:45341", + "issue:43064|issue:44315", + "issue:38617|issue:42757", + "issue:44261|issue:44485", + "issue:44410|issue:45375", + "issue:44704|issue:45103", + "issue:43502|issue:43992", + "issue:44062|issue:44117", + "issue:43986|issue:44117", + "issue:38617|issue:44351", + "issue:43582|issue:45341", + "issue:44038|issue:44410", + "issue:43097|issue:43873", + "issue:43992|issue:44484", + "issue:43986|issue:45290", + "issue:43066|issue:43906", + "issue:43825|issue:44655", + "issue:38617|issue:43097", + "issue:43352|issue:44380", + "issue:43097|issue:43901", + "issue:43012|issue:43550", + "issue:43012|issue:44368", + "issue:42886|issue:43381" + ], + "prepared_review_unit_hash": "26b5ed52c893d95a760af23e2b1b323c442b944cecd0431c6fe67f2990de8383", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 34, + "serialized_chars": 37881, + "estimated_input_tokens": 9471, + "estimated_eval_tokens": 19198 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is heterogeneous: most pairs are only superficially similar (imports, warnings, tokenizer/Qwen/GLM mentions) but do not appear to describe the same underlying bug or change.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43873 is the most active unresolved bug here, with the broadest practical impact and the clearest ongoing discussion, even though it is not a true duplicate hub for the rest of the set.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44117", + "right": "issue:44987", + "accept": false, + "reason": "Both are loading-related, but one is about TOKENIZER_MAPPING_NAMES returning None and the other about a transformers>=5.1.0 model load failure; different failure modes and code paths." + }, + { + "left": "issue:43986", + "right": "issue:45341", + "accept": false, + "reason": "AutoProcessor/video-model dependency crash vs a testing_utils bug; no shared underlying bug." + }, + { + "left": "issue:43931", + "right": "issue:44410", + "accept": false, + "reason": "Both involve Qwen-family models, but one is a Qwen3-VL weight-shape mismatch and the other is missing layer projections in qwen3next; distinct bugs." + }, + { + "left": "issue:43335", + "right": "issue:45071", + "accept": false, + "reason": "SwitchTransformers sparse-layer config bug vs PretrainedConfig type checking regression; unrelated." + }, + { + "left": "issue:43408", + "right": "issue:44368", + "accept": false, + "reason": "Both are warnings, but they concern different model families and different config mismatches." + }, + { + "left": "issue:43931", + "right": "issue:44038", + "accept": false, + "reason": "Same broad Qwen3-VL area, but one is a shape mismatch and the other is a MoE load bug in transformers 5.0; not the same issue." + }, + { + "left": "issue:43986", + "right": "issue:45372", + "accept": false, + "reason": "Both are processor-loading failures, but the missing dependency/import is different and the affected processors are unrelated." + }, + { + "left": "issue:43874", + "right": "issue:44263", + "accept": false, + "reason": "Both are GLM-related, but one is a missing processor method and the other is a torch.split/indexer bug; different code paths." + }, + { + "left": "issue:44829", + "right": "issue:45245", + "accept": false, + "reason": "FlashAttention training degeneration and category-count limit are unrelated bugs." + }, + { + "left": "issue:43976", + "right": "issue:45071", + "accept": false, + "reason": "Python version compatibility issue vs config type-checking regression; no shared fix." + }, + { + "left": "issue:44464", + "right": "issue:45071", + "accept": false, + "reason": "Compiled chunked generation inconsistency vs PretrainedConfig typing; unrelated." + }, + { + "left": "issue:43531", + "right": "issue:45341", + "accept": false, + "reason": "Qwen3-MoE sliding_window behavior vs a testing utility typo/bug; not the same bug." + }, + { + "left": "issue:43064", + "right": "issue:44315", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state corruption vs Liger kernel not applied with model_init; different training paths." + }, + { + "left": "issue:38617", + "right": "issue:42757", + "accept": false, + "reason": "Both are ImportErrors, but they are for different missing symbols in different modules." + }, + { + "left": "issue:44261", + "right": "issue:44485", + "accept": false, + "reason": "MLA rms_norm_eps precision issue vs GLM-5 RoPE implementation; unrelated model internals." + }, + { + "left": "issue:44410", + "right": "issue:45375", + "accept": false, + "reason": "Both are Qwen config/model bugs, but one is missing attention projections and the other is a missing vision config field." + }, + { + "left": "issue:44704", + "right": "issue:45103", + "accept": false, + "reason": "AutoProcessor kwargs forwarding bug vs auto_docstring crash under future annotations; different layers and fixes." + }, + { + "left": "issue:43502", + "right": "issue:43992", + "accept": false, + "reason": "Local-files-only network requests vs missing embed_tokens weight on load; separate loading bugs." + }, + { + "left": "issue:44062", + "right": "issue:44117", + "accept": false, + "reason": "Tokenizer AddedToken argument duplication vs TOKENIZER_MAPPING_NAMES assuming non-None; both tokenizer-related but not the same defect." + }, + { + "left": "issue:43986", + "right": "issue:44117", + "accept": false, + "reason": "Processor-loading crash without torchvision vs tokenizer mapping None-handling; different subsystems and symptoms." + }, + { + "left": "issue:38617", + "right": "issue:44351", + "accept": false, + "reason": "Both are import-name errors, but they concern different missing public symbols and likely separate API regressions." + }, + { + "left": "issue:43582", + "right": "issue:45341", + "accept": false, + "reason": "Apple Silicon caching allocator TypeError vs testing_utils bug; unrelated." + }, + { + "left": "issue:44038", + "right": "issue:44410", + "accept": false, + "reason": "Both reference Qwen3, but one is a model load failure and the other is missing layer projections; not the same underlying bug." + }, + { + "left": "issue:43097", + "right": "issue:43873", + "accept": false, + "reason": "Config deprecation/removal warning issue vs quantization offloading bug; unrelated." + }, + { + "left": "issue:43992", + "right": "issue:44484", + "accept": false, + "reason": "Missing encoder weight on load vs default max_shard_size question; different concerns." + }, + { + "left": "issue:43986", + "right": "issue:45290", + "accept": false, + "reason": "AutoProcessor video-model dependency crash vs chat-template tokenization crash with tool calls; different code paths." + }, + { + "left": "issue:43066", + "right": "issue:43906", + "accept": false, + "reason": "A title referencing an isolated reproduction is not enough to match a specific tokenizer-decoder bug; the concrete failure descriptions differ." + }, + { + "left": "issue:43825", + "right": "issue:44655", + "accept": false, + "reason": "Pipeline task messaging vs saving Pipeline objects; separate pipeline issues." + }, + { + "left": "issue:38617", + "right": "issue:43097", + "accept": false, + "reason": "ImportError for a missing config symbol vs removal of a config option; not the same bug." + }, + { + "left": "issue:43352", + "right": "issue:44380", + "accept": false, + "reason": "FlashAttention 2 support error for Nemotron vs GPT2 attention scaling ignored under SDPA/FlashAttention; different models and fixes." + }, + { + "left": "issue:43097", + "right": "issue:43901", + "accept": false, + "reason": "Removed tie_embeddings setting vs text-classification docs mismatch; unrelated." + }, + { + "left": "issue:43012", + "right": "issue:43550", + "accept": false, + "reason": "bfloat16 compile warning vs Bamba torch.compile SDPA failure; different symptoms and likely different fixes." + }, + { + "left": "issue:43012", + "right": "issue:44368", + "accept": false, + "reason": "PyTorch precision warning during compile vs tie_word_embeddings warning; both warnings but not the same behavior." + }, + { + "left": "issue:42886", + "right": "issue:43381", + "accept": false, + "reason": "Offline cache-loading failure vs gradient checkpointing in eval mode; unrelated runtime paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 37, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:36010", + "issue:36683", + "issue:38617", + "issue:42757", + "issue:42907", + "issue:42994", + "issue:43012", + "issue:43064", + "issue:43066", + "issue:43208", + "issue:43299", + "issue:43352", + "issue:43475", + "issue:43502", + "issue:43525", + "issue:43526", + "issue:43550", + "issue:43582", + "issue:43618", + "issue:43756", + "issue:43761", + "issue:43824", + "issue:43867", + "issue:43901", + "issue:43937", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44117", + "issue:44188", + "issue:44220", + "issue:44230", + "issue:44263", + "issue:44380", + "issue:44483", + "issue:44485", + "issue:44492", + "issue:44514", + "issue:44655", + "issue:44683", + "issue:44792", + "issue:44938", + "issue:45216", + "issue:45290", + "issue:45381", + "issue:45440", + "issue:45479" + ], + "soft_pairs": [ + "issue:43208|issue:43550", + "issue:43937|issue:44792", + "issue:42907|issue:44655", + "issue:44380|issue:45479", + "issue:43525|issue:44117", + "issue:43352|issue:43976", + "issue:43064|issue:44380", + "issue:43208|issue:43526", + "issue:44483|issue:44514", + "issue:43618|issue:44380", + "issue:36683|issue:43582", + "issue:43618|issue:43761", + "issue:43012|issue:44683", + "issue:44483|issue:45290", + "issue:42757|issue:43824", + "issue:38617|issue:43502", + "issue:44038|issue:44230", + "issue:44483|issue:45381", + "issue:43299|issue:43867", + "issue:43901|issue:44655", + "issue:42994|issue:44655", + "issue:43756|issue:44485", + "issue:36683|issue:43475", + "issue:43867|issue:45216", + "issue:44220|issue:44655", + "issue:43352|issue:43937", + "issue:43066|issue:45381", + "issue:44077|issue:44380", + "issue:36683|issue:44077", + "issue:44263|issue:45440", + "issue:43064|issue:44492", + "issue:43976|issue:44938", + "issue:36683|issue:43525", + "issue:36010|issue:42757", + "issue:44188|issue:44380", + "issue:43618|issue:45216" + ], + "prepared_review_unit_hash": "07e4f140947f9edc3fd706ededa0158057371a599de9b9bc7ba3e83f0e5a3663", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 36, + "serialized_chars": 37554, + "estimated_input_tokens": 9389, + "estimated_eval_tokens": 19034 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "The cluster is highly heterogeneous: most items are distinct model-specific regressions, import errors, save/load bugs, or unrelated runtime issues. None of the suggested soft edges look like the same underlying bug or change, so I would reject all of them.", + "confidence": 0.97, + "canonical_issue_reason": "Issue 44188 is the strongest single anchor because it is the most active open issue and a core runtime regression, but the set is too diverse for any issue to truly represent the whole cluster.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44188 is the best available cluster representative: it is open, high-activity, and centered on core execution behavior. Even so, it does not subsume the other items well.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43208", + "right": "issue:43550", + "accept": false, + "reason": "Different bugs and model families: xLSTM training issues vs Bamba torch.compile/SDPA failure." + }, + { + "left": "issue:43937", + "right": "issue:44792", + "accept": false, + "reason": "GLM-5 invalid GenerationConfig vs janus test failure; unrelated failure modes." + }, + { + "left": "issue:42907", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but one is dequantized Ministrals/Devstrals and the other is Pipeline.save_pretrained; different code paths." + }, + { + "left": "issue:44380", + "right": "issue:45479", + "accept": false, + "reason": "GPT2 attention scaling under SDPA/FlashAttention vs degenerate zero loss in sequence classification; unrelated." + }, + { + "left": "issue:43525", + "right": "issue:44117", + "accept": false, + "reason": "Missing pad_token_id on Llama4Config is unrelated to TOKENIZER_MAPPING_NAMES returning None." + }, + { + "left": "issue:43352", + "right": "issue:43976", + "accept": false, + "reason": "Flash Attention 2 support issue vs Python version compatibility issue; not the same bug." + }, + { + "left": "issue:43064", + "right": "issue:44380", + "accept": false, + "reason": "FSDP2/PEFT optimizer-state mismatch is unrelated to GPT2 attention scaling behavior." + }, + { + "left": "issue:43208", + "right": "issue:43526", + "accept": false, + "reason": "xLSTM training bugs vs BeitImageProcessorFast label reduction bug; different subsystems." + }, + { + "left": "issue:44483", + "right": "issue:44514", + "accept": false, + "reason": "Both are chat-related, but one is /v1/chat/completions request handling and the other is apply_chat_template batching/padding." + }, + { + "left": "issue:43618", + "right": "issue:44380", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to GPT2 scaling under SDPA/FlashAttention." + }, + { + "left": "issue:36683", + "right": "issue:43582", + "accept": false, + "reason": "Gemma3Config missing vocab_size vs Apple Silicon TypeError in caching_allocator_warmup; unrelated." + }, + { + "left": "issue:43618", + "right": "issue:43761", + "accept": false, + "reason": "Both involve CLIP, but one is attentions not assigned and the other is hidden_states missing; different concrete failures." + }, + { + "left": "issue:43012", + "right": "issue:44683", + "accept": false, + "reason": "PyTorch warning on bfloat16 compilation is not the same as flex_attention failing on torch>=2.9." + }, + { + "left": "issue:44483", + "right": "issue:45290", + "accept": false, + "reason": "Different chat surfaces: API request rejection vs apply_chat_template crash on assistant tool-call messages." + }, + { + "left": "issue:42757", + "right": "issue:43824", + "accept": false, + "reason": "Import errors, but for different symbols and packages: huggingface_hub.is_offline_mode vs transformers.Qwen2_5_VLForConditionalGeneration." + }, + { + "left": "issue:38617", + "right": "issue:43502", + "accept": false, + "reason": "Import failure from configuration_utils is unrelated to local_files_only still triggering network requests." + }, + { + "left": "issue:44038", + "right": "issue:44230", + "accept": false, + "reason": "Same model family, but one is a loading regression and the other is an fp8 support request; not the same fix." + }, + { + "left": "issue:44483", + "right": "issue:45381", + "accept": false, + "reason": "Chat-completions request rejection is unrelated to Qwen2.5-VL video vision_position_ids being wrong." + }, + { + "left": "issue:43299", + "right": "issue:43867", + "accept": false, + "reason": "Qwen3VL-Moe loading breakage is not the same as a sorted-state_dict load error." + }, + { + "left": "issue:43901", + "right": "issue:44655", + "accept": false, + "reason": "Docs mismatch for return_all_scores is unrelated to inability to save Pipeline objects." + }, + { + "left": "issue:42994", + "right": "issue:44655", + "accept": false, + "reason": "Both involve saving, but quantized model save failure and Pipeline.save_pretrained failure are different artifacts and bugs." + }, + { + "left": "issue:43756", + "right": "issue:44485", + "accept": false, + "reason": "Smollm3 RoPE-layer issue and GLM-5 RoPE implementation discussion are different model-specific problems." + }, + { + "left": "issue:36683", + "right": "issue:43475", + "accept": false, + "reason": "Different missing-attribute regressions on different models: Gemma3Config.vocab_size vs Sam3VisionEncoderOutput.fpn_position_embeddings." + }, + { + "left": "issue:43867", + "right": "issue:45216", + "accept": false, + "reason": "Both concern saving/loading checkpoints, but the sorted state_dict load error and the Qwen3.5 save_pretrained regression are distinct." + }, + { + "left": "issue:44220", + "right": "issue:44655", + "accept": false, + "reason": "Audio feature extraction bug vs Pipeline saving bug; no shared code path." + }, + { + "left": "issue:43352", + "right": "issue:43937", + "accept": false, + "reason": "Flash Attention support for NemotronH is unrelated to GLM-5 GenerationConfig validation." + }, + { + "left": "issue:43066", + "right": "issue:45381", + "accept": false, + "reason": "Wrong tokenizer decoder type in v5 is unrelated to Qwen2.5-VL video position IDs." + }, + { + "left": "issue:44077", + "right": "issue:44380", + "accept": false, + "reason": "patchtsmixer post_init validation and GPT2 attention scaling are unrelated." + }, + { + "left": "issue:36683", + "right": "issue:44077", + "accept": false, + "reason": "Gemma3Config missing vocab_size is unrelated to patchtsmixer post_init allowance." + }, + { + "left": "issue:44263", + "right": "issue:45440", + "accept": false, + "reason": "Both are MoE-related, but they affect different models and different internals." + }, + { + "left": "issue:43064", + "right": "issue:44492", + "accept": false, + "reason": "Optimizer-state corruption and a cache-strategy typo are unrelated." + }, + { + "left": "issue:43976", + "right": "issue:44938", + "accept": false, + "reason": "Both are Python compatibility issues, but they target different interpreter versions and likely different causes." + }, + { + "left": "issue:36010", + "right": "issue:42757", + "accept": false, + "reason": "Different import failures from different modules and symbols." + }, + { + "left": "issue:44188", + "right": "issue:44380", + "accept": false, + "reason": "Different attention/backend regressions: allow_is_bidirectional_skip under torch.compile vs GPT2 scaling under SDPA/FlashAttention." + }, + { + "left": "issue:43618", + "right": "issue:45216", + "accept": false, + "reason": "CLIP attentions regression is unrelated to the Qwen3.5 save_pretrained checkpoint bug." + } + ] + }, + "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": 13, + "review_unit_index": 38, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:28282", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:41669", + "issue:42548", + "issue:42890", + "issue:43010", + "issue:43064", + "issue:43065", + "issue:43232", + "issue:43352", + "issue:43502", + "issue:43575", + "issue:43618", + "issue:43644", + "issue:43723", + "issue:43867", + "issue:43931", + "issue:43976", + "issue:43992", + "issue:44062", + "issue:44075", + "issue:44077", + "issue:44112", + "issue:44222", + "issue:44242", + "issue:44246", + "issue:44351", + "issue:44360", + "issue:44387", + "issue:44484", + "issue:44556", + "issue:44857", + "issue:44938", + "issue:44964", + "issue:44977", + "issue:45071", + "issue:45125", + "issue:45278", + "issue:45310", + "issue:45341", + "issue:45357", + "issue:45375", + "issue:45381", + "issue:45440", + "issue:45478" + ], + "soft_pairs": [ + "issue:43992|issue:45310", + "issue:36683|issue:44242", + "issue:43010|issue:45278", + "issue:36683|issue:43352", + "issue:41669|issue:44246", + "issue:43352|issue:43931", + "issue:43867|issue:45375", + "issue:36683|issue:43064", + "issue:43992|issue:45478", + "issue:44938|issue:45071", + "issue:43867|issue:45357", + "issue:43010|issue:43976", + "issue:44938|issue:45381", + "issue:44360|issue:45440", + "issue:43618|issue:44077", + "issue:43644|issue:44938", + "issue:36296|issue:43575", + "issue:43867|issue:44484", + "issue:43618|issue:44857", + "issue:42890|issue:44964", + "issue:42890|issue:44112", + "issue:28282|issue:33453", + "issue:43618|issue:44977", + "issue:43502|issue:44062", + "issue:43010|issue:44075", + "issue:42890|issue:45341", + "issue:43723|issue:44556", + "issue:44387|issue:44556", + "issue:43010|issue:43232", + "issue:44222|issue:45125", + "issue:43010|issue:43065", + "issue:42548|issue:44351", + "issue:45125|issue:45478" + ], + "prepared_review_unit_hash": "f2149912b8b6170e28f17d509ed7cd0e0d5a0188463ad8cbd902664207207d0c", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 47, + "item_count": 47, + "soft_pair_count": 33, + "serialized_chars": 36385, + "estimated_input_tokens": 9097, + "estimated_eval_tokens": 18450 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Most items are unrelated standalone issues; the only obvious exact duplicate in the set is the Qwen3.5 MoE from_pretrained regression (45310/45478), but it is not one of the listed soft edges. No soft pair meets a strict same-bug/same-change duplicate threshold.", + "confidence": 0.58, + "canonical_issue_reason": "issue:45310 is the clearest representative issue in this set: it is a concrete, narrowly scoped regression report for Qwen3.5 MoE from_pretrained, and issue:45478 is an exact-title duplicate of it.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:45310 is the best global representative because it describes a specific reproducible bug, is well-scoped, and has a near-identical duplicate, making it a strong canonical triage target.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43992", + "right": "issue:45310", + "accept": false, + "reason": "Different code paths: UMT5Encoder missing embed_tokens.weight vs Qwen3.5 MoE from_pretrained error." + }, + { + "left": "issue:36683", + "right": "issue:44242", + "accept": false, + "reason": "Gemma3Config vocab_size attribute error is unrelated to load-balancing loss not being added." + }, + { + "left": "issue:43010", + "right": "issue:45278", + "accept": false, + "reason": "Decorator/no_grad cache update bug is unrelated to broad import errors after upgrading transformers." + }, + { + "left": "issue:36683", + "right": "issue:43352", + "accept": false, + "reason": "Config attribute bug and Flash Attention 2 unsupported error are different problems." + }, + { + "left": "issue:41669", + "right": "issue:44246", + "accept": false, + "reason": "Both mention import slowdowns, but one is specifically about import-star usage in models while the other is a generic intermittent import-time complaint; not clearly the same underlying bug." + }, + { + "left": "issue:43352", + "right": "issue:43931", + "accept": false, + "reason": "Flash Attention support check vs Qwen3-VL weight shape mismatch are unrelated." + }, + { + "left": "issue:43867", + "right": "issue:45375", + "accept": false, + "reason": "State_dict sorting/load error is not the same as a missing config field dropped by strict parsing." + }, + { + "left": "issue:36683", + "right": "issue:43064", + "accept": false, + "reason": "Gemma3Config vocab_size bug and FSDP2 optimizer-state mismatch are unrelated." + }, + { + "left": "issue:43992", + "right": "issue:45478", + "accept": false, + "reason": "UMT5Encoder missing embedding weights is unrelated to the Qwen3.5 MoE from_pretrained regression." + }, + { + "left": "issue:44938", + "right": "issue:45071", + "accept": false, + "reason": "Python 3.14 import/load failure is not the same as PretrainedConfig type-checking breakage." + }, + { + "left": "issue:43867", + "right": "issue:45357", + "accept": false, + "reason": "Model load error from state_dict ordering is unrelated to incorrect save_pretrained visual encoder keys." + }, + { + "left": "issue:43010", + "right": "issue:43976", + "accept": false, + "reason": "Generation kwargs/cache update bug is unrelated to a Python-version compatibility issue." + }, + { + "left": "issue:44938", + "right": "issue:45381", + "accept": false, + "reason": "Interpreter load failure and Qwen2.5-VL video position-id bug are different issues." + }, + { + "left": "issue:44360", + "right": "issue:45440", + "accept": false, + "reason": "DSA indexer missing ReLU and DeepseekV3MoE divergence are unrelated code paths." + }, + { + "left": "issue:43618", + "right": "issue:44077", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to patchtsmixer post_init eligibility." + }, + { + "left": "issue:43644", + "right": "issue:44938", + "accept": false, + "reason": "Non-persistent buffer junk fill is unrelated to Python 3.14 import failure." + }, + { + "left": "issue:36296", + "right": "issue:43575", + "accept": false, + "reason": "Tensor parallel training bug and tp-related OOM on Qwen2-57B are not the same concrete defect." + }, + { + "left": "issue:43867", + "right": "issue:44484", + "accept": false, + "reason": "State_dict sorting load error and save_pretrained shard-size question are unrelated." + }, + { + "left": "issue:43618", + "right": "issue:44857", + "accept": false, + "reason": "Missing CLIPOutput attentions and LwDetrImageLoss AMP crash are different bugs." + }, + { + "left": "issue:42890", + "right": "issue:44964", + "accept": false, + "reason": "SamHQ test seed flakiness is unrelated to Phi-4 multimodal model loading failure." + }, + { + "left": "issue:28282", + "right": "issue:33453", + "accept": false, + "reason": "PyTorch-missing AutoModel ImportError and tokenizer loading regression are different issues." + }, + { + "left": "issue:43618", + "right": "issue:44977", + "accept": false, + "reason": "CLIPOutput attentions regression is unrelated to Qwen3.5 flash-attention generation issues." + }, + { + "left": "issue:43502", + "right": "issue:44062", + "accept": false, + "reason": "local_files_only network leakage and AddedToken keyword collision are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:44075", + "accept": false, + "reason": "Cache update no_grad behavior and SGD args being ignored are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:45341", + "accept": false, + "reason": "SamHQ test flakiness is unrelated to the testing_utils bug." + }, + { + "left": "issue:43723", + "right": "issue:44556", + "accept": false, + "reason": "Tokenizer loading regression and checkpoint reload regression after version upgrade are different load paths." + }, + { + "left": "issue:44387", + "right": "issue:44556", + "accept": false, + "reason": "Int4 CUDA reserved-memory OOM and checkpoint reload failure are unrelated." + }, + { + "left": "issue:43010", + "right": "issue:43232", + "accept": false, + "reason": "Cache update no_grad bug and generation kwargs after sync_gpus are different generation issues." + }, + { + "left": "issue:44222", + "right": "issue:45125", + "accept": false, + "reason": "FP8 save_pretrained moe and missing _tp_plan for tensor parallelism are different Qwen/MoE problems." + }, + { + "left": "issue:43010", + "right": "issue:43065", + "accept": false, + "reason": "Cache/layer update decoration bug is unrelated to the Sam3PixelDecoder dummy Conv2d issue." + }, + { + "left": "issue:42548", + "right": "issue:44351", + "accept": false, + "reason": "cannot import PreTrainedModel and cannot import HybridCache are different export/import failures." + }, + { + "left": "issue:45125", + "right": "issue:45478", + "accept": false, + "reason": "Missing tensor-parallel plan and Qwen3.5 MoE from_pretrained error are different problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 39, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:30990", + "issue:31795", + "issue:33357", + "issue:33453", + "issue:34689", + "issue:36296", + "issue:38175", + "issue:38617", + "issue:41628", + "issue:42548", + "issue:42757", + "issue:42831", + "issue:42890", + "issue:43010", + "issue:43012", + "issue:43232", + "issue:43299", + "issue:43502", + "issue:43526", + "issue:43576", + "issue:43673", + "issue:43824", + "issue:43827", + "issue:43874", + "issue:43901", + "issue:44075", + "issue:44222", + "issue:44230", + "issue:44273", + "issue:44315", + "issue:44351", + "issue:44355", + "issue:44360", + "issue:44368", + "issue:44373", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44655", + "issue:44749", + "issue:44871", + "issue:44908", + "issue:44961", + "issue:45003", + "issue:45071", + "issue:45125", + "issue:45375", + "issue:45468" + ], + "soft_pairs": [ + "issue:44355|issue:44749", + "issue:45125|issue:45375", + "issue:30990|issue:31795", + "issue:43502|issue:44485", + "issue:43010|issue:44368", + "issue:44075|issue:44871", + "issue:44315|issue:44360", + "issue:42890|issue:44373", + "issue:38175|issue:43502", + "issue:43299|issue:44556", + "issue:43502|issue:44961", + "issue:44355|issue:44655", + "issue:44355|issue:45071", + "issue:44222|issue:44230", + "issue:42831|issue:43874", + "issue:42890|issue:43232", + "issue:42757|issue:43576", + "issue:33453|issue:34689", + "issue:33453|issue:36296", + "issue:43827|issue:44961", + "issue:43576|issue:43824", + "issue:43576|issue:45071", + "issue:43673|issue:44492", + "issue:38617|issue:43576", + "issue:33357|issue:33453", + "issue:44961|issue:45468", + "issue:31795|issue:38617", + "issue:41628|issue:42548", + "issue:43576|issue:44351", + "issue:44908|issue:44961", + "issue:36296|issue:43901", + "issue:44485|issue:44961", + "issue:44273|issue:45003", + "issue:36296|issue:43526", + "issue:36296|issue:43012", + "issue:41628|issue:44355" + ], + "prepared_review_unit_hash": "6c5616a16fd5287a8357d223f5e30adc7ba0b02decc7abcf7694edf7fe2df22f", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 36, + "serialized_chars": 37906, + "estimated_input_tokens": 9477, + "estimated_eval_tokens": 19210 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly heterogeneous: loading/import regressions, doc issues, model-specific bugs, and unrelated test/perf reports. None of the soft pairs look like the same underlying bug or change, so all soft edges are rejected.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "issue:43576 is the broadest v5-facing regression report and has external references, so it is the most reasonable representative if one issue must be picked.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44355", + "right": "issue:44749", + "accept": false, + "reason": "Different bugs: compiled-file import/runtime errors vs slowdown when filtering after version upgrade." + }, + { + "left": "issue:45125", + "right": "issue:45375", + "accept": false, + "reason": "Different model-parallel/config problems: missing _tp_plan vs a config field being dropped." + }, + { + "left": "issue:30990", + "right": "issue:31795", + "accept": false, + "reason": "Loading hang in Sentence Transformers vs documentation confusion in forward() are unrelated." + }, + { + "left": "issue:43502", + "right": "issue:44485", + "accept": false, + "reason": "local_files_only causing network calls is unrelated to GLM-5 RoPE implementation." + }, + { + "left": "issue:43010", + "right": "issue:44368", + "accept": false, + "reason": "Cache update no_grad decoration vs tie_word_embeddings warning are different code paths." + }, + { + "left": "issue:44075", + "right": "issue:44871", + "accept": false, + "reason": "SGD args ignored is unrelated to Gemma-3 eos_token_id configuration mismatch." + }, + { + "left": "issue:44315", + "right": "issue:44360", + "accept": false, + "reason": "Liger Kernel application issue vs DSA indexer ReLU discussion are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:44373", + "accept": false, + "reason": "Missing set_seed in tests vs wrong docstring for position_ids are not the same bug." + }, + { + "left": "issue:38175", + "right": "issue:43502", + "accept": false, + "reason": "SIGLIP2 zero probabilities vs unexpected API requests are unrelated." + }, + { + "left": "issue:43299", + "right": "issue:44556", + "accept": false, + "reason": "Both mention loading, but one is Qwen3VL MoE loading and the other is checkpoint reload compatibility after upgrade; not the same fix." + }, + { + "left": "issue:43502", + "right": "issue:44961", + "accept": false, + "reason": "local_files_only network leakage is unrelated to the 'racoon' issue." + }, + { + "left": "issue:44355", + "right": "issue:44655", + "accept": false, + "reason": "Compiled-file errors vs saving Pipeline objects are different failures." + }, + { + "left": "issue:44355", + "right": "issue:45071", + "accept": false, + "reason": "Compiled-file execution errors are unrelated to PretrainedConfig type-checking regressions." + }, + { + "left": "issue:44222", + "right": "issue:44230", + "accept": false, + "reason": "FP8 save_pretrained bug vs FP8 inference/support request are related only broadly to fp8, not the same concrete issue." + }, + { + "left": "issue:42831", + "right": "issue:43874", + "accept": false, + "reason": "FP8 accuracy issue vs missing image-processor method are different problems." + }, + { + "left": "issue:42890", + "right": "issue:43232", + "accept": false, + "reason": "Seeded test flakiness vs generation kwargs after sync_gpus are unrelated." + }, + { + "left": "issue:42757", + "right": "issue:43576", + "accept": false, + "reason": "Missing huggingface_hub import vs broken transformers env command are different failures." + }, + { + "left": "issue:33453", + "right": "issue:34689", + "accept": false, + "reason": "Tokenizer loading regression vs specific Llama 3.2 Vision model loading breakage are not the same bug." + }, + { + "left": "issue:33453", + "right": "issue:36296", + "accept": false, + "reason": "Tokenizer loading regression is unrelated to tensor-parallel training behavior." + }, + { + "left": "issue:43827", + "right": "issue:44961", + "accept": false, + "reason": "Docs still referencing pipeline() vs 'racoon' are unrelated." + }, + { + "left": "issue:43576", + "right": "issue:43824", + "accept": false, + "reason": "Broken v5 env command and a missing model import are both import/API regressions but not the same concrete issue." + }, + { + "left": "issue:43576", + "right": "issue:45071", + "accept": false, + "reason": "CLI/env breakage is unrelated to PretrainedConfig type-checking behavior." + }, + { + "left": "issue:43673", + "right": "issue:44492", + "accept": false, + "reason": "Missing GenerationMixin cache in chunked_prefill vs a typo in cache strategies are different issues." + }, + { + "left": "issue:38617", + "right": "issue:43576", + "accept": false, + "reason": "ImportError for layer_type_validation is unrelated to the env command regression." + }, + { + "left": "issue:33357", + "right": "issue:33453", + "accept": false, + "reason": "MacOS bus error with a community CLIP model vs tokenizer loading regression are unrelated." + }, + { + "left": "issue:44961", + "right": "issue:45468", + "accept": false, + "reason": "'racoon' provides no evidence of being the same as the Gemma-4 audio positional encoding bug." + }, + { + "left": "issue:31795", + "right": "issue:38617", + "accept": false, + "reason": "Doc confusion in model.forward vs missing layer_type_validation import are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:42548", + "accept": false, + "reason": "Different missing exports: AutoImageProcessor vs PreTrainedModel." + }, + { + "left": "issue:43576", + "right": "issue:44351", + "accept": false, + "reason": "Broken env command and missing HybridCache import are separate API/export problems." + }, + { + "left": "issue:44908", + "right": "issue:44961", + "accept": false, + "reason": "Scheduler kwargs bug vs 'racoon' are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43901", + "accept": false, + "reason": "Tensor-parallel training bug vs TextClassificationPipeline docs are unrelated." + }, + { + "left": "issue:44485", + "right": "issue:44961", + "accept": false, + "reason": "GLM-5 RoPE implementation issue vs 'racoon' are unrelated." + }, + { + "left": "issue:44273", + "right": "issue:45003", + "accept": false, + "reason": "Lazy loading malfunction and unsafe sys.modules access may be adjacent, but there is no evidence they are the same concrete bug." + }, + { + "left": "issue:36296", + "right": "issue:43526", + "accept": false, + "reason": "Tensor-parallel training bug vs BeitImageProcessorFast reduce_labels bug are unrelated." + }, + { + "left": "issue:36296", + "right": "issue:43012", + "accept": false, + "reason": "Tensor-parallel training bug vs bfloat16 compilation warning are unrelated." + }, + { + "left": "issue:41628", + "right": "issue:44355", + "accept": false, + "reason": "Missing AutoImageProcessor import vs compiled-file execution errors are different problems." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 40, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:31795", + "issue:33453", + "issue:34689", + "issue:37428", + "issue:41669", + "issue:42371", + "issue:42890", + "issue:43352", + "issue:43408", + "issue:43502", + "issue:43519", + "issue:43576", + "issue:43606", + "issue:43673", + "issue:43704", + "issue:43723", + "issue:43827", + "issue:43976", + "issue:44038", + "issue:44077", + "issue:44246", + "issue:44261", + "issue:44263", + "issue:44273", + "issue:44295", + "issue:44297", + "issue:44315", + "issue:44351", + "issue:44360", + "issue:44393", + "issue:44485", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44655", + "issue:44704", + "issue:44829", + "issue:44861", + "issue:44908", + "issue:44945", + "issue:45071", + "issue:45092", + "issue:45125", + "issue:45230", + "issue:45310", + "issue:45341", + "issue:45468", + "issue:45478" + ], + "soft_pairs": [ + "issue:44261|issue:44360", + "issue:44623|issue:45230", + "issue:43576|issue:44556", + "issue:41669|issue:43673", + "issue:44393|issue:45071", + "issue:44038|issue:45125", + "issue:44556|issue:45092", + "issue:42371|issue:44492", + "issue:42890|issue:44908", + "issue:44704|issue:44908", + "issue:44246|issue:44273", + "issue:43408|issue:44315", + "issue:44297|issue:45230", + "issue:43576|issue:44908", + "issue:43352|issue:44393", + "issue:44908|issue:45071", + "issue:43976|issue:44393", + "issue:44295|issue:45230", + "issue:41669|issue:44351", + "issue:41669|issue:45341", + "issue:44492|issue:44945", + "issue:41669|issue:44492", + "issue:44315|issue:44829", + "issue:44393|issue:45478", + "issue:44077|issue:44861", + "issue:31795|issue:33453", + "issue:43827|issue:44908", + "issue:41669|issue:44908", + "issue:43723|issue:44393", + "issue:44655|issue:44861", + "issue:43576|issue:44273", + "issue:44360|issue:45468", + "issue:45230|issue:45310", + "issue:44273|issue:44908", + "issue:41669|issue:43704", + "issue:43519|issue:44485", + "issue:43519|issue:44908", + "issue:31795|issue:34689", + "issue:31795|issue:37428", + "issue:44315|issue:45092", + "issue:44273|issue:45071", + "issue:43704|issue:44485", + "issue:44485|issue:45468", + "issue:42371|issue:43704", + "issue:43502|issue:43519", + "issue:43519|issue:44623", + "issue:43519|issue:43704", + "issue:44263|issue:45468", + "issue:44261|issue:45468", + "issue:43606|issue:45468", + "issue:42890|issue:43704" + ], + "prepared_review_unit_hash": "507882254f35c684cb25de9cb34cd424665c6dd06c48c5b01b8cc02315062963", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 48, + "item_count": 48, + "soft_pair_count": 51, + "serialized_chars": 42381, + "estimated_input_tokens": 10596, + "estimated_eval_tokens": 21448 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a false-positive similarity group: the items share broad themes like v5 regressions, loading, imports, tokenizer/processor APIs, or model-family names, but they describe different bugs and change requests. No soft edge looks like the same underlying issue, so all candidate merges should be rejected.", + "confidence": 0.88, + "canonical_issue_reason": "Issue 44556 is the closest representative of the cluster because it is a concrete, version-bounded v5 regression with a clear reload failure and actionable reproduction. The rest of the items are too heterogeneous to share one true canonical issue.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 44556 is the strongest standalone issue here: specific symptom, explicit upgrade boundary, and a reproducible compatibility regression. It is more actionable than the broader or more ambiguous reports.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44261", + "right": "issue:44360", + "accept": false, + "reason": "Different bugs: one is a precision issue from missing rms_norm_eps, the other is a ReLU omission in a DSA indexer." + }, + { + "left": "issue:44623", + "right": "issue:45230", + "accept": false, + "reason": "One is a specific save_pretrained file omission; the other is an uninformative generic bug report with no concrete matching symptom." + }, + { + "left": "issue:43576", + "right": "issue:44556", + "accept": false, + "reason": "Different failure modes: broken transformers env command versus checkpoint reload incompatibility after upgrading to v5." + }, + { + "left": "issue:41669", + "right": "issue:43673", + "accept": false, + "reason": "Import-star performance regressions and missing GenerationMixin cache are unrelated code paths." + }, + { + "left": "issue:44393", + "right": "issue:45071", + "accept": false, + "reason": "Qwen3-VL bbox hallucination is unrelated to a PretrainedConfig type-checking break in v5.4.0." + }, + { + "left": "issue:44038", + "right": "issue:45125", + "accept": false, + "reason": "Different model-specific issues: Qwen3-VL-Moe versus missing _tp_plan for Qwen3_5Moe tensor parallelism." + }, + { + "left": "issue:44556", + "right": "issue:45092", + "accept": false, + "reason": "Both concern v5 loading compatibility, but they affect different checkpoint types and stem from different initialization paths." + }, + { + "left": "issue:42371", + "right": "issue:44492", + "accept": false, + "reason": "TF32 behavior settings and a cache-strategy typo are unrelated." + }, + { + "left": "issue:42890", + "right": "issue:44908", + "accept": false, + "reason": "Test seeding failures in sam_hq do not match an inverse_sqrt scheduler kwargs bug." + }, + { + "left": "issue:44704", + "right": "issue:44908", + "accept": false, + "reason": "AutoProcessor kwarg forwarding to cached_file is unrelated to scheduler argument handling." + }, + { + "left": "issue:44246", + "right": "issue:44273", + "accept": false, + "reason": "Import-time slowness and lazy-loading malfunction are adjacent but not the same bug." + }, + { + "left": "issue:43408", + "right": "issue:44315", + "accept": false, + "reason": "sam3_video/sam3_tracker mismatch is unrelated to Liger Kernel not being applied via model_init." + }, + { + "left": "issue:44297", + "right": "issue:45230", + "accept": false, + "reason": "Tokenizer_class mismatch on save_pretrained is not the same as a vague generic bug report." + }, + { + "left": "issue:43576", + "right": "issue:44908", + "accept": false, + "reason": "A broken env command and scheduler kwargs being ignored are unrelated regressions." + }, + { + "left": "issue:43352", + "right": "issue:44393", + "accept": false, + "reason": "Flash Attention 2 support error for NemotronH is unrelated to Qwen3-VL bbox output errors." + }, + { + "left": "issue:44908", + "right": "issue:45071", + "accept": false, + "reason": "Scheduler kwargs handling and PretrainedConfig type checking are different subsystems and bugs." + }, + { + "left": "issue:43976", + "right": "issue:44393", + "accept": false, + "reason": "Python version compatibility and Qwen3-VL output errors have no concrete overlap." + }, + { + "left": "issue:44295", + "right": "issue:45230", + "accept": false, + "reason": "Reading position_ids after buffer registration is unrelated to a generic bug report." + }, + { + "left": "issue:41669", + "right": "issue:44351", + "accept": false, + "reason": "Import-star cleanup and HybridCache import failure are different import-related problems." + }, + { + "left": "issue:41669", + "right": "issue:45341", + "accept": false, + "reason": "Import performance work and a testing_utils bug do not describe the same defect." + }, + { + "left": "issue:44492", + "right": "issue:44945", + "accept": false, + "reason": "A cache-strategy typo is unrelated to incorrect LLM output under pipeline parallelism." + }, + { + "left": "issue:41669", + "right": "issue:44492", + "accept": false, + "reason": "Removing import * and a typo in cache strategies are not the same issue." + }, + { + "left": "issue:44315", + "right": "issue:44829", + "accept": false, + "reason": "Liger Kernel application during model_init is unrelated to flash_attention_3 training collapse." + }, + { + "left": "issue:44393", + "right": "issue:45478", + "accept": false, + "reason": "Different model families and symptoms: Qwen3-VL bbox errors versus Qwen3.5 Moe from_pretrained failure." + }, + { + "left": "issue:44077", + "right": "issue:44861", + "accept": false, + "reason": "post_init policy for patchtsmixer and _get_tied_weight_keys list handling are different code paths." + }, + { + "left": "issue:31795", + "right": "issue:33453", + "accept": false, + "reason": "Documentation confusion around forward args is unrelated to a tokenizer loading regression." + }, + { + "left": "issue:43827", + "right": "issue:44908", + "accept": false, + "reason": "Docs still mentioning pipeline() and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:44908", + "accept": false, + "reason": "Import-star usage and inverse_sqrt scheduler kwargs are unrelated regressions." + }, + { + "left": "issue:43723", + "right": "issue:44393", + "accept": false, + "reason": "Tokenizer loading in v5 is unrelated to Qwen3-VL bounding-box output issues." + }, + { + "left": "issue:44655", + "right": "issue:44861", + "accept": false, + "reason": "Saving Pipeline objects and tied-weights-key handling are different APIs with different failures." + }, + { + "left": "issue:43576", + "right": "issue:44273", + "accept": false, + "reason": "Broken env command and lazy loading malfunction are both v5-related but not the same concrete bug." + }, + { + "left": "issue:44360", + "right": "issue:45468", + "accept": false, + "reason": "DSA indexer ReLU omission is unrelated to Gemma4AudioRelPositionalEncoding." + }, + { + "left": "issue:45230", + "right": "issue:45310", + "accept": false, + "reason": "A placeholder bug report does not establish the same root cause as the Qwen3.5 Moe from_pretrained error." + }, + { + "left": "issue:44273", + "right": "issue:44908", + "accept": false, + "reason": "Lazy loading problems and scheduler kwargs handling are unrelated." + }, + { + "left": "issue:41669", + "right": "issue:43704", + "accept": false, + "reason": "Import-time cleanup and VRAM leakage in multithreaded loading are different bugs." + }, + { + "left": "issue:43519", + "right": "issue:44485", + "accept": false, + "reason": "Qwen3VL timestamp calculation and GLM-5 RoPE implementation are unrelated model/processor issues." + }, + { + "left": "issue:43519", + "right": "issue:44908", + "accept": false, + "reason": "Timestamp math in a processor is unrelated to scheduler configuration." + }, + { + "left": "issue:31795", + "right": "issue:34689", + "accept": false, + "reason": "Forward-arg documentation confusion is unrelated to a Llama 3.2 model loading regression." + }, + { + "left": "issue:31795", + "right": "issue:37428", + "accept": false, + "reason": "Documentation wording and an ImportError for flash attention support are different classes of issue." + }, + { + "left": "issue:44315", + "right": "issue:45092", + "accept": false, + "reason": "Model_init not applying Liger Kernel is not the same as old remote-code checkpoints failing under meta initialization." + }, + { + "left": "issue:44273", + "right": "issue:45071", + "accept": false, + "reason": "Lazy loading malfunction and PretrainedConfig type checking are unrelated." + }, + { + "left": "issue:43704", + "right": "issue:44485", + "accept": false, + "reason": "VRAM leakage in dataloader threads is unrelated to GLM-5 RoPE math." + }, + { + "left": "issue:44485", + "right": "issue:45468", + "accept": false, + "reason": "GLM-5 RoPE implementation and Gemma4AudioRelPositionalEncoding are different model-specific bugs." + }, + { + "left": "issue:42371", + "right": "issue:43704", + "accept": false, + "reason": "TF32 API migration advice is unrelated to a VRAM leak in Qwen3ForCausalLM." + }, + { + "left": "issue:43502", + "right": "issue:43519", + "accept": false, + "reason": "local_files_only network requests and Qwen3VL timestamp errors are unrelated failures." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 2, + "component_count": 13, + "review_unit_index": 41, + "review_unit_count": 41, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-15354-362", + "nodes": [ + "issue:31515", + "issue:31795", + "issue:33453", + "issue:36296", + "issue:36683", + "issue:38617", + "issue:42913", + "issue:43066", + "issue:43298", + "issue:43525", + "issue:43572", + "issue:43618", + "issue:44230", + "issue:44492", + "issue:44556", + "issue:44623", + "issue:44861", + "issue:44908", + "issue:44998", + "issue:45290", + "issue:45310", + "issue:45341" + ], + "soft_pairs": [ + "issue:42913|issue:43066", + "issue:44908|issue:44998", + "issue:44861|issue:45290", + "issue:44998|issue:45341", + "issue:44492|issue:44998", + "issue:43618|issue:44998", + "issue:44556|issue:44998", + "issue:43525|issue:44861", + "issue:43572|issue:44861", + "issue:31515|issue:38617", + "issue:44230|issue:44623", + "issue:31515|issue:36296", + "issue:31515|issue:31795", + "issue:44230|issue:45310", + "issue:31515|issue:33453", + "issue:31515|issue:36683", + "issue:43298|issue:44230" + ], + "prepared_review_unit_hash": "0bfa4f0163b55a2ca1d55f16044d944c4291f1439bbf2e4d99123e9095f93e13", + "component_budget": { + "node_count": 362, + "item_count": 362, + "soft_pair_count": 1351, + "serialized_chars": 697234, + "estimated_input_tokens": 174309, + "estimated_eval_tokens": 348874 + }, + "budget": { + "node_count": 22, + "item_count": 22, + "soft_pair_count": 17, + "serialized_chars": 17367, + "estimated_input_tokens": 4342, + "estimated_eval_tokens": 8940 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "Mostly a loose thematically related set of loading/tokenizer/config regression issues, but the soft pairs do not describe the same underlying bug. No PRs are present.", + "confidence": 0.83, + "canonical_issue_reason": "issue:44556 is the clearest, most self-contained regression report in the set: checkpoint reload breaks after upgrading from v4.57 to v5.2/v5.3. It best represents the broad loading/version-compatibility theme, even though the rest are not true duplicates.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44556 is the strongest standalone triage candidate because it has a concrete versioned repro and a specific failure mode. The other issues are either more specialized, more vague, or clearly unrelated despite surface similarity.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:42913", + "right": "issue:43066", + "accept": false, + "reason": "Both mention tokenizer v5 behavior, but one is a general behavior change from v4 to v5 and the other is a specific wrong decoder type problem. Not the same bug." + }, + { + "left": "issue:44908", + "right": "issue:44998", + "accept": false, + "reason": "Completely unrelated; scheduler kwargs vs a nonsensical unrelated issue title. No shared concrete bug." + }, + { + "left": "issue:44861", + "right": "issue:45290", + "accept": false, + "reason": "Different code paths: tied-weight key extraction crash vs chat template handling of tool-call assistant messages." + }, + { + "left": "issue:44998", + "right": "issue:45341", + "accept": false, + "reason": "No meaningful overlap; unrelated generic issue versus a testing_utils bug." + }, + { + "left": "issue:44492", + "right": "issue:44998", + "accept": false, + "reason": "A typo in cache strategies is not the same as the other issue; no shared underlying defect." + }, + { + "left": "issue:43618", + "right": "issue:44998", + "accept": false, + "reason": "CLIP attentions assignment regression is unrelated to the other issue." + }, + { + "left": "issue:44556", + "right": "issue:44998", + "accept": false, + "reason": "Checkpoint reload/version upgrade regression does not match the unrelated issue." + }, + { + "left": "issue:43525", + "right": "issue:44861", + "accept": false, + "reason": "Different model/config attribute errors in different subsystems; not the same bug." + }, + { + "left": "issue:43572", + "right": "issue:44861", + "accept": false, + "reason": "StableLmConfig pad_token_idx regression and tied-weights crash are different failures." + }, + { + "left": "issue:31515", + "right": "issue:38617", + "accept": false, + "reason": "Checkpoint loading slowness and an ImportError for layer_type_validation are unrelated." + }, + { + "left": "issue:44230", + "right": "issue:44623", + "accept": false, + "reason": "fp8 inference support for a vision-language model is not the same as missing files from processor.save_pretrained." + }, + { + "left": "issue:31515", + "right": "issue:36296", + "accept": false, + "reason": "from_pretrained load performance and tensor-parallel training are different problems." + }, + { + "left": "issue:31515", + "right": "issue:31795", + "accept": false, + "reason": "Loading slow checkpoint and documentation confusion in model.forward are not the same underlying issue." + }, + { + "left": "issue:44230", + "right": "issue:45310", + "accept": false, + "reason": "Both involve Qwen-family models, but one is fp8 support and the other is a from_pretrained error; not the same concrete bug." + }, + { + "left": "issue:31515", + "right": "issue:33453", + "accept": false, + "reason": "Checkpoint loading slowness and tokenizer loading regression are different failure modes." + }, + { + "left": "issue:31515", + "right": "issue:36683", + "accept": false, + "reason": "from_pretrained loading slowdown is unrelated to Gemma3Config missing vocab_size." + }, + { + "left": "issue:43298", + "right": "issue:44230", + "accept": false, + "reason": "Continuous batching for audio models and fp8 Qwen VL inference support are distinct issues with different code paths." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 3, + "component_count": 13, + "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": "f9a3a55792236d149b1577ae1deb073396836987c469fd94f49585aee6d8da0e", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1590, + "estimated_input_tokens": 398, + "estimated_eval_tokens": 1052 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two distinct feature-request issues about adding support for more architectures, but they concern different features (Flash Attention 2 vs GGUF) and should not be deduplicated.", + "confidence": 0.98, + "canonical_issue_reason": "issue:26350 is the more established thread (older and much higher discussion activity), but it is not a duplicate of issue:33260.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:26350 is the better representative of the pair because it has substantially more discussion and visibility.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:26350", + "right": "issue:33260", + "accept": false, + "reason": "They share only a generic \"support for more architectures\" theme; one is about Flash Attention 2 and the other about GGUF, so they are different feature requests and 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": 4, + "component_count": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-27088-4", + "nodes": [ + "issue:27088", + "issue:43834", + "issue:44365", + "issue:44370" + ], + "soft_pairs": [ + "issue:27088|issue:44370", + "issue:27088|issue:44365", + "issue:27088|issue:43834", + "issue:43834|issue:44370", + "issue:43834|issue:44365", + "issue:44365|issue:44370" + ], + "prepared_review_unit_hash": "29e19d4a4c3511a0f79f4ad09609f48e8b520fa651a382a95a5096271f83b0ab", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4454, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4454, + "estimated_input_tokens": 1114, + "estimated_eval_tokens": 2484 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are all documentation translation request issues, but they are not the same underlying task: #27088 is specifically for Turkish, while the others are generic template issues for unspecified languages. The high text similarity appears to come from the shared issue template, not duplicate content.", + "confidence": 0.96, + "canonical_issue_reason": "#27088 is the only concrete, language-specific issue (Turkish). The other issues are generic placeholders and do not describe the same exact request.", + "canonical_pr_reason": null, + "best_issue_reason": "#27088 is the best representative issue because it is specific and actionable, whereas the others are broad template entries without a concrete target language.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:27088", + "right": "issue:44370", + "accept": false, + "reason": "Same translation-request template, but #27088 is Turkish and #44370 is unspecified; different underlying requests." + }, + { + "left": "issue:27088", + "right": "issue:44365", + "accept": false, + "reason": "Shared template wording only; one is a Turkish docs translation request and the other is a generic language placeholder." + }, + { + "left": "issue:27088", + "right": "issue:43834", + "accept": false, + "reason": "Both are translation issues, but they target different languages / one is generic, so they are not duplicates." + }, + { + "left": "issue:43834", + "right": "issue:44370", + "accept": false, + "reason": "Near-identical placeholders, but they do not identify the same concrete language task." + }, + { + "left": "issue:43834", + "right": "issue:44365", + "accept": false, + "reason": "Template similarity only; no evidence they refer to the same specific translation work." + }, + { + "left": "issue:44365", + "right": "issue:44370", + "accept": false, + "reason": "Both are generic translation placeholders, but there is no concrete shared language target to establish duplication." + } + ] + }, + "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": 13, + "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-36817-13", + "nodes": [ + "issue:36817", + "issue:37015", + "issue:41211", + "issue:42503", + "issue:42738", + "issue:43671", + "issue:43837", + "issue:43909", + "issue:44121", + "issue:44205", + "issue:45027", + "issue:45177", + "issue:45322" + ], + "soft_pairs": [ + "issue:36817|issue:41211", + "issue:36817|issue:43671", + "issue:37015|issue:43837", + "issue:37015|issue:44121", + "issue:37015|issue:43909", + "issue:37015|issue:42503", + "issue:37015|issue:45027", + "issue:41211|issue:43671", + "issue:42503|issue:43837", + "issue:42503|issue:43671", + "issue:42503|issue:42738", + "issue:42503|issue:45027", + "issue:42738|issue:45027", + "issue:43671|issue:43837", + "issue:43671|issue:45027", + "issue:43671|issue:45177", + "issue:43671|issue:44121", + "issue:43671|issue:43909", + "issue:43837|issue:43909", + "issue:43837|issue:45322", + "issue:43837|issue:44121", + "issue:43909|issue:44205", + "issue:44121|issue:45322", + "issue:43837|issue:45027", + "issue:43837|issue:45177" + ], + "prepared_review_unit_hash": "f71a878ee9e7b8b02ded57ea5d61c29897926f71e62ca549d747943458c1407a", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 25, + "serialized_chars": 16010, + "estimated_input_tokens": 4003, + "estimated_eval_tokens": 8262 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 25, + "serialized_chars": 16010, + "estimated_input_tokens": 4003, + "estimated_eval_tokens": 8262 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These issues are all broad model-addition requests, but they target different models and capabilities (BERT variants, audio/TTS/ASR, OCR, sparse transformers, etc.). They share a template-level similarity, not the same underlying change, so none should be deduplicated or merged into one canonical issue.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:36817", + "right": "issue:41211", + "accept": false, + "reason": "EuroBert vs DEIMv2 are different model requests; same broad \u201cadd model\u201d theme only." + }, + { + "left": "issue:36817", + "right": "issue:43671", + "accept": false, + "reason": "EuroBert config addition and Qwen3-TTS support are unrelated model types and code paths." + }, + { + "left": "issue:37015", + "right": "issue:43837", + "accept": false, + "reason": "NeoBERT and Qwen3-ASR are distinct models; not the same feature or fix." + }, + { + "left": "issue:37015", + "right": "issue:44121", + "accept": false, + "reason": "NeoBERT and weight-sparse transformer support are different architectures and implementation tasks." + }, + { + "left": "issue:37015", + "right": "issue:43909", + "accept": false, + "reason": "NeoBERT vs LFM2.5 Audio is unrelated model support, despite similar request wording." + }, + { + "left": "issue:37015", + "right": "issue:42503", + "accept": false, + "reason": "NeoBERT and ModernVBERT are different model additions, not one duplicate request." + }, + { + "left": "issue:37015", + "right": "issue:45027", + "accept": false, + "reason": "NeoBERT and Voxtral TTS support target different modalities and code paths." + }, + { + "left": "issue:41211", + "right": "issue:43671", + "accept": false, + "reason": "DEIMv2 and Qwen3-TTS are unrelated model requests; only the request template is similar." + }, + { + "left": "issue:42503", + "right": "issue:43837", + "accept": false, + "reason": "ModernVBERT models and Qwen3-ASR support are different features and model families." + }, + { + "left": "issue:42503", + "right": "issue:43671", + "accept": false, + "reason": "ModernVBERT and Qwen3-TTS are separate model-support requests with no shared concrete change." + }, + { + "left": "issue:42503", + "right": "issue:42738", + "accept": false, + "reason": "ModernVBERT and BERT-like models with RoPE are related only at a high level; not the same requested addition." + }, + { + "left": "issue:42503", + "right": "issue:45027", + "accept": false, + "reason": "ModernVBERT and Voxtral TTS support are different model families and implementation areas." + }, + { + "left": "issue:42738", + "right": "issue:45027", + "accept": false, + "reason": "BERT-like RoPE support and Voxtral TTS support are not the same underlying change." + }, + { + "left": "issue:43671", + "right": "issue:43837", + "accept": false, + "reason": "Qwen3-TTS and Qwen3-ASR are both Qwen3 requests, but they are different modalities and separate support work." + }, + { + "left": "issue:43671", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-TTS support and LFM2.5 Audio support are different models, not duplicates." + }, + { + "left": "issue:43671", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-TTS support and weight-sparse transformer support are unrelated changes." + }, + { + "left": "issue:43671", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-TTS and Voxtral TTS are distinct model integrations with different code paths." + }, + { + "left": "issue:43671", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-TTS and DeepSeek-OCR2 are unrelated model additions." + }, + { + "left": "issue:43837", + "right": "issue:43909", + "accept": false, + "reason": "Qwen3-ASR and LFM2.5 Audio are separate audio-model requests, not one concrete fix." + }, + { + "left": "issue:43837", + "right": "issue:45322", + "accept": false, + "reason": "Qwen3-ASR and EUPE are different model families and different implementation targets." + }, + { + "left": "issue:43837", + "right": "issue:44121", + "accept": false, + "reason": "Qwen3-ASR support and weight-sparse transformer support do not describe the same change." + }, + { + "left": "issue:43909", + "right": "issue:44205", + "accept": false, + "reason": "LFM2.5 Audio and SAM3-LiteText are unrelated model requests across different modalities." + }, + { + "left": "issue:44121", + "right": "issue:45322", + "accept": false, + "reason": "Weight-sparse transformer support and EUPE addition are different model-architecture requests." + }, + { + "left": "issue:43837", + "right": "issue:45027", + "accept": false, + "reason": "Qwen3-ASR and Voxtral TTS are separate speech-model support requests." + }, + { + "left": "issue:43837", + "right": "issue:45177", + "accept": false, + "reason": "Qwen3-ASR and DeepSeek-OCR2 are unrelated model additions." + } + ] + }, + "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": 13, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-37934-6", + "nodes": [ + "issue:37934", + "issue:42936", + "issue:43240", + "issue:43668", + "issue:43813", + "issue:44031" + ], + "soft_pairs": [ + "issue:37934|issue:43668", + "issue:42936|issue:43240", + "issue:42936|issue:44031", + "issue:42936|issue:43813", + "issue:42936|issue:43668", + "issue:43240|issue:43668", + "issue:43668|issue:44031", + "issue:43668|issue:43813" + ], + "prepared_review_unit_hash": "f1e950ab26a8e3227ccedd5060213ad47cbee1171dddcad42dcb1a1bab340343", + "component_budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6293, + "estimated_input_tokens": 1574, + "estimated_eval_tokens": 3404 + }, + "budget": { + "node_count": 6, + "item_count": 6, + "soft_pair_count": 8, + "serialized_chars": 6293, + "estimated_input_tokens": 1574, + "estimated_eval_tokens": 3404 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "No duplicate cluster here: these issues cover unrelated problems across model configs, tokenization warnings, loss wiring, and a typo. None of the soft-similarity links look like the same underlying bug.", + "confidence": 0.95, + "canonical_issue_reason": "No clear canonical issue; the reports are heterogeneous and do not describe the same defect or code path.", + "canonical_pr_reason": null, + "best_issue_reason": "No single best issue for deduping because this set does not form a duplicate cluster; each item appears to be a distinct bug or typo.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:37934", + "right": "issue:43668", + "accept": false, + "reason": "Different model areas and bug types: Llama4TextL2Norm behavior question vs ModernBERT config type hint." + }, + { + "left": "issue:42936", + "right": "issue:43240", + "accept": false, + "reason": "Mask2Former ignore_value wiring is unrelated to kwargs not being passed into a loss function." + }, + { + "left": "issue:42936", + "right": "issue:44031", + "accept": false, + "reason": "Mask2Former loss/ignore_value behavior is unrelated to tokenizer regex warnings." + }, + { + "left": "issue:42936", + "right": "issue:43813", + "accept": false, + "reason": "A model loss-parameter issue is not the same as a typo in a quantization attribute name." + }, + { + "left": "issue:42936", + "right": "issue:43668", + "accept": false, + "reason": "These are separate issues: one about ignore_value usage, the other about a config type annotation." + }, + { + "left": "issue:43240", + "right": "issue:43668", + "accept": false, + "reason": "Loss-function kwargs propagation and config type hints do not describe the same bug." + }, + { + "left": "issue:43668", + "right": "issue:44031", + "accept": false, + "reason": "Incorrect type hint in a config class is unrelated to tokenizer regex warnings after a version bump." + }, + { + "left": "issue:43668", + "right": "issue:43813", + "accept": false, + "reason": "Type-hint correction and a spelling error in a quantization field are distinct fixes." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 7, + "component_count": 13, + "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": "152c7c4502d4952876d0b1c12aedba74685aa45e5052ccf05fe1d4b9fcfa9072", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1480, + "estimated_input_tokens": 370, + "estimated_eval_tokens": 996 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two issues do not look like duplicates: one is a terse request about block decoding, the other is a specific request for DeepSeek v3.2 support. The only linkage is weak textual similarity, not a shared underlying problem.", + "confidence": 0.97, + "canonical_issue_reason": "No clear canonical issue: the issues describe different topics and do not appear to report the same bug or feature request.", + "canonical_pr_reason": null, + "best_issue_reason": "Issue 43037 is the better representative of the cluster only because it is more specific and actionable, but it is not a duplicate of 41084.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:41084", + "right": "issue:43037", + "accept": false, + "reason": "Different subjects: block decoding implementation vs DeepSeek v3.2 support. Similar wording is too weak to indicate 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": 8, + "component_count": 13, + "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": "4827160b2fc43dac15415836da2e6ae3728ad9186c7e56ce47b13be177f9d6f6", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1584, + "estimated_input_tokens": 396, + "estimated_eval_tokens": 1048 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related only in the broad sense that both mention Trainer customization, but they describe different problems: one is about `remove_unused_column` behavior, the other about FLOPS estimation for custom trainers. I would not cluster them as duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": "Neither issue is a duplicate of the other; if forced to pick one representative, #43598 is the clearer, more specific customization request around `remove_unused_column`.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43598", + "right": "issue:43600", + "accept": false, + "reason": "Different underlying bugs/features: #43598 is about `remove_unused_column` in Trainer, while #43600 is about estimating FLOPS for custom trainers. Shared Trainer context is too broad to treat as the same issue." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 9, + "component_count": 13, + "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": "3ed92323afdf1c8dd0bd0747c10a07d0c6fe83f048fb3157bdc7b41bca4e84ab", + "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 issues are about different Qwen model capability gaps: one is model registration for Qwen3-Omni, the other is sequence classification support for Qwen3.5. They are not the same underlying bug/change, so the soft duplicate link should be rejected.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43684", + "right": "issue:44405", + "accept": false, + "reason": "Different model families and different requested functionality (AutoModel registration vs AutoModelForSequenceClassification support); not mergeable as one duplicate fix." + } + ] + }, + "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": 13, + "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": "af74e354797a6e25483db37f945d0059fa1d72af2f9f9e138305c8d4b010815c", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 4264, + "estimated_input_tokens": 1066, + "estimated_eval_tokens": 2388 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The three \"huggingface \" issues form a weak duplicate-looking cluster, but the evidence is only soft similarity with no deterministic accept flags. \"Fett\" still looks separate. Keep the cluster representative as 43845 if needed, but do not treat the pairwise links as confirmed duplicates.", + "confidence": 0.73, + "canonical_issue_reason": "issue:43845 is the earliest of the near-identical low-information \"huggingface \" reports, so it is the best representative for that tentative cluster, though the linkage evidence is weak.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:43845 is the strongest global issue representative because it is the earliest of the repeated low-context reports and fits the tentative duplicate set best.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:43835", + "right": "issue:43845", + "accept": false, + "reason": "Only soft similarity and different titles; not enough evidence they are the same underlying issue." + }, + { + "left": "issue:43835", + "right": "issue:43846", + "accept": false, + "reason": "Only soft similarity with no shared concrete details or target; keep separate." + }, + { + "left": "issue:43835", + "right": "issue:43859", + "accept": false, + "reason": "Different titles and no substantive overlapping evidence of the same bug." + }, + { + "left": "issue:43845", + "right": "issue:43846", + "accept": false, + "reason": "They look similar, but the packet only provides soft similarity and deterministic_accept is false, so this should stay tentative at most." + }, + { + "left": "issue:43845", + "right": "issue:43859", + "accept": false, + "reason": "High textual similarity, but no concrete evidence beyond soft similarity; do not accept as a confirmed duplicate pair." + }, + { + "left": "issue:43846", + "right": "issue:43859", + "accept": false, + "reason": "Indistinguishable low-information titles, but still only soft similarity with no deterministic acceptance; reject as unconfirmed." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The cluster summary is mostly grounded, but the soft-edge verdicts are too strong. All pairwise links are supported only by soft_similarity, and the packet marks every pair as deterministic_accept=false, so accepting issue:43845 with issue:43846 and issue:43859 overstates the evidence. The conservative choice here would be to keep those pairs rejected or at most tentative, not accepted." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 11, + "component_count": 13, + "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": "aa2ee6e0cc0fc5d4a7fa47dca1935fec1664025e226a1d2469f7387c5bce3958", + "component_budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 + }, + "budget": { + "node_count": 5, + "item_count": 5, + "soft_pair_count": 9, + "serialized_chars": 5788, + "estimated_input_tokens": 1447, + "estimated_eval_tokens": 3150 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All items are low-signal and appear unrelated spam or a generic documentation request. No soft pair has enough evidence to be treated as the same underlying bug or change.", + "confidence": 0.23, + "canonical_issue_reason": "issue:44418 is the only item with a non-placeholder, somewhat substantive title; the others are clearly spam-like and provide no usable bug/change signal.", + "canonical_pr_reason": null, + "best_issue_reason": "issue:44418 is the best available issue for triage because it is the most descriptive and least obviously spammy, even though it still lacks concrete technical detail.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:44371", + "right": "issue:44572", + "accept": false, + "reason": "Both are generic '' issues with no concrete bug, code path, or change described; identical placeholder titles are not enough to infer the same underlying problem." + }, + { + "left": "issue:44371", + "right": "issue:44573", + "accept": false, + "reason": "Both look like separate spam submissions rather than the same bug or change; there is no shared technical subject beyond the placeholder title." + }, + { + "left": "issue:44371", + "right": "issue:44418", + "accept": false, + "reason": "A spam placeholder issue and a documentation enhancement request are not the same underlying problem." + }, + { + "left": "issue:44371", + "right": "issue:44937", + "accept": false, + "reason": "Generic spam versus a separate spam-like promotional title; no evidence of one concrete issue or fix." + }, + { + "left": "issue:44418", + "right": "issue:44572", + "accept": false, + "reason": "Documentation enhancement suggestion and placeholder spam are unrelated." + }, + { + "left": "issue:44418", + "right": "issue:44573", + "accept": false, + "reason": "Documentation enhancement suggestion and placeholder spam are unrelated." + }, + { + "left": "issue:44572", + "right": "issue:44573", + "accept": false, + "reason": "Both are '' issues, but there is no concrete shared bug/change to justify treating them as duplicates." + }, + { + "left": "issue:44572", + "right": "issue:44937", + "accept": false, + "reason": "Placeholder spam and a promotional-looking title do not describe the same underlying defect or change." + }, + { + "left": "issue:44573", + "right": "issue:44937", + "accept": false, + "reason": "Placeholder spam and a promotional-looking title do not describe the same underlying defect or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 12, + "component_count": 13, + "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": "b7cc7e33e64627c0dafc580a2ec2edef1e41dd4c17ce47cc53a41d9cf21e186f", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 1646, + "estimated_input_tokens": 412, + "estimated_eval_tokens": 1080 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two closed issues have identical titles and a soft-similarity signal, so they are plausibly related and may be duplicates, but the available packet does not expose body content to confirm that they describe the same issue.", + "confidence": 0.79, + "canonical_issue_reason": "#45145 is the better canonical issue because it has an inbound reference, giving it slightly more linkage context than #45151.", + "canonical_pr_reason": null, + "best_issue_reason": "#45145 is the strongest representative of the issue cluster because it has the same title as #45151 and slightly more reference context.", + "best_pr_reason": null, + "soft_edge_verdicts": [ + { + "left": "issue:45145", + "right": "issue:45151", + "accept": false, + "reason": "Likely related/possible duplicate based on identical titles and soft similarity, but the packet does not show body content or other concrete evidence to confirm they are the same underlying issue." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The duplicate-issue inference is directionally plausible, but the analyst overstates the evidence. The packet only supports that the two closed issues have identical titles, similar body lengths, and a soft-similarity signal; it does not expose body content, so claims like \"strongly overlapping body/theme\" and \"same title/content\" are not fully grounded. Also, the soft pair was not deterministically accepted, so the verdict should be phrased more conservatively (e.g., \"likely related/possible duplicate\" rather than definitive duplicate). The confidence of 0.98 is too strong for the available evidence." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "issue", + "component_index": 13, + "component_count": 13, + "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": "pull_request", + "component_index": 1, + "component_count": 25, + "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:42668", + "pull_request:43251", + "pull_request:43254", + "pull_request:43424", + "pull_request:43823", + "pull_request:43995", + "pull_request:44827", + "pull_request:45078", + "pull_request:45170", + "pull_request:45213", + "pull_request:45435", + "pull_request:45454", + "pull_request:45489" + ], + "soft_pairs": [ + "pull_request:39895|pull_request:43424", + "pull_request:42668|pull_request:43823", + "pull_request:42668|pull_request:45078", + "pull_request:43251|pull_request:43823", + "pull_request:43251|pull_request:44827", + "pull_request:43424|pull_request:44827", + "pull_request:43424|pull_request:45435", + "pull_request:43424|pull_request:43823", + "pull_request:43823|pull_request:45078", + "pull_request:43823|pull_request:45454", + "pull_request:43823|pull_request:45435", + "pull_request:43823|pull_request:43995", + "pull_request:43995|pull_request:45078", + "pull_request:43995|pull_request:45435", + "pull_request:43995|pull_request:45213", + "pull_request:43995|pull_request:44827", + "pull_request:44827|pull_request:45489", + "pull_request:44827|pull_request:45170", + "pull_request:43823|pull_request:44827", + "pull_request:45078|pull_request:45454", + "pull_request:45078|pull_request:45435", + "pull_request:45213|pull_request:45435", + "pull_request:44827|pull_request:45213", + "pull_request:43823|pull_request:45213", + "pull_request:45078|pull_request:45213", + "pull_request:45454|pull_request:45489", + "pull_request:43823|pull_request:45489", + "pull_request:43251|pull_request:43254" + ], + "prepared_review_unit_hash": "e77ec4c3ee556115959682801a06929e14789ded98d08ec7fc9209aba713ea5e", + "component_budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 28, + "serialized_chars": 41462, + "estimated_input_tokens": 10366, + "estimated_eval_tokens": 20988 + }, + "budget": { + "node_count": 14, + "item_count": 14, + "soft_pair_count": 28, + "serialized_chars": 41462, + "estimated_input_tokens": 10366, + "estimated_eval_tokens": 20988 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Only one soft-edge pair is a true duplicate: 43251 and 43254 both patch `fixed_cross_entropy` in `src/transformers/loss/loss_utils.py` to forward supported `nn.functional.cross_entropy` kwargs. All other pairs are unrelated changes across different models, utilities, or tests; in particular, 43823 is a MobileLLM addition and should not be described as Gemma/Mistral-related.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43251 is the clearest representative of the duplicate loss fix: it directly patches `fixed_cross_entropy`, is explicitly tied to issue 43240, and makes the same concrete kwargs pass-through change as 43254 with minimal scope.", + "best_issue_reason": null, + "best_pr_reason": "PR 43251 is the best standalone representative in this cluster because it addresses the concrete cross-entropy kwargs bug directly, with an explicit issue link and the narrowest, most focused patch.", + "soft_edge_verdicts": [ + { + "left": "pull_request:39895", + "right": "pull_request:43424", + "accept": false, + "reason": "Different features in unrelated areas: VideoPrism model addition vs executorch dynamic-shape export testing." + }, + { + "left": "pull_request:42668", + "right": "pull_request:43823", + "accept": false, + "reason": "Different changes: processor-pretrained robustness work vs a new MobileLLM model addition." + }, + { + "left": "pull_request:42668", + "right": "pull_request:45078", + "accept": false, + "reason": "Both involve auto/processor code, but they fix different behaviors: processor-from-pretrained robustness vs tokenizer conversion handling." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43823", + "accept": false, + "reason": "Cross-entropy kwargs forwarding is unrelated to adding the MobileLLM model." + }, + { + "left": "pull_request:43251", + "right": "pull_request:44827", + "accept": false, + "reason": "Different code paths and purposes: loss utility kwargs vs Mistral4 test/model fixes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:44827", + "accept": false, + "reason": "Executorch export tests are unrelated to Mistral4 model/test changes." + }, + { + "left": "pull_request:43424", + "right": "pull_request:45435", + "accept": false, + "reason": "Different targets: executorch dynamic-shape export tests vs a Whisper tokenization bounds fix." + }, + { + "left": "pull_request:43424", + "right": "pull_request:43823", + "accept": false, + "reason": "Executorch export testing and MobileLLM model addition are unrelated changes." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45078", + "accept": false, + "reason": "MobileLLM model addition and tokenizer conversion error handling are different changes with no shared bug." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45454", + "accept": false, + "reason": "MobileLLM addition is unrelated to the Gemma4 text-only training fix; they touch different models and code paths." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45435", + "accept": false, + "reason": "MobileLLM model addition is unrelated to the Whisper special-token indexing fix." + }, + { + "left": "pull_request:43823", + "right": "pull_request:43995", + "accept": false, + "reason": "New model addition vs Falcon output-collection refactor are different artifacts." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45078", + "accept": false, + "reason": "Falcon output-interface refactoring is unrelated to tokenizer conversion logic." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45435", + "accept": false, + "reason": "Falcon refactoring and Whisper tokenization fixes address different code paths." + }, + { + "left": "pull_request:43995", + "right": "pull_request:45213", + "accept": false, + "reason": "Falcon model refactoring and the model-creation skill PR are unrelated." + }, + { + "left": "pull_request:43995", + "right": "pull_request:44827", + "accept": false, + "reason": "Falcon refactor is unrelated to the Mistral4 test/model fixes." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45489", + "accept": false, + "reason": "Both are model PRs, but they fix different concrete behaviors in different model families and are not mergeable as one change." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45170", + "accept": false, + "reason": "Mistral4 test/model changes are unrelated to the CLIP-like `layrnorm` renaming fix." + }, + { + "left": "pull_request:43823", + "right": "pull_request:44827", + "accept": false, + "reason": "MobileLLM addition and Mistral4 fixes are unrelated changes." + }, + { + "left": "pull_request:45078", + "right": "pull_request:45454", + "accept": false, + "reason": "Tokenizer conversion handling is unrelated to Gemma4 text-only training behavior." + }, + { + "left": "pull_request:45078", + "right": "pull_request:45435", + "accept": false, + "reason": "Different tokenizer issues: conversion-required handling vs special-token Unicode splitting bounds." + }, + { + "left": "pull_request:45213", + "right": "pull_request:45435", + "accept": false, + "reason": "A model-creation skill PR is unrelated to a Whisper tokenizer bug fix." + }, + { + "left": "pull_request:44827", + "right": "pull_request:45213", + "accept": false, + "reason": "Mistral4 fixes are unrelated to the model-creation skill scaffolding." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45213", + "accept": false, + "reason": "MobileLLM model addition and the model-creation skill are different artifacts, not the same change." + }, + { + "left": "pull_request:45078", + "right": "pull_request:45213", + "accept": false, + "reason": "Tokenizer behavior changes are unrelated to the skill/documentation PR." + }, + { + "left": "pull_request:45454", + "right": "pull_request:45489", + "accept": false, + "reason": "Gemma4 text-only training behavior and Gemma3n/Gemma4 cache-sharing alignment are separate concrete fixes." + }, + { + "left": "pull_request:43823", + "right": "pull_request:45489", + "accept": false, + "reason": "MobileLLM addition is unrelated to Gemma3n/Gemma4 cache-sharing changes." + }, + { + "left": "pull_request:43251", + "right": "pull_request:43254", + "accept": true, + "reason": "Same file and same underlying bug: both forward supported kwargs to `nn.functional.cross_entropy` inside `fixed_cross_entropy`." + } + ] + }, + "evaluator_result": { + "accept": false, + "feedback": "The 43251/43254 duplicate call is well grounded, but several soft-edge reasons are not conservative enough. In particular, PR 43823 is a MobileLLM addition, so descriptions like \"Gemma/Mistral-related code\" or \"Both touch Gemma-family models\" are factually off for pairs such as 43823|44827 and 43823|45454. Since the instructions require precise, conservative verdicts, I can\u2019t accept this output as-is." + }, + "evaluator_used": true, + "retried": true, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 2, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-40546-2", + "nodes": [ + "pull_request:40546", + "pull_request:42781" + ], + "soft_pairs": [ + "pull_request:40546|pull_request:42781" + ], + "prepared_review_unit_hash": "165162363feb23aa86a51b40e0ca400a1ea9cbb4de8adb5dce4eb9f8b182e2c9", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7225, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 7225, + "estimated_input_tokens": 1807, + "estimated_eval_tokens": 3870 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The two pull requests are related VibeVoice work, but they are not duplicates: 40546 is the base VibeVoice implementation, while 42781 adds a realtime variant plus the acoustic tokenizer and generation changes. They share many files, but they target different concrete capabilities.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "40546 is the foundational VibeVoice implementation and the broader, more complete base change set; 42781 is a draft extension for realtime support that depends on and expands the core model work.", + "best_issue_reason": null, + "best_pr_reason": "40546 is the best canonical PR for the underlying VibeVoice addition because it represents the main initial implementation rather than a specialized follow-on feature.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40546", + "right": "pull_request:42781", + "accept": false, + "reason": "Reject: 40546 implements core VibeVoice support, while 42781 adds realtime-specific functionality and an acoustic tokenizer. Shared filenames alone do not make them the same underlying change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 3, + "component_count": 25, + "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": "4d1fe8af026fa7dcb39a3c5cd04bddc45fe43e4ee16b2d67322e0d37441df2e2", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4675, + "estimated_input_tokens": 1169, + "estimated_eval_tokens": 2594 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These two pull requests are unrelated: one adds covariate-based forecasting support to TimesFM, while the other adds DINOv3 image classification support. The only overlap is generic model/docs/tests work, which is not enough to treat them as duplicates.", + "confidence": 0.98, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR: the PRs address different models and different code paths, so there is no single underlying change to collapse them into.", + "best_issue_reason": null, + "best_pr_reason": "Neither PR is a duplicate of the other; both are independent feature additions in separate subsystems, so neither is a better canonical representative for this cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:40755", + "right": "pull_request:41224", + "accept": false, + "reason": "Different models and different functionality: TimesFM covariate forecasting vs. DINOv3 image classification support. Shared docs/tests patterns do not indicate the same underlying bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 4, + "component_count": 25, + "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": "bf853edcac36b9d46727ab7091e6d4edc2525d28b28d3cc612346441fafe02a8", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5934, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both PRs are for adding DEIMv2 to Transformers. PR #44339 is a broader, more complete implementation that subsumes the earlier partial add in #41356, so they look like the same underlying feature work rather than separate changes.", + "confidence": 0.92, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #44339 is the more complete DEIMv2 implementation: it adds the model plus loss utilities, auto-class registration, conversion script, modular file, and config updates. It clearly subsumes the narrower earlier PR #41356.", + "best_issue_reason": null, + "best_pr_reason": "PR #44339 is the best representative because it covers the full DEIMv2 integration and is the more complete, maintainable landing point for this feature.", + "soft_edge_verdicts": [ + { + "left": "pull_request:41356", + "right": "pull_request:44339", + "accept": true, + "reason": "Same underlying change: both add DEIMv2 support. #44339 expands on the same model/files and looks like a superset of #41356 rather than a different fix." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: both PRs target the same issue (#41211), share several DEIMv2 files, and the later PR clearly adds more DEIMv2-related pieces. The soft-edge accept is conservative enough here, and the characterization as related feature work rather than separate changes is supported by the evidence." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 5, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-41797-2", + "nodes": [ + "pull_request:41797", + "pull_request:45186" + ], + "soft_pairs": [ + "pull_request:41797|pull_request:45186" + ], + "prepared_review_unit_hash": "6d04fc6904485368e707c98b27ea5ee42ca9f1d07b519dcfb9591d7a77958124", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6723, + "estimated_input_tokens": 1681, + "estimated_eval_tokens": 3618 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6723, + "estimated_input_tokens": 1681, + "estimated_eval_tokens": 3618 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate model-addition PRs with overlapping boilerplate and registry/docs files, but they implement different models and do not look mergeable as one 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:41797", + "right": "pull_request:45186", + "accept": false, + "reason": "Different underlying changes: one adds DeepSeek OCR, the other adds Isaac. Shared auto/docs plumbing is expected for new-model PRs, not evidence of the same fix." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 6, + "component_count": 25, + "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-42166-2", + "nodes": [ + "pull_request:42166", + "pull_request:45426" + ], + "soft_pairs": [ + "pull_request:42166|pull_request:45426" + ], + "prepared_review_unit_hash": "94e4688c750ab2029f402f9934224319a87ddbf6b98c0638d34a504db00e35b1", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5408, + "estimated_input_tokens": 1352, + "estimated_eval_tokens": 2960 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5408, + "estimated_input_tokens": 1352, + "estimated_eval_tokens": 2960 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "The cluster contains two unrelated feature PRs: one adds InternVLFlash, the other adds AXK1. The overlap is limited to shared auto-registration plumbing, which is not enough to treat them as the same change.", + "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:42166", + "right": "pull_request:45426", + "accept": false, + "reason": "Different model integrations with different filenames, configs, and tests; shared auto-registry edits are generic boilerplate, not the same bug or change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 7, + "component_count": 25, + "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-42227-4", + "nodes": [ + "pull_request:42227", + "pull_request:43291", + "pull_request:43665", + "pull_request:44815" + ], + "soft_pairs": [ + "pull_request:42227|pull_request:43291", + "pull_request:42227|pull_request:43665", + "pull_request:42227|pull_request:44815", + "pull_request:43291|pull_request:43665", + "pull_request:43291|pull_request:44815", + "pull_request:43665|pull_request:44815" + ], + "prepared_review_unit_hash": "60886efd49f4cdce90a921d77f190fb8abb077efb094f314bb3ddf20601769ca", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10168, + "estimated_input_tokens": 2542, + "estimated_eval_tokens": 5340 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 10168, + "estimated_input_tokens": 2542, + "estimated_eval_tokens": 5340 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are not true duplicates; they span unrelated Whisper pipeline/tokenizer work, model test tweaks, and an FP8 dequant/loading fix. The only plausible overlap is broad \u201cWhisper\u201d terminology, but the concrete code paths differ enough to reject merges.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR for duplicate-triage: the cluster is heterogeneous, but PR #42227 is the strongest standalone representative because it has a clear bug fix, explicit issue linkage, and a focused code-path change.", + "best_issue_reason": null, + "best_pr_reason": "PR #42227 is the best representative artifact overall: it addresses a concrete Whisper ASR bug, includes a targeted implementation change, and is tied to an explicit tracking issue.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42227", + "right": "pull_request:43291", + "accept": false, + "reason": "Both mention Whisper, but #42227 fixes ASR pipeline return_language forwarding while #43291 changes Whisper tokenizer/tests; different bugs and not plausibly one PR." + }, + { + "left": "pull_request:42227", + "right": "pull_request:43665", + "accept": false, + "reason": "Unrelated model test edits (CLIP/ViT) versus a Whisper pipeline fix; no shared concrete code-path problem." + }, + { + "left": "pull_request:42227", + "right": "pull_request:44815", + "accept": false, + "reason": "Whisper ASR language-return fix versus FP8 dequant/model-loading changes; completely different subsystems and failure modes." + }, + { + "left": "pull_request:43291", + "right": "pull_request:43665", + "accept": false, + "reason": "Whisper tokenizer/test cleanup and CLIP/ViT test edits are unrelated; same broad category of tests does not imply the same bug." + }, + { + "left": "pull_request:43291", + "right": "pull_request:44815", + "accept": false, + "reason": "Tokenizer/test changes for Whisper are unrelated to FP8 quantization/dequantization and loading logic." + }, + { + "left": "pull_request:43665", + "right": "pull_request:44815", + "accept": false, + "reason": "CLIP/ViT test modifications do not overlap with FP8 dequant/model-loading changes in any concrete fixable way." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 8, + "component_count": 25, + "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": "d3213d288c938d92f5c2515d30871f5451ff2f2c63e2b501c780e76ca8d8012d", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5935, + "estimated_input_tokens": 1484, + "estimated_eval_tokens": 3224 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are two separate workflow dependency upgrade PRs: one for GitHub Actions runtime compatibility, one for docker/build-push-action version bumps. They are related in theme but not the same change.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 42975 is the broader and more self-contained workflow action runtime upgrade, spanning many files and explicitly targeting Node 24 compatibility; PR 42976 is a narrower docker action version refresh.", + "best_issue_reason": null, + "best_pr_reason": "PR 42975 is the best standalone representative of the cluster because it has the clearer release goal and wider workflow impact, while still being a single coherent upgrade.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42975", + "right": "pull_request:42976", + "accept": false, + "reason": "Reject: they update different actions for different purposes (Node 24-compatible checkout/cache/upload-artifact vs docker/build-push-action v6). Similar maintenance theme, but not the same underlying bug or mergeable into one PR without conflating separate dependency bumps." + } + ] + }, + "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": 25, + "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": "2adba3260fe6ddb17db1c74a72dca8dfb0816ba0d9d68d54ef8c0e6bd06cd2ad", + "component_budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14136, + "estimated_input_tokens": 3534, + "estimated_eval_tokens": 7324 + }, + "budget": { + "node_count": 4, + "item_count": 4, + "soft_pair_count": 6, + "serialized_chars": 14136, + "estimated_input_tokens": 3534, + "estimated_eval_tokens": 7324 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are all model-addition work with overlapping framework scaffolding, but they target different models and different code paths: ViT NEPA, Molmo, Molmo2, and CTSM. Shared changes in docs/auto mappings are expected and do not indicate duplication.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No canonical PR here: the four pull requests are for distinct model implementations, not the same bug or change.", + "best_issue_reason": null, + "best_pr_reason": "PR 43451 (Molmo2) is the broadest and most developed of the set, with the largest surface area and most review activity, so it is the best representative PR for the cluster if one must be chosen.", + "soft_edge_verdicts": [ + { + "left": "pull_request:42978", + "right": "pull_request:43448", + "accept": false, + "reason": "Different model additions: ViT NEPA vs Molmo. Similar auto/doc scaffolding, but not the same concrete feature or fix." + }, + { + "left": "pull_request:42978", + "right": "pull_request:43451", + "accept": false, + "reason": "Different model additions: ViT NEPA vs Molmo2. Shared Transformers boilerplate does not make them duplicates." + }, + { + "left": "pull_request:42978", + "right": "pull_request:45490", + "accept": false, + "reason": "Different model additions: ViT NEPA vs CTSM. They touch the same framework areas but implement unrelated models." + }, + { + "left": "pull_request:43448", + "right": "pull_request:43451", + "accept": false, + "reason": "Molmo and Molmo2 are separate model families with separate implementations, configs, and docs; not one mergeable change." + }, + { + "left": "pull_request:43448", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo vs CTSM are unrelated model additions. Overlap is only in shared registration/documentation plumbing." + }, + { + "left": "pull_request:43451", + "right": "pull_request:45490", + "accept": false, + "reason": "Molmo2 and CTSM are distinct model implementations, so 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": 10, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43488-3", + "nodes": [ + "pull_request:43488", + "pull_request:43532", + "pull_request:45476" + ], + "soft_pairs": [ + "pull_request:43488|pull_request:43532", + "pull_request:43488|pull_request:45476", + "pull_request:43532|pull_request:45476" + ], + "prepared_review_unit_hash": "06fe7f1e2581430c87d02dcbb4fd86cffcbd1606e113c7cf246a675904577eea", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7468, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7468, + "estimated_input_tokens": 1867, + "estimated_eval_tokens": 3990 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three PRs are unrelated test/synthetic changes: one tweaks packaging/versioning, one rewires comment-driven CI workflows, and one adds a caller workflow plus unrelated tokenization/test edits. They only overlap superficially on a shared test file, not on the same bug or change.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single canonical PR stands out because the three PRs target different code paths and are not duplicates of one another; if forced, 43532 is the most substantial CI-related change, but it is still not representative of the others.", + "best_issue_reason": null, + "best_pr_reason": "43532 is the most coherent standalone change set, centered on CI/comment workflow behavior, but it should not be treated as a cluster canonical because it does not match the other PRs' scope.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43488", + "right": "pull_request:43532", + "accept": false, + "reason": "Shared vit test filename is incidental; 43488 is setup/version/dependency churn, while 43532 is workflow and notification plumbing. Different code paths and different intent." + }, + { + "left": "pull_request:43488", + "right": "pull_request:45476", + "accept": false, + "reason": "43488 changes packaging/version/deps, while 45476 adds a PR CI workflow and touches CLIP/chinese-CLIP test scaffolding. No same underlying bug or mergeable combined fix." + }, + { + "left": "pull_request:43532", + "right": "pull_request:45476", + "accept": false, + "reason": "Both mention CI/test infrastructure and a vit test file, but 43532 is about self-comment/self-scheduled workflow logic, whereas 45476 adds a new PR-CI caller plus unrelated tokenizer/test edits. 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": 11, + "component_count": 25, + "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": "6bdb893c182097484f72bc42915f4cff7c3528777ec9df7dad238d7b1c891b13", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 6082, + "estimated_input_tokens": 1521, + "estimated_eval_tokens": 3298 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are related but distinct feature PRs: one adds Promptable Concept Segmentation, the other Promptable Visual Segmentation. They share docs/pipeline registration scaffolding, but they do not fix the same code-path or implement the same change.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43613 is the better representative because it is the broader/more integrated segmentation pipeline addition, touching more model backends and pipeline plumbing.", + "best_issue_reason": null, + "best_pr_reason": "PR #43613 is the stronger standalone PR for this cluster because it has wider implementation coverage and is the more complete segmentation feature addition.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43612", + "right": "pull_request:43613", + "accept": false, + "reason": "Different features: concept segmentation vs visual segmentation. Similar docs and shared pipeline scaffolding are 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": 12, + "component_count": 25, + "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": "d834eba410e4169fbe9ffb39bf75ac9b94756881af61c350df06e0363c190f6c", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4291, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4291, + "estimated_input_tokens": 1073, + "estimated_eval_tokens": 2402 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two unrelated Trainer extensibility PRs were clustered together only because both touch `src/transformers/trainer.py`: one adds `_loss_is_scaled_for_ga` to control gradient-accumulation loss scaling, the other factors out signature-column detection into `_get_signature_columns` for column filtering. They address different code paths and are not duplicates.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43651 is the better representative of the cluster because it makes a concrete behavior change in the training step and is the more substantial, user-facing Trainer fix.", + "best_issue_reason": null, + "best_pr_reason": "PR 43651 is the strongest single PR to keep as the cluster representative; it has the clearest end-user impact and a more self-contained behavioral change than the refactoring-only helper added in 43663.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43651", + "right": "pull_request:43663", + "accept": false, + "reason": "Reject: these modify different Trainer mechanisms. 43651 changes gradient accumulation loss scaling in `training_step`, while 43663 introduces a helper for model signature/column filtering. Shared file alone is not enough to treat them as the same change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 13, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43656-3", + "nodes": [ + "pull_request:43656", + "pull_request:43836", + "pull_request:43842" + ], + "soft_pairs": [ + "pull_request:43656|pull_request:43842", + "pull_request:43836|pull_request:43842" + ], + "prepared_review_unit_hash": "f8042d67d07e88b89a8f2b2110cde9bd6f0ec3c6023b8d75a2bd591059b7bca8", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 5993, + "estimated_input_tokens": 1499, + "estimated_eval_tokens": 3254 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 5993, + "estimated_input_tokens": 1499, + "estimated_eval_tokens": 3254 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three pull requests target the same TypeAdapter import/type-annotation NameError in `src/transformers/cli/serve.py`, but they vary in specificity. The most targeted fix is PR #43842; the others are broader or earlier variants of the same underlying problem.", + "confidence": 0.88, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #43842 is the cleanest and most focused fix: it guards the `TypeAdapter` import under `TYPE_CHECKING` and quotes the runtime annotation, directly addressing the import-time NameError with minimal unrelated change.", + "best_issue_reason": null, + "best_pr_reason": "PR #43842 is the best representative because it precisely fixes the concrete failure mode (`TypeAdapter` NameError when pydantic is absent) with the smallest, most self-contained diff.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43656", + "right": "pull_request:43842", + "accept": true, + "reason": "Both address the same `TypeAdapter` NameError in the CLI serve path and appear to be alternative fixes for the same import-time annotation problem in the same file." + }, + { + "left": "pull_request:43836", + "right": "pull_request:43842", + "accept": true, + "reason": "Both are fixes for the same `TypeAdapter` NameError in `serve.py`; despite #43836 being broader, the core change is the same concrete annotation/import issue and could plausibly be merged into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: all three PRs are in `src/transformers/cli/serve.py` and each is clearly aimed at the same TypeAdapter-related import/type-annotation NameError, with PR #43842 being the most focused fix. The pairwise accept calls are also conservative enough given the shared file and bug pattern." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 14, + "component_count": 25, + "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-43757-3", + "nodes": [ + "pull_request:43757", + "pull_request:45500", + "pull_request:45506" + ], + "soft_pairs": [ + "pull_request:45500|pull_request:45506", + "pull_request:43757|pull_request:45500", + "pull_request:43757|pull_request:45506" + ], + "prepared_review_unit_hash": "4447f433e0b642137eee441f92a2d823a184fb435e4ee91eb8c6146e666e83ef", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7783, + "estimated_input_tokens": 1946, + "estimated_eval_tokens": 4148 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 7783, + "estimated_input_tokens": 1946, + "estimated_eval_tokens": 4148 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Two PRs are effectively the same GPT-OSS GGUF-loading fix, with 45506 the later, non-draft version; 43757 is only an earlier partial fallback and not the same full change.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45506 is the clearest canonical PR: it is the later finalized version of the GPT-OSS GGUF-loading work, includes the full support across loader/integration/tests/docs, and explicitly supersedes #43757.", + "best_issue_reason": null, + "best_pr_reason": "PR #45506 is the best representative of the cluster because it is the complete, non-draft implementation of GPT-OSS GGUF support and directly subsumes the earlier draft PR.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45500", + "right": "pull_request:45506", + "accept": true, + "reason": "Same concrete GPT-OSS GGUF loading change set: same target issue, same files, and the later PR is just the finalized version of the earlier draft." + }, + { + "left": "pull_request:43757", + "right": "pull_request:45500", + "accept": false, + "reason": "43757 only adds a fallback for the gpt-oss architecture in one loader path, while 45500 implements broader full GGUF loading support across loader, integration, docs, and tests." + }, + { + "left": "pull_request:43757", + "right": "pull_request:45506", + "accept": false, + "reason": "Same as above: 43757 is a partial workaround, not the same underlying full GGUF-loading fix as 45506." + } + ] + }, + "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": 25, + "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": 16, + "component_count": 25, + "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-43838-3", + "nodes": [ + "pull_request:43838", + "pull_request:43973", + "pull_request:45283" + ], + "soft_pairs": [ + "pull_request:43838|pull_request:45283", + "pull_request:43973|pull_request:45283" + ], + "prepared_review_unit_hash": "7105e8e9fca39b9892a98470abd8d783453f3812684244a1327f48556aa95486", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 8699, + "estimated_input_tokens": 2175, + "estimated_eval_tokens": 4606 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 2, + "serialized_chars": 8699, + "estimated_input_tokens": 2175, + "estimated_eval_tokens": 4606 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three unrelated model-support PRs (Qwen3-ASR, Lfm2.5 Audio, and Qwen3.5 GGUF loading). They share a broad audio/LLM-adjacent ecosystem but not the same underlying bug or code-path change, so the soft links should be rejected.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 43838 is the strongest representative: it adds a full Qwen3-ASR integration end-to-end (model/config/processor/autos/docs/tests) and explicitly targets issue 43837.", + "best_issue_reason": null, + "best_pr_reason": "PR 43838 is the best single representative of the cluster because it is the most complete and clearly scoped model-support implementation, with substantial code, docs, and test coverage.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43838", + "right": "pull_request:45283", + "accept": false, + "reason": "Different changes: Qwen3-ASR model integration vs Qwen3.5 GGUF loading support. They do not fix the same concrete code-path problem." + }, + { + "left": "pull_request:43973", + "right": "pull_request:45283", + "accept": false, + "reason": "Different model additions and different mechanisms: Lfm2.5 Audio support vs Qwen3.5 GGUF mapping/loading. Same broad area only, not 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": 17, + "component_count": 25, + "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:44018|pull_request:44068", + "pull_request:44066|pull_request:44071", + "pull_request:44066|pull_request:44068", + "pull_request:44018|pull_request:44066", + "pull_request:44019|pull_request:44722", + "pull_request:44086|pull_request:44722", + "pull_request:44019|pull_request:44085", + "pull_request:44018|pull_request:44019", + "pull_request:44019|pull_request:44071", + "pull_request:44068|pull_request:44116", + "pull_request:44068|pull_request:44722", + "pull_request:44018|pull_request:44116", + "pull_request:44071|pull_request:44722", + "pull_request:44018|pull_request:44086", + "pull_request:44018|pull_request:44085", + "pull_request:44068|pull_request:44086", + "pull_request:44071|pull_request:44085", + "pull_request:44059|pull_request:44086", + "pull_request:44085|pull_request:44086", + "pull_request:44059|pull_request:44722", + "pull_request:44098|pull_request:44101", + "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:44059|pull_request:44071", + "pull_request:44017|pull_request:44071", + "pull_request:43997|pull_request:44073", + "pull_request:44059|pull_request:44068", + "pull_request:44073|pull_request:44101", + "pull_request:44018|pull_request:44059", + "pull_request:44056|pull_request:44101", + "pull_request:44073|pull_request:44161", + "pull_request:44101|pull_request:44161", + "pull_request: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:44010|pull_request:44073", + "pull_request:44018|pull_request:44129", + "pull_request:43997|pull_request:44074", + "pull_request:43997|pull_request:44129", + "pull_request:44010|pull_request:44018", + "pull_request:44071|pull_request:44129", + "pull_request:43996|pull_request:44161", + "pull_request:43996|pull_request:44017", + "pull_request:44010|pull_request:44044", + "pull_request:44044|pull_request:44074", + "pull_request:44066|pull_request:44722", + "pull_request:44085|pull_request:44722", + "pull_request:44085|pull_request:44129", + "pull_request:43996|pull_request:44007", + "pull_request:43996|pull_request:44129", + "pull_request:44044|pull_request:44129", + "pull_request:44044|pull_request:44154" + ], + "prepared_review_unit_hash": "8f094394d71060f2893abafe4167f03e441f178cf5ad94cf3565d8d91bcd2805", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84896, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 + }, + "budget": { + "node_count": 27, + "item_count": 27, + "soft_pair_count": 72, + "serialized_chars": 75094, + "estimated_input_tokens": 18774, + "estimated_eval_tokens": 37804 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "This cluster is mostly a family of model-specific refactors for standardized output tracing, all loosely tied to issue #43979, but the PRs are not all duplicates of each other. The clearest duplicate subgroup is the GPT-J/CodeGen line (#44066, #44085, #44722), while the GPT-Neo line (#44018, #44068) is separate.", + "confidence": 0.67, + "canonical_issue_reason": "All items are pull requests; the only issue reference is the external tracker #43979, so there is no issue artifact in this cluster to choose as canonical.", + "canonical_pr_reason": "PR #44722 is the strongest representative of the GPT-J/CodeGen subgroup: it is the later, broader revision, explicitly targets #43979, and overlaps the same GPT-J output-tracing code path as #44066 and #44085.", + "best_issue_reason": "No issue node is present here; #43979 is only referenced externally, so there is no in-cluster issue candidate to rank as best.", + "best_pr_reason": "#44722 is the best global representative among the accepted duplicates because it captures the same GPT-J tracing refactor as #44066/#44085 while being the most complete revision in that subgroup. The rest of the cluster is split across separate model-specific refactors.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43996", + "right": "pull_request:44085", + "accept": false, + "reason": "Both are output-tracing refactors, but they touch different model families (CVT/FNet vs GPT-J) and different code paths, so they are not the same concrete change." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44044", + "accept": false, + "reason": "Different models and implementations; this is a thematic similarity only, not the same underlying bug or patch." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44085", + "accept": true, + "reason": "Both target GPT-J output tracing in the same file and same code path; #44066 also includes CodeGen, but the GPT-J refactor overlaps enough to plausibly merge as one change." + }, + { + "left": "pull_request:44007", + "right": "pull_request:44072", + "accept": false, + "reason": "These are separate image-model refactors (RegNet/ResNet/RT-DETR vs EfficientNet) and do not patch the same concrete implementation." + }, + { + "left": "pull_request:44072", + "right": "pull_request:44722", + "accept": false, + "reason": "Different subsystems and files; they share the generic output-tracing theme but not the same code-path problem." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44072", + "accept": false, + "reason": "GPT-J/CodeGen vs EfficientNet are unrelated model paths; same refactor style is not enough to treat them as duplicates." + }, + { + "left": "pull_request:44013", + "right": "pull_request:44044", + "accept": false, + "reason": "MobileNetV2 and DeBERTa v2 are different models with different output-handling code; not mergeable as one fix." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J/CodeGen output tracing and MGP-STR output tracing are separate implementations in different model stacks." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44068", + "accept": true, + "reason": "Both are GPT-Neo output-tracing refactors in the same model/file family and appear to be the same underlying change with slightly different revisions." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-J/CodeGen and MPT are different model code paths; similarity is only at the architectural refactor level." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44068", + "accept": false, + "reason": "Different models (GPT-J vs GPT-Neo), so despite similar wording they are not the same concrete patch." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44066", + "accept": false, + "reason": "GPT-Neo and GPT-J/CodeGen are distinct model implementations; the overlap is only the shared tracing pattern." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44722", + "accept": false, + "reason": "ResNet output tracing and GPT-J/CodeGen output tracing are unrelated concrete changes." + }, + { + "left": "pull_request:44086", + "right": "pull_request:44722", + "accept": false, + "reason": "MGP-STR and GPT-J/CodeGen are separate model-specific changes; not the same underlying bug or patch." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44085", + "accept": false, + "reason": "Different model families and code paths; they do not fix the same concrete issue." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44019", + "accept": false, + "reason": "GPT-Neo vs ResNet are separate model implementations, so these are not duplicates." + }, + { + "left": "pull_request:44019", + "right": "pull_request:44071", + "accept": false, + "reason": "ResNet and MPT output tracing are unrelated patches in different files." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo and Flaubert/XLM are different model stacks; same theme, different code paths." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-Neo output tracing is a separate subgroup from GPT-J/CodeGen output tracing." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44116", + "accept": false, + "reason": "GPT-Neo vs Flaubert/XLM are different concrete implementations, not one shared fix." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-J and MGP-STR are unrelated model code paths." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44086", + "accept": false, + "reason": "GPT-2 and MGP-STR are different models with different tracing logic; not the same change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44086", + "accept": false, + "reason": "The overlap is only broad output-tracing refactoring; the affected code paths are different model families." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44722", + "accept": false, + "reason": "GPT-2 output tracing is a separate implementation from GPT-J/CodeGen." + }, + { + "left": "pull_request:44098", + "right": "pull_request:44101", + "accept": false, + "reason": "ViLT and Flaubert/XLM are different model families and not the same concrete patch." + }, + { + "left": "pull_request:44018", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are distinct model implementations; this is not a duplicate fix." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44722", + "accept": false, + "reason": "SegFormer and GPT-J/CodeGen patches are unrelated beyond the shared decorator-style refactor theme." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44161", + "accept": false, + "reason": "SegFormer and LongT5 touch different model code paths and cannot be treated as one change." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44073", + "accept": false, + "reason": "SegFormer and VisualBert are separate models; same general pattern does not make them duplicates." + }, + { + "left": "pull_request:44068", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-Neo and MPT are different model families with different output-capture implementations." + }, + { + "left": "pull_request:44071", + "right": "pull_request:44086", + "accept": false, + "reason": "MPT and MGP-STR are unrelated concrete patches." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44071", + "accept": false, + "reason": "GPT-2 and MPT are separate model implementations and do not target the same code path." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44071", + "accept": false, + "reason": "SegFormer and MPT are different model families; not mergeable as one PR." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44073", + "accept": false, + "reason": "RegNet and VisualBert are different model code paths, so this is only a thematic similarity." + }, + { + "left": "pull_request:44059", + "right": "pull_request:44068", + "accept": false, + "reason": "GPT-2 and GPT-Neo are distinct models; same refactor style does not imply duplicate content." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44101", + "accept": false, + "reason": "MPNet and Flaubert/XLM are separate implementations with different output handling." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44161", + "accept": false, + "reason": "VisualBert and LongT5 are unrelated model-specific changes." + }, + { + "left": "pull_request:44101", + "right": "pull_request:44161", + "accept": false, + "reason": "Flaubert/XLM and LongT5 touch different code paths and are not the same fix." + }, + { + "left": "pull_request:44017", + "right": "pull_request:44018", + "accept": false, + "reason": "SegFormer and GPT-Neo are different models; shared refactor language is insufficient." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44056", + "accept": false, + "reason": "RegNet and MPNet are separate model implementations with no shared concrete patch." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44071", + "accept": false, + "reason": "SqueezeBert and MPT are unrelated model code paths." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44073", + "accept": false, + "reason": "SqueezeBert and VisualBert are different models, so they are not duplicates." + }, + { + "left": "pull_request:44066", + "right": "pull_request:44722", + "accept": true, + "reason": "Both touch the same GPT-J output-tracing code path, and #44722 is a broader later revision that overlaps the same concrete change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44722", + "accept": true, + "reason": "Both are GPT-J tracing refactors in the same file; #44722 is the more complete version of the same underlying change." + }, + { + "left": "pull_request:44085", + "right": "pull_request:44129", + "accept": false, + "reason": "GPT-J and SpeechT5 are different model families with different output-tracing logic." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44007", + "accept": false, + "reason": "CVT/FNet and RegNet/ResNet/RT-DETR are different model code paths; not the same underlying fix." + }, + { + "left": "pull_request:43996", + "right": "pull_request:44129", + "accept": false, + "reason": "CVT/FNet and SpeechT5 are unrelated implementations beyond the generic tracing refactor theme." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44129", + "accept": false, + "reason": "DeBERTa v2 and SpeechT5 patch different model code paths." + }, + { + "left": "pull_request:44044", + "right": "pull_request:44154", + "accept": false, + "reason": "DeBERTa v2 and VITS are separate model implementations and not the same concrete change." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: it correctly treats this as a mostly model-specific output-tracing refactor cluster tied to #43979, keeps the GPT-J/CodeGen and GPT-Neo subgroups separate, and does not overmerge unrelated model families. The accepted soft edges also look mergeable as single PRs for the same bugfix within the same model/file family." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 17, + "component_count": 25, + "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": "4a3c93a7dc00b9269dfc12f1a55dc5b2b82b9c738f3431b16e68fb1b00d6d730", + "component_budget": { + "node_count": 28, + "item_count": 28, + "soft_pair_count": 78, + "serialized_chars": 84896, + "estimated_input_tokens": 21224, + "estimated_eval_tokens": 42704 + }, + "budget": { + "node_count": 8, + "item_count": 8, + "soft_pair_count": 6, + "serialized_chars": 17123, + "estimated_input_tokens": 4281, + "estimated_eval_tokens": 8818 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": true, + "analyst_result": { + "summary": "These PRs all belong to the same broad housekeeping theme\u2014migrating individual model implementations to standardized output capture\u2014but they affect different models and code paths, so they are not duplicates of one another. None of the soft edges should be merged under strict triage.", + "confidence": 0.94, + "canonical_issue_reason": null, + "canonical_pr_reason": "No PR is a duplicate of the others; if one must be chosen as the cluster representative, 44073 is the clearest exemplar because it explicitly wires in both capture_outputs and can_return_tuple plus model-specific output hooks.", + "best_issue_reason": null, + "best_pr_reason": "44073 is the best single representative of the theme because it shows the standardized output-tracing refactor in a concrete, self-contained model implementation. It is more representative than the smaller or more specialized variants, but it is still not a duplicate of them.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44024", + "right": "pull_request:44076", + "accept": false, + "reason": "Different models (FocalNet vs ImageGPT) and different code paths; same refactor theme is not enough to treat them as the same change." + }, + { + "left": "pull_request:44073", + "right": "pull_request:44074", + "accept": false, + "reason": "VisualBert and TextNet are unrelated model implementations; these are parallel standardized-output refactors, not one concrete bug fix or mergeable duplicate PR." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44073", + "accept": false, + "reason": "MPNet and VisualBert touch different architectures and output-capture wiring; they are not the same underlying change." + }, + { + "left": "pull_request:44056", + "right": "pull_request:44074", + "accept": false, + "reason": "MPNet vs TextNet are separate model refactors, not a single fix to one code path." + }, + { + "left": "pull_request:43997", + "right": "pull_request:44010", + "accept": false, + "reason": "RegNet and SqueezeBert are different models with different implementation details; this is only a thematic similarity." + }, + { + "left": "pull_request:44010", + "right": "pull_request:44129", + "accept": false, + "reason": "SqueezeBert and SpeechT5 modify different model stacks and output semantics; not a duplicate or mergeable single 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": 18, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-43998-13", + "nodes": [ + "pull_request:43998", + "pull_request:43999", + "pull_request:44000", + "pull_request:44001", + "pull_request:44002", + "pull_request:44003", + "pull_request:44004", + "pull_request:44025", + "pull_request:44026", + "pull_request:44027", + "pull_request:44028", + "pull_request:44029", + "pull_request:44030" + ], + "soft_pairs": [ + "pull_request:43998|pull_request:44002", + "pull_request:43998|pull_request:44000", + "pull_request:43998|pull_request:44001", + "pull_request:43998|pull_request:44003", + "pull_request:43998|pull_request:44029", + "pull_request:43999|pull_request:44001", + "pull_request:43999|pull_request:44002", + "pull_request:43999|pull_request:44025", + "pull_request:43999|pull_request:44026", + "pull_request:43999|pull_request:44027", + "pull_request:44000|pull_request:44001", + "pull_request:44000|pull_request:44002", + "pull_request:44000|pull_request:44003", + "pull_request:44000|pull_request:44025", + "pull_request:44001|pull_request:44002", + "pull_request:44001|pull_request:44003", + "pull_request:44001|pull_request:44025", + "pull_request:44002|pull_request:44029", + "pull_request:44002|pull_request:44003", + "pull_request:44003|pull_request:44004", + "pull_request:43998|pull_request:44004", + "pull_request:44001|pull_request:44004", + "pull_request:44002|pull_request:44004", + "pull_request:44000|pull_request:44004", + "pull_request:44025|pull_request:44026", + "pull_request:44025|pull_request:44027", + "pull_request:44025|pull_request:44028", + "pull_request:44025|pull_request:44029", + "pull_request:44025|pull_request:44030", + "pull_request:44026|pull_request:44029", + "pull_request:44026|pull_request:44027", + "pull_request:44026|pull_request:44028", + "pull_request:44026|pull_request:44030", + "pull_request:44027|pull_request:44028", + "pull_request:44027|pull_request:44029", + "pull_request:44027|pull_request:44030", + "pull_request:44028|pull_request:44029", + "pull_request:44028|pull_request:44030", + "pull_request:44029|pull_request:44030" + ], + "prepared_review_unit_hash": "9a231e6af7b9802507e4562d09b573c6f1466a88f7f2b426ee9f2efba0b5cdc8", + "component_budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39781, + "estimated_input_tokens": 9946, + "estimated_eval_tokens": 20148 + }, + "budget": { + "node_count": 13, + "item_count": 13, + "soft_pair_count": 39, + "serialized_chars": 39781, + "estimated_input_tokens": 9946, + "estimated_eval_tokens": 20148 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All items are separate model-specific output-tracing refactors. They share the same cleanup pattern, but they do not fix the same concrete code path and should not be deduplicated into one PR.", + "confidence": 0.97, + "canonical_issue_reason": null, + "canonical_pr_reason": "No single PR is a true duplicate canonical here; the cluster is a set of parallel refactors across different model implementations.", + "best_issue_reason": null, + "best_pr_reason": "PR 44025 is the best representative of the cluster theme because it is a focused multi-file output-tracing refactor, but it is still distinct from the others.", + "soft_edge_verdicts": [ + { + "left": "pull_request:43998", + "right": "pull_request:44002", + "accept": false, + "reason": "Both are output-tracing refactors, but they touch different model code paths (`timm_backbone` vs `upernet`) and are not mergeable as one fix." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44000", + "accept": false, + "reason": "Same refactor theme, but different models and forward paths (`timm_backbone` vs `vision_text_dual_encoder`)." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44001", + "accept": false, + "reason": "Shared cleanup pattern only; `timm_backbone` and `univnet` are unrelated model implementations." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44003", + "accept": false, + "reason": "Both are refactors, but they affect different architectures (`timm_backbone` vs `falcon_mamba`/`mamba`) and different concrete code paths." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44029", + "accept": false, + "reason": "These are separate model-specific output-tracing changes (`timm_backbone` vs `rwkv`), not the same underlying bug or change." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44001", + "accept": false, + "reason": "Both modernize output handling, but `mobilenet_v1` and `univnet` are distinct implementations and cannot be merged into one PR." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44002", + "accept": false, + "reason": "Same high-level pattern, but `mobilenet_v1` and `upernet` are different code paths and model families." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44025", + "accept": false, + "reason": "Both are output-tracing refactors, but they target different models (`mobilenet_v1` vs `depth_anything`) and different forward signatures." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44026", + "accept": false, + "reason": "These are separate model refactors (`mobilenet_v1` vs `vision_encoder_decoder`), not one concrete fix." + }, + { + "left": "pull_request:43999", + "right": "pull_request:44027", + "accept": false, + "reason": "Shared refactor style only; `mobilenet_v1` and `speech_encoder_decoder` are unrelated code paths." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44001", + "accept": false, + "reason": "Both are output-tracing refactors, but they change different model stacks (`vision_text_dual_encoder` vs `univnet`)." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44002", + "accept": false, + "reason": "Different models and different forward implementations (`vision_text_dual_encoder` vs `upernet`)." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44003", + "accept": false, + "reason": "Although both are tracing refactors, they apply to unrelated model code paths and cannot plausibly be merged into one PR." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44025", + "accept": false, + "reason": "Same theme, but `vision_text_dual_encoder` and `depth_anything` are different model families with separate fixes." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44002", + "accept": false, + "reason": "Both are output-tracing cleanup, but they affect different models (`univnet` vs `upernet`) and separate forward contracts." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44003", + "accept": false, + "reason": "These are not the same bug/change; they target different model implementations (`univnet` vs `mamba`/`falcon_mamba`)." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44025", + "accept": false, + "reason": "Different model families and different code paths; same refactor pattern is not enough for a duplicate." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44029", + "accept": false, + "reason": "`upernet` and `rwkv` are unrelated model implementations; this is only a thematic similarity." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44003", + "accept": false, + "reason": "Shared output-tracing cleanup, but the underlying forward logic lives in different models and is not mergeable as one PR." + }, + { + "left": "pull_request:44003", + "right": "pull_request:44004", + "accept": false, + "reason": "Different subsystems entirely (`mamba`/`falcon_mamba` vs `codegen`), so these are not the same concrete change." + }, + { + "left": "pull_request:43998", + "right": "pull_request:44004", + "accept": false, + "reason": "Both touch output handling, but `timm_backbone` and `codegen` are unrelated code paths." + }, + { + "left": "pull_request:44001", + "right": "pull_request:44004", + "accept": false, + "reason": "These are separate refactors in different model families (`univnet` vs `codegen`)." + }, + { + "left": "pull_request:44002", + "right": "pull_request:44004", + "accept": false, + "reason": "`upernet` and `codegen` do not share the same underlying fix; only the refactor motif matches." + }, + { + "left": "pull_request:44000", + "right": "pull_request:44004", + "accept": false, + "reason": "Different models and different forward contracts; not a duplicate PR pair." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44026", + "accept": false, + "reason": "Both are output-tracing refactors, but they target different wrappers (`depth_anything` vs `vision_encoder_decoder`) and distinct forward paths." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44027", + "accept": false, + "reason": "`depth_anything` and `speech_encoder_decoder` are separate model-specific changes, not one shared fix." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model implementations (`depth_anything` vs `superpoint`), so this is not the same underlying bug or change." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44029", + "accept": false, + "reason": "These are parallel output-tracing refactors in different models (`depth_anything` vs `rwkv`)." + }, + { + "left": "pull_request:44025", + "right": "pull_request:44030", + "accept": false, + "reason": "Different code paths and model families (`depth_anything` vs `dpr`); same pattern, not same PR." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44029", + "accept": false, + "reason": "Both are refactors around model outputs, but `vision_encoder_decoder` and `rwkv` are unrelated implementations." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44027", + "accept": false, + "reason": "`vision_encoder_decoder` and `speech_encoder_decoder` are similar wrapper types, but they are distinct code paths and separate changes." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44028", + "accept": false, + "reason": "Different model families (`vision_encoder_decoder` vs `superpoint`); not one concrete bugfix." + }, + { + "left": "pull_request:44026", + "right": "pull_request:44030", + "accept": false, + "reason": "Separate model-specific refactors (`vision_encoder_decoder` vs `dpr`), so they should not be merged as duplicates." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44028", + "accept": false, + "reason": "These touch different models (`speech_encoder_decoder` vs `superpoint`) and do not represent the same underlying change." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44029", + "accept": false, + "reason": "Unrelated model code paths (`speech_encoder_decoder` vs `rwkv`), despite sharing the same refactor theme." + }, + { + "left": "pull_request:44027", + "right": "pull_request:44030", + "accept": false, + "reason": "Different model-specific output handling changes; not a duplicate pair." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44029", + "accept": false, + "reason": "`superpoint` and `rwkv` are unrelated code paths, so this is only superficial similarity." + }, + { + "left": "pull_request:44028", + "right": "pull_request:44030", + "accept": false, + "reason": "Different models and different forward implementations (`superpoint` vs `dpr`)." + }, + { + "left": "pull_request:44029", + "right": "pull_request:44030", + "accept": false, + "reason": "Both are model-specific refactors, but they affect different architectures (`rwkv` vs `dpr`) and are not mergeable into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 19, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44662-3", + "nodes": [ + "pull_request:44662", + "pull_request:45075", + "pull_request:45401" + ], + "soft_pairs": [ + "pull_request:44662|pull_request:45401", + "pull_request:44662|pull_request:45075", + "pull_request:45075|pull_request:45401" + ], + "prepared_review_unit_hash": "1ae9d8cacd76a707d18b86d11c11f686fb0d37db393af09f8a8ed476def6f407", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 11071, + "estimated_input_tokens": 2768, + "estimated_eval_tokens": 5792 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 11071, + "estimated_input_tokens": 2768, + "estimated_eval_tokens": 5792 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are three separate model-integration pull requests for different models (PenguinVL, DeepSeek-OCR-2, Voxtral TTS). They share common Transformers plumbing files, but the implementations, model classes, and targets are distinct, so they are not duplicates.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45075 is the most mature-looking standalone integration: it is not marked draft, has substantial review activity, and includes a full model package plus tests. It is not a duplicate of the others, but it is the best single representative if one must be chosen from this cluster.", + "best_issue_reason": null, + "best_pr_reason": "PR #45075 is the strongest representative for the cluster because it is a complete, non-draft model addition with substantial review activity. The other PRs are also standalone model additions, but they are separate changes for different architectures and should not be merged into one.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44662", + "right": "pull_request:45401", + "accept": false, + "reason": "Different models and code paths: PenguinVL vs Voxtral TTS. Shared auto/docs registration files are expected for independent model additions and do not make these the same change." + }, + { + "left": "pull_request:44662", + "right": "pull_request:45075", + "accept": false, + "reason": "Different model integrations: PenguinVL and DeepSeek-OCR-2. Overlap is limited to shared Transformers registry/docs boilerplate, not the same underlying bug or feature." + }, + { + "left": "pull_request:45075", + "right": "pull_request:45401", + "accept": false, + "reason": "DeepSeek-OCR-2 and Voxtral TTS are unrelated model additions with different implementations and targets. The shared files are generic registration updates, not evidence of one concrete mergeable change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 20, + "component_count": 25, + "review_unit_index": 1, + "review_unit_count": 1, + "status": "reviewed", + "reason": null, + "source": "llm", + "cache_hit": false, + "model": "gpt-5.4-mini?service_tier=flex", + "cluster_id": "cluster-44979-2", + "nodes": [ + "pull_request:44979", + "pull_request:45363" + ], + "soft_pairs": [ + "pull_request:44979|pull_request:45363" + ], + "prepared_review_unit_hash": "3d300441ed7501b66471d4b3bfff21be52eb6f034b5b703b375894d6cb13dffd", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4407, + "estimated_input_tokens": 1102, + "estimated_eval_tokens": 2460 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are related to model/kernel fusion, but they address different layers of the stack: one introduces a generic module fusion API, while the other adds a specific n-to-1 kernel-fusion path via KernelConfig and integration changes. They are not duplicates.", + "confidence": 0.95, + "canonical_issue_reason": null, + "canonical_pr_reason": null, + "best_issue_reason": null, + "best_pr_reason": "PR 45363 is the more concrete end-user feature: it wires fusion into existing loading/integration code paths and includes an example, making it the better standalone representative of the cluster.", + "soft_edge_verdicts": [ + { + "left": "pull_request:44979", + "right": "pull_request:45363", + "accept": false, + "reason": "Different concrete changes: 44979 adds a generic Module Fusion API and tests, while 45363 implements n-to-1 kernel fusion through KernelConfig and integration plumbing. Related topic, but not the same code-path change." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 21, + "component_count": 25, + "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": "11ee7051d443ee9ae38db57d12f6b21717af7160b6450dfca79d45f7071e2e98", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4414, + "estimated_input_tokens": 1104, + "estimated_eval_tokens": 2464 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "Both pull requests target the same Mistral regex patch in the same file, with essentially identical code changes; PR 45317 is the more complete version because it also adds tests.", + "confidence": 0.99, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45317 is the better canonical PR because it makes the same fix as 45086 and includes accompanying test coverage, so it is the more complete representative of the duplicate pair.", + "best_issue_reason": null, + "best_pr_reason": "PR 45317 is the best representative of the cluster: same underlying fix as 45086, plus tests, making it the more merge-ready and informative duplicate target.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45086", + "right": "pull_request:45317", + "accept": true, + "reason": "Same concrete bug fix in the same code path (_patch_mistral_regex / tokenization_utils_tokenizers.py); the diffs are effectively the same, and the second PR just adds tests, so they plausibly could be merged 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": 22, + "component_count": 25, + "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": "8ffe091d15de12d586690691501506c4eab49839b4bd71825ec8ffce82e4bbeb", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5398, + "estimated_input_tokens": 1350, + "estimated_eval_tokens": 2956 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are closely related heterogeneous-feature PRs, but they are not the same change: one adds per-layer config plumbing, the other adds broader heterogeneous model/runtime support. They overlap in config utilities and tests, yet they target different parts of the codepath and could reasonably land as separate PRs.", + "confidence": 0.91, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR #45332 is the broader umbrella change, covering heterogeneous modeling/runtime support in addition to config plumbing, while #45333 is the narrower config-only foundation.", + "best_issue_reason": null, + "best_pr_reason": "PR #45332 is the best representative of the cluster because it contains the larger end-to-end heterogeneous support work and subsumes the config changes used by #45333.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45332", + "right": "pull_request:45333", + "accept": false, + "reason": "Related feature work, but not the same concrete change: #45333 adds per-layer config support, while #45332 adds heterogeneous modeling/cache/runtime support. They overlap in shared helpers, yet they are separate slices of the feature and not safely mergeable 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": 23, + "component_count": 25, + "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-45351-3", + "nodes": [ + "pull_request:45351", + "pull_request:45509", + "pull_request:45515" + ], + "soft_pairs": [ + "pull_request:45351|pull_request:45509", + "pull_request:45351|pull_request:45515", + "pull_request:45509|pull_request:45515" + ], + "prepared_review_unit_hash": "12efe7296ae86d20a9691bb3beafb427279ef36b4754af0e6721218cdd6f9c1f", + "component_budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5821, + "estimated_input_tokens": 1456, + "estimated_eval_tokens": 3168 + }, + "budget": { + "node_count": 3, + "item_count": 3, + "soft_pair_count": 3, + "serialized_chars": 5821, + "estimated_input_tokens": 1456, + "estimated_eval_tokens": 3168 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "All three pull requests target the same crash in `get_device_properties()` when CUDA/ROCm is detected but `torch.cuda.is_available()` is false. They are close variants of the same fix, with PR 45351 the most complete/canonical version.", + "confidence": 0.9, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45351 is the strongest canonical PR: it explicitly links the tracking issue, guards the CUDA/ROCm capability call with `torch.cuda.is_available()`, and keeps the XPU path intact in the same function.", + "best_issue_reason": null, + "best_pr_reason": "PR 45351 is the best representative of the cluster because it is the most complete and coherent fix for the exact crash path, while the other two are narrower variants of the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45351", + "right": "pull_request:45509", + "accept": true, + "reason": "Same underlying bug in `get_device_properties()`: avoid calling CUDA capability APIs when CUDA is present but unavailable. Both PRs patch the same code path and could plausibly be consolidated." + }, + { + "left": "pull_request:45351", + "right": "pull_request:45515", + "accept": true, + "reason": "Same crash fix in the same function, with nearly identical logic around `torch.cuda.is_available()`. These are alternate implementations of the same change." + }, + { + "left": "pull_request:45509", + "right": "pull_request:45515", + "accept": true, + "reason": "Both target the same `get_device_properties()` availability guard for CUDA/ROCm. They address the same concrete failure mode and are mergeable into one PR." + } + ] + }, + "evaluator_result": { + "accept": true, + "feedback": "The summary is grounded in the packet: all three PRs touch the same `get_device_properties()` path in `src/transformers/testing_utils.py` and address the CUDA/ROCm availability crash. Calling PR 45351 the most canonical/comprehensive version is reasonable because it has the explicit issue link and preserves the XPU branch in the same refactor. The pairwise soft-edge accepts are also defensible as the PRs look like close variants of the same fix." + }, + "evaluator_used": true, + "retried": false, + "accepted_nontrivial_soft_edge": true, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 24, + "component_count": 25, + "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": "09cbc2edb152fd0953da59dc909841a68059cdd04df707cfd3f0c1fa39f448ec", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5306, + "estimated_input_tokens": 1327, + "estimated_eval_tokens": 2910 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 5306, + "estimated_input_tokens": 1327, + "estimated_eval_tokens": 2910 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These PRs are both typing-related, but they address different changes: 45415 is a broad repo-wide type-checking sweep, while 45425 adds specific typing protocols/helpers around modeling_utils/PEFT. They do not appear to fix the same concrete code-path problem, so I would not cluster them as duplicates.", + "confidence": 0.93, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45425 is the more focused and concrete typing change; it better represents the cluster than the broad type-checking sweep in 45415.", + "best_issue_reason": null, + "best_pr_reason": "PR 45425 is the best representative because it makes a specific typing/API adjustment in a narrower area, whereas 45415 is a large mechanical type-checking pass.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45415", + "right": "pull_request:45425", + "accept": false, + "reason": "Shared typing-related files are not enough here: 45415 is a broad annotation/type-checking sweep across many modules, while 45425 adds targeted typing protocols and modeling_utils integration. They are not the same underlying change and would not plausibly merge into one PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + }, + { + "label": "pull_request", + "component_index": 25, + "component_count": 25, + "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-45470-2", + "nodes": [ + "pull_request:45470", + "pull_request:45487" + ], + "soft_pairs": [ + "pull_request:45470|pull_request:45487" + ], + "prepared_review_unit_hash": "1c2d791880bc098c4a5b373854e202acece21f45d93cc94dea699fc37380d767", + "component_budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4769, + "estimated_input_tokens": 1193, + "estimated_eval_tokens": 2642 + }, + "budget": { + "node_count": 2, + "item_count": 2, + "soft_pair_count": 1, + "serialized_chars": 4769, + "estimated_input_tokens": 1193, + "estimated_eval_tokens": 2642 + }, + "overflow_policy": "truncate_then_skip", + "trimmed": false, + "aggressively_trimmed": false, + "split": false, + "analyst_result": { + "summary": "These are not duplicates: one PR is a targeted test skip for Sam3LiteText flash-attn dispatch behavior, while the other is a broad model-parallel fix across multiple model implementations. They do not appear to fix the same concrete bug or code path.", + "confidence": 0.96, + "canonical_issue_reason": null, + "canonical_pr_reason": "PR 45487 is the stronger canonical PR because it makes an actual code-path fix across model implementations, whereas 45470 only skips a test in one model suite.", + "best_issue_reason": null, + "best_pr_reason": "PR 45487 is the best representative of the cluster since it addresses a substantive runtime/model-parallel issue; 45470 is just a test adjustment and is not the same change.", + "soft_edge_verdicts": [ + { + "left": "pull_request:45470", + "right": "pull_request:45487", + "accept": false, + "reason": "Reject: 45470 only skips a flash-attn dispatch test for Sam3LiteText, while 45487 changes model-parallel handling in several model files. Different underlying problems and not plausibly one merged PR." + } + ] + }, + "evaluator_result": null, + "evaluator_used": false, + "retried": false, + "accepted_nontrivial_soft_edge": false, + "error_kind": null, + "error_message": null + } + ] +} diff --git a/snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/manifest.json b/snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/manifest.json new file mode 100644 index 0000000000000000000000000000000000000000..b1baae99ec649b6409dca991184f26acac5c1fd2 --- /dev/null +++ b/snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/manifest.json @@ -0,0 +1,14 @@ +{ + "analysis_id": "hybrid-model-20260420t094431z", + "artifacts": { + "hybrid": "snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "model": null, + "published_at": "2026-04-20T09:57:25Z", + "repo": "huggingface/transformers", + "schema_version": 1, + "snapshot_id": "20260420T094431Z", + "variant": "hybrid" +} diff --git a/snapshots/20260420T094431Z/manifest.json b/snapshots/20260420T094431Z/manifest.json index ef38522fce89ab94a0b98520a624a6668bdc891e..c32de75a723e3aa7df6933d152bf6ce9b332a15b 100644 --- a/snapshots/20260420T094431Z/manifest.json +++ b/snapshots/20260420T094431Z/manifest.json @@ -32,6 +32,24 @@ "timeline_events": 643 }, "extracted_at": "2026-04-20T09:44:31Z", + "published_analysis": { + "canonical_analysis_id": "hybrid-model-20260420t094431z", + "runs": { + "hybrid-model-20260420t094431z": { + "analysis_id": "hybrid-model-20260420t094431z", + "artifacts": { + "hybrid": "snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.json", + "hybrid_reviews": "snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/analysis-report-hybrid.llm-reviews.json" + }, + "channel": "canonical", + "manifest_path": "snapshots/20260420T094431Z/analysis-runs/hybrid-model-20260420t094431z/manifest.json", + "model": null, + "published_at": "2026-04-20T09:57:25Z", + "variant": "hybrid" + } + }, + "schema_version": 1 + }, "repo": "huggingface/transformers", "snapshot_id": "20260420T094431Z", "watermark": {